EssaysAttachment 6

Command and Control and the Pathologyof Natural Resource ManagementC. S. HOLLING" AND GARY K. MEFFE~"*Department of Zoology, Urtiversit3" of Florida, Gainesvii!e. FL 3261 I, U.S.A."tUnive~iW of Georgia’s Savannah River Ecology Laboratory Drawer E, Aiken. SC 29802, U.S.A.

Abstract: As the hutnan population grou’s and natural resources decline, there is pressure to apply increas-Dig levels of top-down, command-and~:ontrol management to natural resources. This is manifested in at-tempts to control ecosystems and in socioeconomic institutions that respond to erratic or surprisDtg ecosystembehavior with more control. Cornmand and control, bou’et’er, usually results in unforeseen conseqttences forboth natural ecosystems artd human welfare in the form of collapsing resources, social and economic strife,and losses of biological diversity. We describe the "pathology of natural resource management," defined as aloss of system resilience when the range of natural variation in the ~’stern is reduced encapsula2es the uttsttstain-able environmental, social, and economic outcomes of command-and-control resource management. If natu-ral levels of variation in system behavior are reduced through command-and-control, then the system be-comes less resilient to external perturbations, resulting in crises and surprises. We provide several examples ofthis pathology in management. An ultimate pathology emerges when resource management agencies,through initial success with command and control, lose sight of their original purposes, eliminate researchand monitoring, and focus on efficiency of co. ntroL 27~e3, then become isolated from the managed s-ystems andinflexible in structure. SimultaneousI); through overcapitaltzation, society becomes dependent upon com-mand and control, demands it in greater intensity, and ignores the underlying ecological change or collapsethat is developing. Solutions to this pathology cannot dome from further comtnand attd cotjtrol (regttlations)but must come from inn. ovat[t,e approaches involving incentives leadhtg to more resilient ecosystents, moreflexible agencies, more self-reliant industries, and a more knou’ledgeable citizenry. We discuss several aspectsof ecosystem pattern and dynamics at large scales that provide insight into ecosystem resilience, and u’e pro-pose a "Golden Rule" of natural resource management that u’e believe is necessary for sttstainabilit):" man-agement should strive to retain critical O~es and ranges of natural variation in resource systems in order tomaintain their resiliency.

Comando-y-control y la patologia del manejo de los recursos naturales

Resumert: A medida que la pobla~i6n humana crece y los recursos naturales declinan, e.xisten presionespara aplicar niveles crecientes de manejo de recursos naturales verticalistas y de comando-y-control. Esto semanifesta en los intentos de cotttrolar los ecosistemas y en institucion~s socioecon6micas que resflonden a loscomportamientos errdticos o sorpresit,os de los ecosistemas con rods control. Sin embargo, el comando-y-con-trol tiene usualmente resttltados hnprevistos tanto para los ecosistemas natttrales comb para el bienestarmano, tales restdtados roman la forma de recursos que colapsan, conflictos sociales y econ6micos ~, pdrdidasde la diversidad biol6gica. En el presente trabajo, describDnos la "patologla del manejo de los recursos natu-rales" (definida COmb ttna pdrdida de la elasticidad del sistema cuando la magnitttd de la variaci6t~ natttralen el sistema es reducida), qtte condensa los resultados ambientales-sociales-econ6micos htsostenibles produ-cidos por el mattejo tie recursos con ttna 6ptica de cotnando-y-control. Si los niveles de variaci6n natural en elcomportamiento de an sistema sort reducidos a trat’ds de comando-y-controL entonces e[ sistema se bacemenos elastico a las perturbaciones e.xterrlaS, lo cual resulta en crisis y sorpresas. Nosotros prot,eemos de vat-los ejemplos de esta patologia en el manejo. Una patologfa e.xtrema sttrge cuando lets agencias de manejo derecursos, pierden de vista sus prop6sitos originales debido al dxito del uso de comando-3~control, eHminandola investigaci6n y el monitoreo y concentrdndose en la eficiencia y el control. De esta fomna, estas agencias seaislan de los sistemas bajo manejo y se hacen mas ittflexibles en su estructura. Sinztdtaneamente y por mediode [a sobrecapitalizacion, la sociedad se hace rods dependiente del comando-y-control, demandtt con ntayores

Address correspondence to G. K 3ttffe.Paper submitted April 24, 1995: revised manuscript accepted July 20. 1995.

328

Ctm~trrvation B ology, Pagt:s 328-337Volume 10.

E--025’1E-025~ 80

interzsidades e igrtora los cambios ecolK~gicos sttb.t’acetttes o el colttpso qtte se esttt desarrolhtttdo. Ltts soht-ciont,$ a esta Datoh~gia no pttetlert prot’enir de ttlt ma.l’or ~’omattdo-),~otztrol (rt,,~ltttttertto$). sitto qtte dr, bert

prot’entr de estrategias [~tttot,atit,tt$ qtte Dtt’ohtcren ht¢’e~ttit’os qtte llet,en a ecosistetnas mas elasticos, agett.cias ttt~’ts flt’.x’ibl~,s, ittdttstt-ias tttds ~tt~tosttficierttes y tttta ¢’itt~ktdatti, t tttt]s [ttst~tt&kt. Di~’ctttitnos t,~triostOS del Datrdrt ), di~dmit’a de los ecosistemas a gratt e~’caltt qtte proree~t tttttt comprettsi6tz de la elttst&’&htd tlt, lecoststema ), 19rolgOttemos una "re~la de oro" sol)re el manej’o de los recttrsos ~ttttttrale$, la cttttl Cret’ltto$ eSnecesaritt pars la soste~tibilidad." El mattejo debe esfor’~_arse pars retetter los tipos )" magrtitttdes de t’ariaci6tt

ntttttral crfticos eft los sistemas de recttrsos a los efectos tie mantener stt elttsticidad.

In ecolog),, we ba~’e an incredibO" comJ)lex s)’stem with pectafions, and increasingly on more rapidly developedno central dogma like tlaat of molecular bioh~KV to let sho~-~e~ ~centives ~nd controls. ~n the b~haviorsus eren pretend that u’e ba~’e control. Nowhere ~ thismore apparent than in cottse~’atio~t, tt’laere tt’e bat.e Of people, institutions, or nature violate the no~s, de-persuaded ourseh’es tt~at some degree of cotttrol ia" re- 5~e5, or expectations of socie~’, co--and and controlall), nece~’a~ is often sought as the primal" solution in an effo~ to~re~e/d. 1991

move human or ecosystem behaviors to a predeter-An esser~tial parado.r of u’ilde~ess conse~’ation is ~ed, predictable state. Consequently, much of naturalthat u’e seek to prese~’e tt@at mt~t change.

Pickett ~d ~te, 1985 resource management has been an effo~ to control na-ture ~ order to ha~’est its products, reduce its t~eats,~d establish hig~y predicmble outcomes for ~e sho~-

Introduction te~ benefit of humanly. O~ thesis is that adoption ofsuch co--and and ~ntrol has resdted ~ a patholo~

Control is a deeply entrenched aspect of contempom~"that pe~eates much of natu~ resource management

h~an societies: we control human behavior ~ou~and precludes long-te~ susta~abfliW.laws, ~centives, threats, contacts, ~d ageements; wecontrol the effects of env~o~entM vacation by con-st~g s~e dwell~gs; we control variation ~ o~ food Co~and-~d-Control M~ementreso~ces by ~ow~g ~d sto~g a~cul~ products;we control h~an pa~ites ~d pathogens t~ou~~e co~and-and~ontrol approach, when extendedgood hygiene and medical technolo~es. ~ contribute~cfitic~y to treatment of na~ resources, o~en re-to stable societies and h~an health and happ~ess, ~ds~ ~ u~oreseen and undes~ble consequences. A fre-wit~ ce~a~ arenas t~s desire to control is ~de~ablyquent, perhaps u~versal result of co--and and controlto o~ ~dividual and collective benefit. ~s approach~ applied to nat~I reso~ce management is reductionto solv~g problems may be coHectively refe~ed to asof the ~nge of natural va~ation of ~’stemg~the~ stmc-"co--and and control," ~ w~ch a problem is per-~e, ~nction, or both~ an attempt to ~crease theirceived and a solution for its control is developed ~d ~- predictabfli~ or stab~. ~at is, variation t~ou~ t~eplemented. ~e expectation is that the solution is direct,or space (such as ~’stem behavior over t~e, or spatialapprop~ate, feasible, and effective over most relevantheterogenei~’) is reduced. ~us. a co~on theme ofspatial and tempo~l scales, b~st of all, co--and ~dmany resource-management effo~s is to reduce nat~lcontrol is expected to solve the problem either t~ou~bounds of variation in ecological systems to make themcontrol of the processes that lead to the problem (e.g.,more predictable, and thus more reliable, for humangood hygiene to prevent disease, or laws that d~ect hu-need. We dampen extremes of ecosystem behavior orman behavioO or through amelio~tion of the problemchange species composition to a~a~ a predictable flow~er it occ~s (e.g., pha~aceutic~s to ~ dise~e or-of goods and see’ices or to reduce dest~ctive or unde-ganisms, or" p~sons or other puMs~ent of lawbreak-s~ble behavior of those systems. For example, we con-era). ~e co~and-and<ontrol approach ~plicitly a~trol a~cultuml pests t~ough herbicides and pesticides:sumes that the problem is well-bounded, clearly defined,we conve~ natu~l, multi-species, variable-aged forestsrelatively simple, and gene~lly linear with respect to~to monoculture, single-aged plantations; we hunt andcause and effect. But when these same methods of con-~ pre~tors to produce a larger, more reliable supply~ol are applied to a complex, noM~ear, and poorly un-of game species; we suppress F~es and pest outbreaks ~demtood natural world, and when the same predictableforests to ensure a steady lumber supply; we clear for-outcomes are expected but rarely obtained, severe ec~ests for p~ture development and stead)" cattle produc-logical, social, and economic repercussions result, tion, and so fo~h.

Humaniw’s contempom~ interactions with nature are Such eft~ms attempt to replace natu~l ecological con-based on a mLx of slowly developed social no~s and ex-trois, which are largely u~own to us and hig~y com-

Ctm~’atio~ Bit~h~gyV,~lume 10, No. 2. April 19~X~

,,

E--0251 81E-025181

330 patt~lo~" of ,Vatural Resource Management Ho/hng (- .~t~e

plex and variable, with engineered constructs and ma-on efficiency, constancy, and predictabiliw--attributesnipulations that on the surface seem entirel.v within ourat the core of command-and-control desires for fail-safecontrol. The purpose is to turn an unpredictable and ~in-design. The second focuses on persistence, change, andefficient" natural system into one that produces prod- unpredictabiliry--attributes embraced by an adaptive,cts in a predictable and economically efficient way.management philosophy. Hoiling (1973) first empha-

unanticipated environmental problems then arise,sized the consequences of these different definitions forthe a prt’ori expectation of certainW is not met and re-ecological systems in order to draw attention to the par-suits in surprise and crisis--chemical pollution and ero-adoxes between constancy and change or between pre-sion from monocultures, loss of biological diversity fromdictabilit3.’ and unpredictabiliW.tree farms, irruption of herbivore populations after pred-ator removal, conflagrations and properw loss whenfires finally erupt, insect pest outbreaks when sprayingThe Pathologystops, and pollution and erosion from grazing. Such cri-ses and surprises, we argue here, are the inevitable con-For illustrative purposes we offer several examples ofsequences of a command-and-control approach to re-the patholog2,.° of natural resource management in whichnewable resource management, where it is (implicitly orreduction of variation has led to a less resilient system,explicitly) believed that humans can select one compo-in the sense of ecosystem resilience:nent of a self-sustaining natural system and change it to a (1) The loss of genetic variation in small populations isfundamentally different configuration in which the ad- generally thought to result in a less resilient genetic sys-justed STstem remains in that new configuration indefi-tern (Allendorf & Lear~- 1986; Meffe 1986), possibly re-nitely without other, related changes in the larger system,sulting in higher probabilities of population extinction.

We call the result "the pathology of natural resource This is particularly true if the environment changes andmanagement~ (Holling 1986; Hol.llng 1995), a simple butpreviously available genoWpes that wo.uld be appropri-far-reaching observation deemed here as follows: whenate t6 the new environment no longer exist. Of coursethe range of natural variation in a systenz is reduced, there are exceptions: for example, loss of deleterious re-the system loses resilience. That is, a ~-stem in whichcessive alleles is not likely to reduce population resil-natural levels of variation have been reduced throughience and may in fact increase it. But overall, loss of ge-command-and-control activities will be less resilient thannetic variance may lead to lower population resilience inan unaltered system when subsequently faced with ex-ecological or evolutionary time.

perturbations, either of a natural (storms, Ftres, (2) StabiliZation of flows by dams in previousl.v wildlyor human-induced (social or institutional) origin, flooding or ~flashy" southwestern U.S. rivers results in a

We believe this principle applies beyond eco~’stemsnative fish fauna that is less resilient in the face of lava-and is particularly relevant at the intersection of ecologi-sire fish species Oleffe 1984; Minckley & Meffe 198/-3.cal, social, and economic systems. The high flow variation of unregulated rivers inhibits es-

Because much of our focus here is on loss of resil-tablishment of exotic fishes, and the last remainingience, we must explore that concept further. Resiliencestrongholds of southwestern native riverine fishes are allof a system has been defined in two reD" different waysin free-flowing rivers (Mi.nckley & Deacon 1991). Whenin the ecological literature; these differences in def’mi-flow variation is stabilized by dams and the process of vi-lion reflect which of two differer~t aspects of stability olent flooding is removed, the resulting lentic condi-are emphasized. The first definition, and the more tradi-tions favorable to many exotic species allow them totional one, concentrates on stability near art equilibriumflourish and eliminate native fishes that evolved withsteady-state, where resistance to disturbance and speedhigh flow variation. Stabilization of discharge variationof return to the equilibrium are used to measure resil-and the presence of invasive species results in unresil-ience (Pimm 1984 &; O’Neill et al. 1986; Tilman &ientand declining native fish faunas.Down_lag 1994). We call that equilibrium resilience. (3) Suppression of t’tre in f’tre-prone ecosystems is re-The second clef’tuition, and the one of greater relevancemarkably successful in reducing the short-term probabil-here. emphasizes conditions far from any equilibrium inity of fire in the national parks of the United States andwhich instabilities can flip a system into another regimein fire-prone suburban regions. But the consequence isof behavior~to another stability domain (Holling 1973, an accumulation of fuel over large areas that eventually1994). In this case the measurement of resilience is theproduces fires of an intensity, extent, and human costmagnitude of disturbance that can be absorbed or ac-never before encountered (Kilgore 1976; Christensen etcommodated before the system changes its structure byal. 1989). Fire suppression in systems that would fre-changing the variables and processes that control sys-quently experience low-intensiW fires results in the ~’s-tern behavior. We call that ecosystem resiliet~ce becauseterns becoming severely affected by the huge fires thatits significance becomes clearly apparent for large-scalefinally erupt; that is, the systems are not resilient to the

, focuses fires that occur with large fuel loads and fun-systemsover long periods. The first definition major may

Vt~lume I0, No, 2, April I~)tX~

E--0251 82E-025182

:s. damentally change state after the fire. As this and thehunting are enhanced by stocking or predator removal;prcvioua e.xamplc serve to demonstrate, suppression or forest ~res are suppressed for years; floods are mini-

d removat of a natural disturbance generally reduces sys-mized by levees. As a consequence, agencies responsi-e tern resilience, hie for management shift their attention from the origi-¯ ~" (4) Monocultural, energy-intensive farming practices hal social or economic purpose to an other~vise laudable,r ~ are the epitome of reduction of variation and loss of re-effort to increase efficiency and reduce costs--betterc-W silience. Plant species diversity in a natural forest con-and more efficient ways to kill insects, eliminate wolves,-’" verted to a monoculture may go from dozens or hun-rear hatcheD" fish, detect and extinguish fires, or control

dreds to one dominant, plus whatever weeds can escapeflows. Priorities thus shift from research and monitoringthe herbicides. Monocultures are notoriously suscepti-(why "waste" money studying and monitoring apparentble to the effects of drought, flo,~ding, insect or patho:success?) to internal agenQo goals of cost e~ciency andgen outbreaks, and market vagaries. They consequentlyinstitutional su~’ival. The second feature of the pathol-require large inputs of energy (fertilizers, pesticides, her-ogT thus emerges: growing isolation of agency person-

f bicides, irrigation) and often large societal subsidies innel from the systems being managed and insensitivity to~ the form of price supports, guaranteed loans, disaster re-public signals of concern--in short, growing instim-

fief, and surplus buyouts. These monocultur.es are funda-tional myopia and rigidiq,-.mentally unresi.Lient to natural or social perturbations. At the same time, economic activities exploiting the

’ (5) Natural, lateral flow variation (periodic floodplain resource benefit from success and expand in the shortinundation) thi-oughout much of the Mississippi Riverterm, and we witness greater capital investment in activ-drainage has been reduced by charmelization and con-ities such as agricultural production, pulp mills, subur-struction of a series of locks and levees to benefit agri- ban development, and fishing and hunting. That too isculture, shipping, and floodplain development. As a re-laudable within lirn_i, ts: it is the development of humansult, the inextricably combined riverine-social systemopportunity and enterprise. But the result is increasinghas little resilience during extreme storm events, as wit-dependency on continued success in controlling naturenessed in the massive flooding of 1993. Attempted corn-while, unknown to most, nature itself is losing resiliencemand-and-control of the river’s flows, allowing expan- and increasing the likelihood of unexpected events andsire floodplain development, resulted in an unresilienteventual system failure. With dependency comes denial,riverine-social system and unprecedented economicdemands by economic interests to keep and expand sub-destruction, sidles, and pressure for further command and control.

The same phenomenon applies equally well beyondThis third feature provides the f’mal element to the ulti-

Q natural resource management to many aspects 6fhumanmate pathologT of command-and-control resource andexistence. For example bureaucracies are an exercise inenvironmental management. The composite result is in-variance reduction through regulation and control; theircreasingly less resilient and more x-ulnerable ecosystems,purpose is elimination of extreme behavior through reg-more myopic and rigid institutions, and more dependentulation to promote conformity, to a specific set of stan-and selfish economic interests all attempting to maintaindards, which to some degree is certainly desirable in ashort-term success.civilized societT. But deeply entrenched bureaucraciesIf the response to this pathology, by other interests,are chaxacteristically unresilient to new challenges be-such as the environmental community, is exclusively de-cause the system discourages innovation or other behav-mand for tighter regulation and prohibition, then the pa-iDeal variance. This is cleariy"evidenced by merely pre- thologT is deepened, because this applies a command-senting a unique situation to a clerk who has beenand-control solution to a problem initiated by commandnarrowly trained in a highly standardized bureaucracyand control. The result is that lobby groups battle otherand watching the incredulous reply, or by the tTpicallylobby groups and generate the gridlocks and trainnegative response to and occasional punishment of awrecks that axe now regional issues--from salmon,government employee who offers an alternative per- owls, fishing, and logging in the Pacific Northwest, tospec~ive to the standard operating procedure, cod, poverty, and cultural st~rvival in Newfoundland, to

The pathology, of natural resource management in-sugar, urbanization, wildliIe, and water in the Evergladesvolves not just a contraction of the resilience of ecosys-(Gunderson et al. 1995).terns in response to human control: two other features Such problems, with a complex of causes, do not havemake for an ultimate pathology. One feature concernssimple solutions. We "know the goa!: more resilient eco-changes that occur in management agencies, and thesystems, more flexible agencies, more self-reliant indus-other involves changes in economic sectors, tries, and more "knowledgeable citizens. We also know

First, loss of ecosystem resilience is accompanied bythe ingr4dients of the solution, if not the specific wayschanges in the management agencies. The initial phaseto combine and use those ingredients. First. replaceof command and control is neaxly always quite success-nomic subsidies with incentives designed so that resto-

.,~ ful: insect pests are reduced by pesticide use; fishing andration and maintenance of ecosystem resilience is to the

V*~lurn¢ I0. Nt~. 2. April

5 3 2 Patbolo$9" of .\’atural Resource 3la*mgement Ho/linff ~ ,lleffe

berte~’~t o{ economic enterprise. An example is the con-variation provide. Irreversible or slowly reversible statesservation policies that reward farmers for restoration ofexist; once the system flips into .~uch a state, onlyhabitat and soils. Second. develop ways for agencies toplicit management intervention can restore its previousinnovate and learn, and allow them to do so. An exam-self-sustaining state, and even then buccess is not as-p/e is the application of actively adaptive environment sured (’ff’alker 1981). Conclttsion: Critical processesmanagement approaches, where policies become hy-fiotction at radically differettt rates and at spatialpotheses and management actions become the experi-scales cot,ering set’eral orders of mag,titttde, attd thesemerits to test those hypotheses (’Holling 197"8; Waiters rates" and scales" chtster arot¢nd a few dominant1986: Lee 1993; Gunderson et al. 1995). Third, engagequencies.people as active partners in the process of science and (2) Spatial attributes are no(uniform or scale-invariant.policy.. Examples are the various regional and continen-Rather, productivity, and textures are patchy and discon-tal monitoring schemes by which people monitortinuous at all scales, from the leaf to the individual to thechanges in nature~acid rain in the northeast, bird popu-vegetation patch to the landscape to the planet. Therelations along fl~’ays, water quality in bays and rivers,are several different ranges of scales, each with differentMonitoring of ecological change over time and space isattributes of patchiness and texture (Holling 1992). Co,t-critical to a better understanding of our managed re-ClltSion: ScaH, zg ttp from small to large cannot be asource sTstems and must be a central component of anyprocess of simple linear addition: non-ff, tear processesadaptive management scenario. Monitoring provides theorganize the shift from one range of scales to another.data for the management experiment and the basis forNot only do the large and slow control the small attddeciding the success or failure of the approach. Fourr2a,fast, the latter occasionally "reuolt~" to affect the former.develop local partnerships among broad constituencies (3) Ecosystems do not have single equilibria, withthat all stand to gain (or lose) together from good (orfunctions controlled to remain near them. Rather, multi-poor) resource management, ple equilibria, destabilizing forces far from equilibrih,

and absence of equilibria define functionally differentstable states, and movement bet-ween states maintains

The Behavior of Natural Ecosystems an overall structure and diversity. Conclusion: On theone hand, destabilizing forces are important in main-

Our suggestions would be more effective with a betterraining dit’ersity, resilience, and opportunit): On theunderstanding of ecog’stem behavior, structure, and. dy-other hand, stabilizing forces are important in main-namics at all spatial scales from the plant to the planetraining prodttctit’ity and biogeochemical o’cles, and,and at all temporal scales from seconds to millennia. Theet’en when these features are perturbed, they recot’ersurprises and crises crea{ed by the pathology" are not rapidly if the stability domain is not exceeded (e.g., re-onJ.v the consequence of incomplete "knowledge of howcot,cry of lakes from eutrophication or acidification;to control nature’s variabiliW.or improper controls beingSchindler 1990; Schindler et al. 1991).applied. They also include ignorance of the constructit’e (4) Policies and management that apply fLxed rules forrole that variation plays in maintaining the integrity of achieving constant yields independent of scale (e.g.,ecosystem function in the face of unexpected events,constant cat-r3"ing capacity of cattle or wildlife or con-Recently, a group of ecologists working with large-scalestan~ sustainable yield of fish, wood, or waled lead toterrestrial, fresh-water, and marine ecosystems devel-systems that gradually lose resilience~systems that sud-oped a ~’nthesis of their experience with natural, dis-denly break down in the face of disturbances that previ-turbed, and managed ecosystems (Holling et al. 1995).ously could be absorbed (Ho11Lng 1986). Conclttsion:They identified key features of ecosystem structure andsystems are mot,ing targets, with tnttltiffle potentialdynamics that explain why surprise and crisis are an in-futures that are uncertain and unpredictable. Tbere-evitable outcome of command-and-control approaches,fore management has to be flexible, adaptit,e, and ex-They concluded with the following lessons: perimental at scales compatible witf) the scales of criti-

(1) Ecological change is not continuous and gradual; cal ecosystem functions (Holling 1978; ~:’alters 1986;rather, it is episodic, with slow accumulation of naturalLee 1993; Gunderson et aL 1995).capital such as biomass or nutrients, punctuated by sud-den releases and reorganization of that capital as the re-sult of internal or external natural processes or of humanA (:loser Look at Resilienceimposed catastrophes. Rare events, such as hurricanesor the arrival of invading species, can unpredictablyThe features we have discussed are the consequences ofshape structure at critical times or at locations of in-the stabilizing properties of natural ecosystems. In thecreased vulnerabilitT; the effects of these rare events canecological literature these properties have been givenpersist for long periods. Therein lies one of the cus through debates on the meaning and realitT of thesources of new options that environmental diversit3." and resilience of ecosystems.

Ct )n~rr’,’:tt i~ )n Bi~lt)gy

E--0251 84E-025184

Earlier, we briefly defined resilience in two ways. ~ystcm resilience is inte~rative, synthetic, and focuses~’ne.~ vwv> aspects of a system’s stability, have very differ-on multiple scales. The consequences lead not on]y toent consequences for evaluating, understanding, andopposite views of system behavior but to oppositemanaging complexiW and change. Ecosystem resilience, views of s.vstem structure that have major consequencesour preferred definition, focuses on the interplay be-for policy. For example, there is a debate over whether~veen stabilizing and destabilizing properties, which areever)" species is important in ecosystem dynamics andat the heart of present issues of development and the en-function or whether only a smaller subset is involved invironment: global change, biodiversiD" loss, ecosystemserf-organization (Baskin 1994). On the one side is evi-restoration, and sustainable development. Nevertheless.dence from controlled experiments showing that declin-much of present ecological theory" uses the equilibriuming generalized diversity reduces productivity (Naeem etdeFmition of resilience, even though that definition rein-at. 1994), or that reducing numbers of grass species re-forces the pathology, of equilibrium-centered commandduces rates of recovew from drought (Tilman & Down-and control. That is because much of that theor)" draws hag 1994). In such examples, however, the physical limi-predominantly from traditions of deductive mathemati-rations of the experiments limit the conclusions to small-ca! theor)" (Pimm 1984) in wh_ich simplLfied, untouchedscale interactions (’plots ranged I-4 m on a side) overecological systems are imagined, or from traditions ofshort periods and to the set of structuring species thatengineering in which the motive is to design systemshappened to be selected at those scales. In contrast,with a single operating objective (Waide & Webster those who argue that a subset of species control dynam-1976; De AngeLis et. al. 1980; O’Neill et al. 1986), or ics and fi.mction draw their evidence from large-scalefrom small-scale quadrat experiments in nature (Tilmanmanipulations of whole eeoc’stems such as lakes (Schin-& Downing 1994) in wh.ich long-term, large-scale suc- dler 1990), from an understanding of process function atcessional or episodic transformations are not of con-different scales (I-lolling-1992; Levin 1992), from land-cern. That makes the mathematics more tractable, it ac-scape- and ecosystem-scale models (Clark et at. 1979;commodates the engineer’s goal to develop optimalCostanza et at. 1986; Waiters & Gunderson 1994), anddesigns, and it provides the ecologist with a rationale forfrom field measures of disturbed and managed ecosys-utilizing manageable, small sized, and short-term experi-terns (Hughes 1994). These observations address boreal,merits, all reasonable goals. But these traditional con-marine, freshwater, and savarma ecosystems and indi-cepts and techniques make the world appear more sire-care that functional diversiW is determined not by allpie. tractable, and manageabl~ than it really is. Theyspecies but by species involved in a set of structuringcaro" an implicit assumption that there is global stabil-processes ($chindler 1990; Holling et al. 1995). F_xam-iw--that there is only one equilibrium steady-state, or, ifples include the set of grass species and ungulate grazersother operating states e.-rdst, they should be avoided withthat maintain the productiviW and resilience of savarmassafeguards and regulatow controls. They transfer the(Watker et at. 1969) and the tree species and suite of 35command-and-control myopia of exploitive develop-species of insectivorous birds that mediate bud~¢orm out-ment to similarly myopic demands for environmentalbreak dynamics in the eastern boreal forest (HollLng 1988).regulations and prohibitions. Any ecosystem contains hundreds to thousands of spe-

Those who emphasize ecosystem resilience, on thecies interacting among themselves and their physicatother hand, come from traditions of applied mathemat-and chemical environment. But not all those interactionsics and applied resource ecology at the scale of eco~°s-have the same strength or the same direction. That is, al-terns, such as the dynamics and management of freshwa-though ever)’~hing might ultimately be connected to ev-ter systems (Fiering 1982), forests (Clark et al. 1979),er)-thing else ff the web of cormections is followed farfisheries (Waiters 1986), semiarid grasslands Ogratker etenough, the first-order interactions that structure theal. 1969), and interacting populations in nature (Dublinsystem increasingly seem to be confined to a subset ofet al. 1990; Sinclair et al. 1990). Because these studiesbiotic and abiotic variables whose interactions form theare rooted in inductive rather than deductive theory for-"template" (Southwood 1977) or the niches that allow amarion and in experience with the effects of large-scale- great diversiW of living things to,.in a sense, "go alongmanagement disturbances, the reality." of flips from onefor the ride" (Carpenter & Leavitt 1991; Cohen 1991;stable state to another cannot be avoided (Holling 1986).Holling 1992). Those species are affected by the ecosys-Indeed, management and resource exploitation cantern but do not, in turn, notably affect the ecosystem, atoverload waters with nutrients, turn forests into grass- least in ways that our relatively crude methods of mea-lands, trigger collapses in fisheries, and transform saran-surement can detect. At the extremes, therefore, speciesnas into shrub-dominated semideserts, can be regarded either as ~drivers" or as -passengers"

These two different views of resilience reflect t-wo dff- (Walker 1992), although this distinction needs to beferent traditions of ecological science: that of equilib-treated cautiously. The driver role of a species may be-rium resilience is experimental, analytical, and focusescome apparent on]3" ever5." now and then under particu-on small spatial scales and short durations; that of eco-lar conditions that trigger their key structuring function.

Volume It). No. 2. April

E--0251 85E-025185

334 PatboloD’ of S"aturcd Resource Managemem Holling & Meffe 1

This large-scale view of ecosystems highlights whereopold clearly anticipated the pathologT of natural re-the priorit-y for resource management, ecosystem res-source management as elaborated here.toratton, or biodiversity policy should lie. Ecological We ~ully recognize that the particular "rule" we pro-

ge is not incremental and local but sudden and ex-pose has far greater conceptual than prescriptive power.tensive. If change does occur, there may be fundamentalPrescriptions and cookbook approaches generally shouldtransformations from one ecosystem n.’pe to another--be avoided in conservation (Merle & Carroll 1994), if forfrom forest to grassland or grassland to a shrubby semi-no other reason than the systems with which we workdesert, for example (Walker et al. 1969; Holling 1973).are idiosyncratic and endlessly varied. No single, detailedThen control of structure will shift from one set of orga-prescription can be of much use for more than a singlenizing processes and variables to another. It is the diver-system. Furthermore, our rule is operationally vague.sity of overlapping influences within those controls that What is a "critical wpe" or a ~critical range" of variation?defines the resilience to those sudden shifts. That, obviously, is specific to a system and is often not

The fundamental points are that only a small set ofknown with any degree of assurance. Ehrerffeld (1992)sel£-organizing processes made up of biotic and physicalindicated that ~...it is extremely difficult to determine aelements are critical in forming the structure and overallnormal state for communities whose parameters are of-behavior of ecosystems, and that these establish se~ often in a condition of flux because of natural distur-relationships, each of which dominates over a definablebance." Schindler (198/’7) further explained that ~...werange of scales in space and time. Each set includes sev-usually do not know the normal range for any variable,eral species of plants or animals, each species havhagat least for aW time period greater than a few years."similar but overlapping influence to give functional re- Thus, our advice to "retain critical tTpes and ranges ofdundancy. It is that set, operating with abiotic pro- natural variation~ must remain for the present as a man-cesses, that generates and maintains ecosystem resil-agement goal to which to aspire, as a conceptual under-ience. It provides the focus for identifying the wpes and pkm’ling for management, rather than an operational dic-sources of variation that are critical for maintaining therum. In practice this translates to adopting a conservativeintegrity’of a natural ~’stem. approach to changing parameters of systems we under-

Thus, we suggest that an ecosystem-resilience per-stand poorly but that we wish to manage. It means thatspective better reflects the reality of large-scale pro-the default condition, unless clearly proven otherwise,cesses and dynamics and provides the most realisticshould be r.etention of the natural state rather than ma-foundation for addressing the challenging and complexnipulation of system components or d.vnamics. It arguesresource mangement issues of the day. It also providesfor humility when managing large systems (Stanleythe conceptual basis necessary to appreciate and under-1995). It shifts the burden of proof from managing bystand the paradoxes typically encountered in resourcesystem manipulation to managing by mLn_imal interven-management, as well as the pathology we describe here.tion, unless proven otherwise. It also argues strongly for

adaptive management rather than command-and-con-trol prescriptions, and development of consistent and

A Golden Rule for Natural Resource Managementdedicated monitoring of systems, both natural and man-~ aged. Only through long-term data collection can we be-

The various observations presented herein suggest agin to close the knowledge gap in understanding normal~Golden Rule" of natural resource management: A’aturalsystem behavior, parxicularly its variance.resource management should strive to retain critical How would this Golden Rule, at least in concept,.types and ranges of natural variation in ecosystems, modify resource management practices to take into ac-That is, management should facilitate existing processescount the pathology of natural resource management?and variabilities rather than changing or controlling We revisit our earlier examples and indicate howthem. By so doing, ecosystem resilience and the organiz-source management might be altered to adhere to theing processes and structures of ecosystems will be main-Golden Rule:rained, thus better serving not onl.v the natural functions (i) Genetic diversiD" of small populations should beand species diversity of those systems but also the long- retained and not further eroded by management prac-term (although not necessarily short-term) interests ofrices (Schonewald-Cox et al. 1983; Faik & Hotshagerhumanity. This is a more sophisticated way of stating 1990. This includes maintenance of natural gene flowAldo Leopold’s (1949) famous assertion that "A thing is in the wild (Meffe & Vrijenhoek 1988). reserves largeright when it tends to preserve the integrity, stabiliw, enough to maintain large breeding populations or meta-and beauty of the biotic community.. It is wrong when it populations of species of concern, and avoidance oftends otherwise." Because we know more today aboutpopulation crashes, bottlenecks, or inbreeding in cap-the dynamics of ecological communities than Leopoldrive breeding programs.did in the 1940s, we would replace "stability" with ~re- (2) In riverine systems with naturally high variation insilience;" otherwise, this remains sound advice, and Le-discharge, replace stabilization of flows via dams with

E--0251 86E-025186

v,.ater~hed restoration and protection. Begin to removeour penchant to control so many systems through corn-dams to restore the cridca! ecosystem process of dis-mand-and-control techniques, wills a few conspicuouscharge variation. Price water to accurately reflect its ect> exceptions the underlying problem of popttlation growthsystem value in order to stimulate conservation men-is often ignored, ironically, our attempts at commandaures, and remove flood-prone lands from development,and control are usually directed at complex, poorly un-Combine this with a bioregional perspective thatderstood, and nonlinear natural rather than atsystems,matches development practices to natural, regional eco-the fundamental source of the problem~human popula-logical constraints. Develop a combination of regionallion growth and consumption~where control is viable,and nation’al incentives and disincentives that wouldreasonable, and could be effective. A rapidly increasingeliminate ecologically disastrous development such ashuman population and increasing consumption is result-large desert cities that rely on water from far outside of ing in greater demands on and competition for dwin-the region and mining of fossil water with a temporallydling and increasingly damaged natural resources. Thelimited productive capacity, resource problems we encounter today can only multi-

(3) Eliminate policies of fire suppression in naturallyply as the human population grows, which means thatfire-prone ecosystems. Eliminate incentives that encour-the errors of command and control will be compot,nded,age rebuilding in such ecosystems a~er fire destruction,which will only lead to calls for more command andand develop incentives such as tax reductions to sitecontrol bv those who do not Amdamentatly underst:mdnew housing and other developments away from suchthe pathology outlined herein. This highlights the ur-areas, eventually to be designated as wilderness, gencT of quicMy changing our fundamental approaches

(4) Proceed from simple monocultures to more corn- to natural resource management and de~-eloping solu-plex agroecosystems with inte,wated pest managementlions and appropriate models of management behaviorand no-till methods (Carroll et al. 1990). Promote,while time and resources..still permit.though education and economic means, ecological corn- Command-and-contro(management can lead to short-plexity in agriculture, eliminating as much as possibleterm economic returns, but it also increases the vulnera-energetic and societal subsidies, allowing free ecosys-bikity of ecosystem.s to perturbations that other-wisetern services (e.g., diversiD° of predators on pests, soilcould be absorbed. Any move toward truly sustainableconservation through no-till methods) to support agri-human endeavors must incorporate this principle or itculture, cannot succeed. Our observations are also pertinent to

(5) Relocate communities out of floodplains of thethe present move toward ecosystem management in theMississippi River and other large riverine systems: ..use United States and elsewhere. If ecosystem managementthose areas as wildli£e refiages and corridors and as nam-is to be more than another buzzword, then there is noral buffers and recharge zones tbr agroecosystems (as issubstitute for understanding the structure and dynamicsnow being promoted in some parts of the Mississippiof natural ecosystems over spatial and temporal scalestloodplain). Provide disincentives for further floodplaincovering several orders of magnitude. TI-xe role of varia-development, lion in structuring ecosystems and maintaining their re-

(6) Examine bureaucracies to identify underlying rea-silience, and managing within the constraints of thatsons for their general intransigence and brittleness, andstructure and dynamics, is critical We must also modifypromote incentives" for alternative behaviors. Developour institutions and policies to recognize the pathology,incentives and rewards for innovation that place stream-described herein and to root out similar pathologies inlining, local solutions, and concern for customers andinstitutional and policy behaviors. To ignore this is tosustainability above adherence to a command structure,perpeturate the pathology of natural resource manage-

ment and place ecosystems and humanity, at great risk.

ConclusionsAcknowledoMnents

Rather than pursuing short-term gain through commandand control, effective natural resource management thatWe dedicate this paper to Mdo Leopold, who clearly an-promotes long-term system viability must be based on anticipated the ideas herein, and who was wOting aboutunderstanding of the key processes that structure andland pathologies as early as 1935. We thank Daviddove ecosystems, and on acceptance of both the naturalEhrerdeld and Garry Peterson for insightful commentsranges of ecos,vstem variation and the constraints of thaton eariier drafts. G. K. Merle was supported by contractvariation for long-term success and sustainability. This isDE-AC09--6SROO-819 bet-ween the U.S. Deparzmentespecially urgent when the growth of the human popu-of Energy and the UniversiW of Georgia, as well as a sab-lation and its consumption of resources is added to thebatical leave at the National Biological Service I.abot-a-picture, as it always must be (Merle et al. t993). Despite tow, Gainesville, Florida.

336 Patbolo,t9" of Natural Resource Managentent lfolling

Literature Cited Folke. B. O.Jan~m and C. S. Holling. editor:,. Biodivcrr, in." loss:10~t.’al .rod e,dontlmtc~ i:.,,ues. Cambridge Univc~it2:’ Prea,,. New York.

Allendorf. F. W . and R. F. Lear3.’. 1986. Heteroz3.’go.’,it2,." and fitness in Hu.,mhea, T. P. 1994. Cat~trt~phe~, pha>e .q’nt:t~, and large ~cale degrada-natural populations of animals. Pages 57--6 in M. E. Soul~, editor, tion of a Caribbean Coral Reel". Science 265:1547-1"551.Confer’alum btolog3.’: The ~cience of scarcitT and diversit2,,’. Sinauer Kilgore, B. 31. 19-6 Fire management in the natural parks: an over-A~:iates. Sunderland. Massachu.setts. view. Pages 45-5- m Prt)ceedings of TaR Timber~ fire ecolog.v con-

Ba~kin. "t’. 199-~. Ecologist~ dare to ask: How much does diver~iQ" mat- ference. Florida State L’ni\’e~itT Rc~earch Council, Tallaha.~.~ee.ter? Science 264:202-203. Lee. K. N. 1993. Comp:L,S and gyroscope: Integrating .,cience and poll.

Carpenter. 5. R.. and P. R. Leavitt. 1991. Temporal variation in pale- tics for the environment. Island Pres~. Covelo. California.olimnological record arising from atrophic cascade. Ecology 72: Leopold. A. 1949. A Sand CountT almanac and sketches here and277-285. there. Oxford UnivemiQ" Press. Ne~" York.

Carroll. C. R..J.H. Vandermeer. and P. M. Rosset, editors. 1990. Agro- Levin, S. A. 1992. The problem of pattern and scale in ecolobD’. Ecol-ecology. McGraw-Hill. New York. ogy 73:1943-1967.

Christensen, N. L., et al. 1989. Interpreting the Yellowstone fires of Meffe, G. K. 198+. Effects of abiotic disturbance on coexistence of19,’~. BioScience 39:6.-8-685. predator-prey fi.~h species. EcologD" 65:152"5- 1554.

Clark. W. C., D. D. Jones, and C. S. Holling. 1979. Lessons for ecologi- Meffe, G. K. 1986. C~m_,erx’ation genetics and the management ofcal policT design: A case study of ecosystem management. Ecologi- dangered fishes. Fisheries 11:14-23.cal Modeling 7:1-53. Meffe, G. K., and C. R. Carroll. i994. Principles of conservation biol-

Cohen. J. 1991. Tropic topology. Science 251:686-687. ogy. Sinauer ,-~sociates. Sunderland. 3la~-achusetts.Costanza. R., F. H. Sklar. and J. W. Day. Jr. 1986. Modeling spatial and Meffe. G. K.. and R. C. Vrigenho~k. 1988. Conse~’ation genetics in the

temporal succession in the Atchafalaya/Terrebonne marsl-destua- management of desert fishes. Conservation Biology 2:157-169.fine complex in South Louisiana. Pages 387-404 in D. A. Woli’e, ed- Meffe. G. K.. A. H. Ehrlich, and D. Ehrenfeld. 1995. Human popydationitor. Est-uarine variabilitT. Academic Press, New York. control: the missing agenda. Consep.’ation Biolog2," 7:1-3.

DeAngelis. D. L, W. M. Post, and C. C. Travis. 1980. Positive feedback Minckley, W. L., and J. E. Deacon. editors. 1991. Battle. against extinc-in natural systems. 5pringer-Verlag, New York. rion: Native fish managemem in the ,M’nerican West. UniversiW of

Dublin, H. To, A. R. E. Sinclair. andJ. McGlade. 1990. Elephants and fire Arizona Press. Tucson.as causes of multiple stable states in the Ser4ngeti-mara woodlands, Minck!ey, W. L., and G. K. Meffe. 198-. Differential selection by flood-Journal of.Maimal EcoloL~. 59:1147-1164. hag in stream-fish communities of the arid .Maaerican Southwest.

Ekrenfeld. D. 1991. The management of diversiw: a conservation pars- Pages 93- I 0-~ in D. C. Heins and ~. J. Matthews, editors: Evolution-dox. Pages 26-39 in F. H. Bor’man and S. R. Kellert, editors. Ecol- ary and communiW ecology." of North American stream fi~hes. Uni-ogy, economics, ethics: the broken circle. Yale Universiq,’ Press, versity of O -klahoma Press, Norman.New Haven, Connecticut. Naeem, S., L. J. Thompson. S. P. l~wler. J. H. Lawton. and R. M. Wood-

Elxrenfeld D. 1992. Ecosystem health and ecological theories. Pages Fro, 1994. Declirdng biodiversiD" can alter the performance of eco-135-143 in R. Costanza, B. G. Norton, and B. D. Hasketl, editors, systems. Nature 368:-34-737.EcosTstem health: New goals for environmental managemen~ Is- O’Neill. R. V., D. L. DeAngelis. J. B. kS’aide. T. F. H. ,Mien. 1986. A hier-land Press, Covelo, Cali.forrLia. arckical concept of ecosystems. Princeton Universit?- Press, Prince-

Falk, D. A., and K. E. Holsinger. 199l. Genetics and conservation of ton, New Jersey.rare plants. Oxford UrtiversiW Press, New York. Pickett, 5. T. A.. and P. 5. White, editors. 1985. The eco!og2:" of natura!

Fiering. M. B. I982, A screening model to quanti~," resilience. Water disturbance and patch dynamics. Academic Press. Orlando. Florida.Resources Research 18:27-32. Pimm, 5. L. 1984. The comple.’dW and stabilitT of ecosTstems. Nature

Gundet-son. L., C. S. HoRing, and S. Light, editors. 1995. Barriers and 307:321-326.bridges to the renewal of ecosystems and institutions. Columbia Schhadler, D. W. 198-. Detecting ecosystem responses to anthropo-Utliver~iW Press, New York. gertic stress. Canadian Journal of Fisheries and Aquatic Sciences

Holling. C. S. 1973. Resilience and stabilit2,," of ecology systems. Armual 44(suppl. I):6-25.Review of Ecolo~" and Systematics 4:1-’-2-3. Schhadler. D. W. 1990. Experimental perturbations of whole lakes as

Holling, C. S. 1978, Adaptive enviroru’nental assessment and manage- tests of hypotheses concerning eeoc’stem structure and function.merit. Jolm Wiley and Sons, New York. Oikos 57:25-4 I.

Holling, C. S. 1986, The resilience of terrestrial eeoc’stems, local sur- Schindler, D. W., et al. 1991. Freshwater acidification, reversibiliw andprise and g!obal change. Pages 292-317" in \V. C. Clark and R.E. recm’e~’: Comparisons of experimental and atmosphe,’-ically-acidi-Munn. editors. Sustainable development of the biosphere. Cam- fled lakes. Pages 193-226 in F. T. Last and R. Wading, editors.bridge UrtiversitT Press, Cambridge. Acidic deposition: its nature and impacts. Proceedings of the Royal

Holling. C. S. 1988, Temperate h}rest insect outbreaks, tropical defor- SocieW of Edinburgh, Edinburgh.estation and migrator" birds. Memoir~ of the Entomological Society Schonewald-Cox. C. M.. S. M. Chambers. B. MacBt3"de, and L. Thomas.of Canada 146:21-32. editors. 1983. Genetics and conse~’ation: A reference for managing

Holling. C. S. 1992. Cross-scale mot-pholo~-, geometry, and dynamics wild animal and plant populations. Benjamin!Cummings, Menloof ecosy.~tems. Ecological Monographs 62:447-502. Park. California.

Holling, C. S. 1994. New science and new investments for a sustain- Sinclair. A. R. E., P. D. Olsen. and T. D. Redhead. 1990. Can predatorsable biosphere. Pages 57-73 in A. M. Jansson, M. Hammer. C. Folke regulate small mammal populations? Oikos 59:382-392.and R. Costanza, editom. Investing in natural capital. Island Press, Southwood, T. R. E. 1977,. Habitat. the template for ecolo#cal strate-Washington, D.C. gies? Journal of .~oaimal Ecology 46:337-36"3.

Ho!ling, C. S. 1995. What barriers? What bridges? Pages 3-34 in L.H. Stanley. T. R., Jr. 1995. Ecosystem management and the arrogance ofGunderson, C. S. Holling. and S. S. Light, editors. Barriers and humanism. Con~e~’ation Btolo,~.-9:255-262.bridges to the renewal of ecosystems and institutions. Columbia Tilman, D.. andJ. A. Downing. 1994. Biodiver~iW and stability ha gras>UniversiD" Press, New York. lands. Nature 367:363-31~5.

Hotling. C. S., D. W. SclMndlcr, B. W. Walker, and J. Roughgarden. Waide, J. B.. and J. R. Webster. 1976. Engineering g’stems analysis: ap-199.5. BiodiversitT in the functioning tff ecosystems: Art ecological plicabilJ~" to eco~stem.~. Pages 329-372 in B. C. Patton. editor. S.vs-primer and synthesis. Pagc’~ 44-83 in C. Pertings, K. G. M’der, C. terns anal.v~,is and simulation in ecology. Acadtwnic Pre~, New York.

Volume IO. N~. 2. April

E--0251 88E-025188

Holling & Meffe Pathology’ of 3"atnral t~esource Management 337

%x’a\~cr. ~,. ~. ~.98,1.. Is succession a viable concept in Aft-ican ",a’,’atana bili~.." of ,oral-arid .~avanna grazing .~y.,,tcms, Journal of Ecology.’ 69:ecosystems? Page:~ 431-44- in D, C. West. H. H. Shugat-t. and D. B.Botkin. editor~. Forcer succession: concepts and :~pplJcat~on. \’~alt¢~. C. J. 1986. Adaptive m.’tnag~ment or" renewableSpt’inger-Vcrlag, .New York. .McGr-aw Hill, Ne,a," York.

Walker. B, H., 1992. Biological diver’si~ and ecological redundancy.\Va[ter.’,. C., and L. Gunderson 1994. A screening ot" water poliQ"

¯ " natives for" ecological re~toradon in the Everglades. Pages -5--767Con~e~’ation Biology." 6:18-23.in S. M. Davis and J. Ogden, editor’s. Everglades: the eco~.v~tcm and

Walker, B. H., D. Ludwig, C. S. Holling. and R, M. Peterman. 1969. Sta- it~ restoration. St. Lucie Press. Delr’a.v Beach. Florida.

~,’~|un|t: tO, No. 2. Aprtl ITS)6

E--0251 89E-025189