1_A Framework for Understanding (2)

1A Framework for UnderstandingChange

F. Stuart Chapin, III, Carl Folke, and Gary P. Kofinas

Introduction

The world is undergoing unprecedentedchanges in many of the factors that deter-mine both its fundamental properties andtheir influence on society. Throughout humanhistory, people have interacted with andshaped ecosystems for social and economicdevelopment (Turner et al. 1990, Redman1999, Jackson 2001, Diamond 2005). Duringthe last 50 years, however, human activitieshave changed ecosystems more rapidly andextensively than at any comparable period ofhuman history (Steffen et al. 2004, Foley et al.2005, MEA 2005d; Plate 1). Earths climate,for example, is now warmer than at any timein the last 500 (and probably the last 1,300)years (IPCC 2007a), in part because of atmo-spheric accumulation of carbon dioxide (CO2)released by the burning of fossil fuels (Fig. 1.1).Agricultural development largely accountsfor the accumulation of other trace gases that

F.S. Chapin, III ()Institute of Arctic Biology, University of AlaskaFairbanks, Fairbanks, AK 99775, USAe-mail: [email protected]

contribute to climate warming (see Chapter 12).As human population increases, in part dueto improved disease prevention, the increaseddemand for food and natural resources hasled to an expansion of agriculture, forestry,and other human activities, causing large-scaleland-cover change and loss of habitats andbiological diversity. About half the worldspopulation now lives in cities and depends onconnections with rural areas worldwide forfood, water, and waste processing (see Chap-ter 13; Plate 2). In addition, increased humanmobility is spreading plants, animals, diseases,industrial products, and cultural perspectivesmore rapidly than ever before. This increasein global mobility, coupled with increasedconnectivity through global markets and newforms of communication, links the worldseconomies and cultures, so decisions in oneplace often have international consequences.

This globalization of economy, culture, andecology is important because it modifies thelife-support system of the planet (Odum 1989),i.e., the capacity of the planet to meet the needsof all organisms, including people. The dramaticincrease in the extinction rate of species (100-to 1,000-fold in the last two centuries) indicatesthat global changes have been catastrophicfor many species, although some species,

3F.S. Chapin et al. (eds.), Principles of Ecosystem Stewardship,DOI 10.1007/978-0-387-73033-2 1, Springer Science+Business Media, LLC 2009

4 F.S. Chapin et al.

Globalpopulation

Peop

le (b

illion)

0

2

4

6

0

Water use

(10

00 km

)3 yr

1

2

4

6

1800

1750

1850

1900

1950

2000

Northern hemisphere temperature

Tem

p. a

nom

aly

( oC

)

0.5

0

0.5

1.0

Land conversion

% o

f lan

d ar

ea

0

25

50

1800

1750

1850

1900

1950

2000

% o

f fish

erie

s

0

20406080

100Fisheries exploited

Year

Year

Year

Year

1800

1750

1850

1900

1950

2000

1800

1750

1850

1900

1950

2000

Species extinctions

0

# (10

3 )

10

20

30

Human well-being ?

Ecosystemintegrity

? Challenges to ecosystemstewardship

Human drivers

Human impacts

Feedbacksto people

Ecosystemconsequences

Figure 1.1. Challenges to ecosystem stewardship.Changes in human population and resource con-sumption alter climate and land cover, which haveimportant ecosystem consequences such as speciesextinctions and overexploitation of fisheries. These

changes reduce ecosystem integrity and have region-ally variable effects on human well-being, whichfeeds back to further changes in human drivers. Panelinserts redrawn from Steffen et al. (2004).

especially invasive species and some diseaseorganisms, have benefited and expanded theirranges. Human society has both benefited andsuffered from global changes, with increased

food production, increased income and livingstandards (in parts of the world), improvedtreatment of many diseases, and longer lifeexpectancy being offset by deterioration in

1 A Framework for Understanding Change 5

ecosystem services, the benefits that societyreceives from ecosystems. More than halfof the ecosystem services on which societydepends for survival and a good life havebeen degradednot deliberately, but inadver-tently as people seek to meet their materialdesires and needs (MEA 2005d). Change cre-ates both challenges and opportunities. Peoplehave amply demonstrated their capacity to alterthe life-support system of the planet. In thisbook we argue that, with appropriate steward-ship, this human capacity can be mobilized tonot only repair but also enhance the capacity ofEarths life-support system to support societaldevelopment.

The unique feature of the changes describedabove is that they are directional. In otherwords, they show a persistent trend over time(Fig. 1.1). Many of these trends have becomemore pronounced since the mid-twentieth cen-tury and will probably continue or acceler-ate in the coming decades, even if societytakes concerted actions to reduce some ratesof change. This situation creates a dilemmain planning for the future because we cannotassume that the future world will behave aswe have known it in the past or that our pastexperience provides an adequate basis to plan

for the future. This issue is especially acute forsustainable management of natural resources.It is no longer possible to manage systems sothey will remain the same as in the recentpast, which has traditionally been the referencepoint for resource managers and conservation-ists. We must adopt a more flexible approach tomanaging resourcesmanagement to sustainthe functional properties of systems that areimportant to society under conditions wherethe system itself is constantly changing. Man-aging resources to foster resilienceto respondto and shape change in ways that both sustainand develop the same fundamental function,structure, identity, and feedbacksseems cru-cial to the future of humanity and the Earth Sys-tem. Resilience-based ecosystem stewardship isa fundamental shift from steady-state resourcemanagement, which attempted to reduce vari-ability and prevent change, rather than torespond to and shape change in ways that bene-fit society (Table 1.1). We emphasize resilience,a concept that embraces change as a basic fea-ture of the way the world works and devel-ops, and therefore is especially appropriate attimes when changes are a prominent featureof the system. We address ecosystems that pro-vide a suite of ecosystem services rather than a

Table 1.1. Contrasts between steady-state resource management, ecosystem management, and resilience-based ecosystem stewardship.

Resilience-based ecosystemSteady-state resource management Ecosystem management stewardship

Reference state: historic condition Historic condition Trajectory of changeManage for a single resource or

speciesManage for multiple ecosystem

servicesManage for fundamental

socialecological propertiesSingle equilibrium state whose

properties can be sustainedMultiple potential states Multiple potential states

Reduce variability Accept historical range of variability Foster variability and diversityPrevent natural disturbances Accept natural disturbances Foster disturbances that sustain

socialecological propertiesPeople use ecosystems People are part of the

socialecological systemPeople have responsibility to sustain

future optionsManagers define the primary use of

the managed systemMultiple stakeholders work with

managers to define goalsMultiple stakeholders work with

managers to define goalsMaximize sustained yield and

economic efficiencyManage for multiple uses despite

reduced efficiencyMaximize flexibility of future options

Management structure protectscurrent management goals

Management goals respond tochanging human values

Management responds to and shapeshuman values


single resource such as fish or trees. We focus onstewardship, which recognizes managers as anintegral component of the system that theymanage. Stewardship also implies a sense ofresponsibility for the state of the system ofwhich we are a part (Leopold 1949). The chal-lenge is to anticipate change and shape it forsustainability in a manner that does not leadto loss of future options (Folke et al. 2003).Ecosystem stewardship recognizes that soci-etys use of resources must be compatible withthe capacity of ecosystems to provide services,which, in turn, is constrained by the life-supportsystem of the planet (Fig. 1.2).

This chapter introduces a framework forunderstanding and managing resources in aworld where persistent directional changes arebecoming more pronounced. We first presenta framework for studying changeone thatintegrates the physical, ecological, and socialdimensions of change and their interactions. Wethen describe the general properties of systemsthat magnify or resist change. Finally we discussgeneral approaches to sustaining desirable sys-tem properties in a directionally changing worldand present a road map to the remaining chap-ters, which address these issues in greater depth.

Earths life support system

Human societies

EconomiesSu

stai

nab

ility

Figure 1.2. Socialecological sustainability requiresthat societys economy and other human activitiesnot exceed the capacity of ecosystems to provide ser-vices, which, in turn, is constrained by the planetslife-support system. Redrawn from Fischer et al.(2007).

An Integrated SocialEcologicalFramework

Linking Physical, Ecological,and Social Processes

Changes in the Earth System are highlyinterconnected. None of the changes men-tioned above is purely physical, ecological, orsocial. Therefore understanding current andfuture change requires a broad interdisciplinaryframework that draws on the concepts andapproaches of many natural and social sciences.We must understand the world, region, orcommunity as a socialecological system (alsotermed a coupled humanenvironment system)in which people depend on resources and ser-vices provided by ecosystems, and ecosystemdynamics are influenced, to varying degrees, byhuman activities (Berkes et al. 2003, Turner etal. 2003, Steffen et al. 2004). Although the rel-ative importance of social and ecological pro-cesses may vary from forests to farms to cities,the functioning of each of these systems, andof the larger regional system in which they areembedded, is strongly influenced by physical,ecological, economic, and cultural factors. Theyare, therefore, best viewed, not as ecological orsocial systems, but as socialecological systemsthat reflect the interactions of physical, ecologi-cal, and social processes (Westley et al. 2002).

Forests, for example, are sometimes man-aged as ecological systems in which the nitrogeninputs from acid rain or the economic influenceson timber demand are considered exogenousfactors (i.e., factors external to the system beingmanaged) and therefore are not incorporatedinto management planning. Production of lum-ber or paper, on the other hand, is often man-aged as an economic system that must balancethe supply and costs of timber inputs against thedemand for and profits from products withoutconsidering ecological influences on forest pro-duction. Finally, local planners make decisionsabout school budgets and the zoning for devel-opment and recreation, based on assumptionsabout regional water supply, which depends onforest cover, and economic projections, whichare influenced by the economic activity of forestindustries. The system and its components are


more vulnerable to unexpected changes (sur-prises) when each subsystem is managed in iso-lation. These surprises might include harvestrestrictions to protect an endangered species,development of inexpensive lumber supplies onanother continent, or expansion of recreationaldemand for forest use by nearby urban resi-dents. More informed decisions are likely toemerge from integrated approaches that rec-ognize the interdependencies of regional com-

ponents and account for uncertainty in futureconditions (Ludwig et al. 2001). Resource stew-ardship policies must therefore be ecologically,economically, and culturally viable, if they areto provide sustainable solutions.

In studying the response of socialecologicalsystems to directional change, we pay par-ticular attention to the processes that linkecological and social components (Fig. 1.3).The environment affects people through both

Ecological properties

Climate,regional

biota,etc.

Exogenous controls

Exogenous controls

Social properties

Slowvariables

Fast variables Fast variables

Tem

pora

l sca

le

Ecosystem services Human actorsEnvironmental impacts

Socialimpacts

Inst

itu

tio

nal

resp

on

ses

Social-ecologicalSystem

Regionalgovernance

systems,regional economy,

etc.

Soil resources,functional types,

disturbanceregime,

etc.

Soil nitrate,deer density,

fire event,etc.

Communityincome,

population density,

access to resources,

etc

Wealth andinfrastructure,

cultural tiesto the land,

etc.

Slowvariables

Spatial scale

Globe

Figure 1.3. Diagram of a socialecological system(the rectangle) that is affected by ecological (left-hand side) and social properties (right-hand side). Inboth subsystems there is a spectrum of controls thatoperate across a range of temporal and spatial scales.At the regional scale exogenous controls respond toglobal trends and affect slow variables at the scaleof management, which, in turn, influence fast vari-ables that change more quickly. When changes in fastvariables persist over long time periods and large

areas, these effects cumulatively propagate upwardto affect slow variables, regional controls, and even-tually the entire globe. Changes in both slow and fastvariables influence environmental impacts, ecosys-tem services, and social impacts, which, together,are the factors that directly affect the well-being ofhuman actors, who modify both ecological and socialsystems through a variety of institutions. Modifiedfrom Chapin et al. (2006a).


direct environmental events such as floodsand droughts and ecosystem services suchas food and water quality (see Chapter 2).Many economic, political, and cultural pro-cesses also shape human responses to the physi-cal and biological environment (see Chapter 3).Human actors (both individuals and groups)in turn affect their ecological environmentthrough a complex web of social processes (seeChapter 4). Together these linkages betweensocial and ecological processes structure thedynamics of socialecological systems (seeChapter 5).

The concept that society and nature dependon one another is not new. It was wellrecognized by ancient Greek philosophers(Boudouris and Kalimtzis 1999); economistsconcerned with the environmental constraintson human population growth (Malthus 1798);geographers and anthropologists seeking tounderstand global patterns of land use andculture (Rappaport 1967, Butzer 1980); andecologists and conservationists concerned withhuman impacts on the environment (Leopold1949, Carson 1962, Odum 1989). The complex-ity and importance of socialecological inter-actions has led many natural and social sci-ence disciplines to address components of theinteraction to both improve understanding andsolve problems. For example, resource man-

agement considers the actions that agencies orindividuals take to sustain natural resources,but typically pays less attention to the inter-actions among interest groups that influencehow management policies develop or how thepublic will respond to management. Similarly,environmental policy analysis addresses thepotential interactions of environmental policiesdeveloped by different organizations, but typ-ically pays less attention to potential social orecological thresholds (critical levels of driversor state variables that, when crossed, triggerabrupt changes or regime shifts) that determinethe long-term effectiveness of these policies.The breadth of approaches provides a wealth oftools for studying integrated socialecologicalsystems. Disciplinary differences in vocabu-lary, methodology, and standards of what con-stitutes academic rigor can, however, createbarriers to communication (Box 1.1; Wilson1998). The increasing recognition that humanactions are threatening Earths life-support sys-tem has recently generated a sense of urgency inaddressing socialecological systems in a moreintegrated fashion (Berkes et al. 2003, Clarkand Dickson 2003, MEA 2005d). This requiresa system perspective that integrates social andecological processes and is flexible enough toaccommodate the breadth of potential humanactions and responses.

Box 1.1. Challenges to Navigating SocialEcological Barriers and Bridges.

The heading of this box combines the titlesof two seminal books on integrated socialecological systems (Barriers and Bridgesand Navigating Social Ecological Systems;Gunderson et al. 1995, Berkes et al. 2003).These titles capture the essence of the chal-lenges in integrating natural and social sci-ences. In this book we adopt the follow-ing conventions in addressing two importantchallenges in this transdisciplinary integra-tion (i.e., integration that transcends tradi-tional disciplines to formulate problems innew ways).

The same word often means differentthings.

1. To a sociologist, adaptation means thebehavioral adjustment by individuals totheir environment. To an ecologist itmeans the genetic changes in a pop-ulation to adjust to their environment(in contrast to acclimation, which entailsphysiological or behavioral adjustment byindividuals). To an anthropologist adap-tation means the cultural adjustmentto environment, without specifying itsgenetic or behavioral basis. In this bookwe use adaptation in its most generalsense (adjustment to change in environ-ment).


2. To an engineer or ecologist describing sys-tems with a single equilibrium, resilienceis the time required for a system to returnto equilibrium after a perturbation. Tosomeone describing systems with mul-tiple stable states, resilience is capacityof the system to absorb a spectrum ofshocks or perturbations and still retainand further develop the same funda-mental structure, functioning, and feed-backs. We use resilience in the lattersense.

3. Natural scientists describe feedbacks asbeing positive or negative to denotewhether they are amplifying or sta-bilizing, respectively. These words areoften used in the social sciences (andin common usage) to mean good orbad. The terminology is especially con-fusing for socialecological systems,because negative feedbacks are oftensocially desirable (= good) and pos-itive feedbacks socially undesirable (=bad). We therefore avoid these termsand talk about amplifying or stabilizingfeedbacks.

4. Words that represent important conceptsin one discipline may be meaningless orviewed as jargon in another (e.g., post-modern, state factor). We define each tech-nical word the first time it is used and use

only those technical terms that are essentialto convey ideas effectively.

Approaches that are viewed as good sci-ence in one discipline may be viewed withskepticism in another.

1. Some natural scientists use systemsmodels to describe (either quantitativelyor qualitatively) the interactions amongcomponents of a system (such as a socialecological system). Some social scien-tists view this as an inappropriate tool tostudy systems with a strong human ele-ment because it seems too deterministicto describe human actions. We use com-plex adaptive systems as a framework tostudy socialecological systems because itenables us to study the integrated nature ofthe system but recognizes legacies of pastevents and the path dependence of humanagency as fundamental properties of themodel.

2. Some natural scientists rely largely onquantitative data as evidence to test ahypothesis, whereas some social scien-tists make extensive use of qualitativedescriptions of patterns that are lessamenable to quantification. We considerboth approaches essential to understandingthe complex dynamics of socialecologicalsystems.

A Systems Perspective

Systems theory provides a conceptual frame-work to understand the dynamics of integratedsystems. A socialecological system consistsof physical components, including soil, water,and rocks; organisms (plants, microbes, andanimalsincluding people); and the productsof human activities, such as food, money, credit,computers, buildings, and pollution. A socialecological system is like a box or a board game,with explicit boundaries and rules, enabling usto quantify the amount of materials (for exam-ple, carbon, people, or money) in the systemand the factors that influence their flows into,through, and out of the system.

Socialecological systems can be defined atmany scales, ranging from a single householdor community garden to the entire planet. Sys-tems are defined to include those componentsand interactions that a person most wants tounderstand. The size, shape, and boundariesof a socialecological system therefore dependentirely on the problem addressed and theobjectives of study. A watershed that includesall the land draining into a lake, for example,is an appropriate system for studying the con-trols over pollution of the lake. A farm, city,water-management district, state, or countrymight be a logical unit for studying the effectsof government policies. A community, nation,


or the globe might be an appropriate unitfor studying barter and commerce. A neigh-borhood, community, or multinational regionmight be a logical unit for studying culturalchange. Defining the most appropriate unit ofanalysis is challenging because key ecologicaland social processes often differ in scale andlogical boundaries (for example, watershedsand water-management districts; Ostrom 1990,Young 1994). Most socialecological systemsare open systems, in the sense that there areflows of materials, organisms, and informationinto and out of the system. We therefore cannotignore processes occurring outside our definedsystem of analysis, for example, the movementof food and wastes across city boundaries.

Socialecological processes are the intercon-nections among components of a system. Thesemay be primarily ecological (for example, plantproduction, decomposition, wildlife migration),socioeconomic (manufacturing, education,fostering of trust among social groups), or amix of ecological and social processes (plowing,hunting, polluting). The interactions amongmultiple processes govern the dynamics ofsocialecological systems. Two types of inter-actions among components (amplifying andstabilizing feedbacks) are especially impor-tant in defining the internal dynamics ofthe system because they lead to predictableoutcomes (DeAngelis and Post 1991, Chapinet al. 1996). Amplifying feedbacks (termedpositive feedbacks in the systems litera-ture) augment changes in process rates andtend to destabilize the system (Box 1.2).They occur when two interacting componentscause one another to change in the samedirection (both components increase or bothdecrease; Fig. 1.4). A disease epidemic occurs,for example, when a disease infects susceptiblehosts, which produce more disease organisms,which infect more hosts, etc., until some otherset of interactions constrains this spiral ofdisease increase. Overfishing can also lead toan amplifying feedback, when the decline infish stocks gives rise to price supports thatenable fishermen to maintain or increase fish-ing pressure despite smaller catches, leadingto a downward spiral of fish abundance. Otherexamples of amplifying feedbacks include

Cattle

Grass Shrubs

Soil moisture

A D

EB

C+ +

+

++

+

Resource uptakeCompetitionHerbivoryResource exploitationDisturbance cycle

ProcessNature of feedback

A or DA+D

BCE

Livelihoods

Fire

StabilizingStabilizingStabilizing

StabilizingAmplifying

Figure 1.4. Examples of linked amplifying and stabi-lizing feedbacks in socialecological systems. Arrowsshow whether one species, resource, or condition hasa positive or a negative effect on another. The feed-back between two species is stabilizing when thearrows have opposite sign (for example, species 1 hasa positive effect on species 2, but species 2 has a neg-ative effect on species 1). The feedback is amplify-ing, when both species affect one another in the samedirection (for example, more cattle providing moreprofit, which motivates people to raise more cattle;feedback loop C in the diagram).

population growth, erosion of cultural integrityin developing nations, and proliferation ofnuclear weapons.

Stabilizing feedbacks (termed negative feed-backs in the systems literature) tend toreduce fluctuations in process rates, although,if extreme, they can induce chaotic fluctuations.Stabilizing feedbacks occur when two interact-ing components cause one another to changein opposite directions (Fig. 1.4). For example,grazing by cattle reduces the biomass of foragegrasses, whereas the grass has a positive effecton cattle production. Any increase in densityof cattle reduces grass biomass, which then con-strains the food available to cattle, thereby sta-bilizing the sustainable densities of both grass


and cattle at intermediate levels. Other exam-ples of stabilizing feedbacks include prices ofgoods in a competitive market and nutrientsupply to plants in a forest. One of the keysto sustainability is to foster stabilizing feed-backs and constrain amplifying feedbacks that

might otherwise push the system toward somenew state. Conversely, if the current state issocially undesirable, for example, at an aban-doned mine site, carefully selected amplifyingfeedbacks may shift the system to a preferrednew state.

Box 1.2. Dynamics of Temporal Change

The stability and dynamics of a systemdepend on the balance of amplifying and sta-bilizing feedbacks and types and frequenciesof perturbations. The strength and natureof feedbacks largely govern the way a sys-tem responds to change. A system with-out strong feedbacks shows chaotic behav-ior in response to a random perturba-tion. Chaotic behavior is unpredictable anddepends entirely on the nature of the pertur-bation. The behavior of a ball on a surfaceprovides a useful analogy (Fig. 1.5; Hollingand Gunderson 2002, Folke et al. 2004). The

location of the ball represents the state of asystem as a function of some variable such aswater availability. In a chaotic system with-out feedbacks, the surface is flat, and we can-not predict changes in the state (i.e., location)of the system in response to a random per-turbation (Fig. 1.5a). This system structureis analogous to theories that important deci-sions can be described in terms of the poten-tial solutions and actors that happen to bepresent at key moments (garbage-can poli-tics; Cohen et al. 1972, Olsen 2001).

a. b.

c. d.

e. f.

Figure 1.5. The location of the ball represents thestate of a system in relationship to some ecological orsocial variable (e.g., water availability, as representedby the position along the horizontal axis). Changesin the state of the system in response to a perturba-tion depend on the nature of system feedbacks (illus-trated as the shape of the surface). The likelihood

that the system will change its state (location alongthe line) differs if there are (a) no feedbacks, (b)stabilizing feedbacks, (c) amplifying feedbacks, (d)alternative stable states, (de) changes in the internalfeedback structure (complex adaptive system), and(ef) response of a complex adaptive system to per-sistent directional changes in a control variable.

A system dominated by stabilizing feed-backs tends to be stable because the interac-tions occurring within the system minimizethe changes in the system in response toperturbations. Using our analogy, stabiliz-ing feedbacks create a bowl-like depressionin the surface so the ball tends to return tothe same location after a random perturba-

tion (Fig. 1.5b). The resilience of the sys-tem, in this cartoon, is the likelihood that itwill remain in the same state despite per-turbations. This analogy characterizes theperspective of a balanced view of nature, inwhich there is a carrying capacity (maximumquantity) of fish, game, or trees that the envi-ronment can support, allowing managers to


regulate harvest to achieve a maximum sus-tained yield. This view is often based on con-siderable depth of biological understandingbut is incomplete (Holling and Gunderson2002).

A system dominated by amplifying feed-backs tends to be unstable because the ini-tial change is amplified by interactions occur-ring within the system. Amplifying feedbackstend to push the system toward some newstate by making the depressions less deepor creating elevated areas on the surface(Fig. 1.5c). This analogy characterizes theview that small is beautiful and that any tech-nology is bad because it causes change. Thereare certainly many examples where technol-ogy has led to unfavorable outcomes, butthis worldview, like the others, is incomplete(Holling and Gunderson 2002).

Many systems can be characterized byalternative stable states, each of which isplausible in a given environment. Neighbor-hoods in US cities, for example, are likely tobe either residential or industrial but unlikelyto be an even mix of the two. In the sur-face analogy, alternative stable states rep-resent multiple depressions in the surface(Fig. 1.5d). A system is likely to return toits original state (=depression) after a smallperturbation, but a larger disturbance mightincrease the likelihood that it will shift tosome alternative state. In other words, thesystem exhibits a nonlinear response to theperturbation and shifts to a new state if somethreshold is exceeded. There may also bepathways of system development, such as thestages of forest succession, in which the inter-nal dynamics of the system cause it to movereadily from one state to another. Some of

these depressions may be deep and representirreversible traps. Others may be shallow, sothe system readily shifts from one state toanother through time. This worldview incor-porates components of all the previous per-spectives but is still incomplete.

The previous cartoons of nature implythat the stability landscape is static. How-ever, each transition influences the internaldynamics of a complex adaptive system andtherefore the probability of subsequent tran-sitions, so the shape of the surface is con-stantly changing (Fig. 1.5e). Reductions inAtlantic cod populations due to overfishing,for example, increased pressures for estab-lishment of aquaculture and charter fishingbusinesses, which then made it less likelythat industrial-scale cod fishing would returnto the North Atlantic. This analogy of astability landscape that is constantly evolv-ing suggests that precise predictions of thefuture state of the system are impossibleand focuses attention on understanding thedynamics of change as a basis for stewardship(Gunderson and Holling 2002).

Now imagine that rather than havinga random perturbation in some importantstate variable like water availability, thisparameter changes directionally. This ele-ment of directionality increases the likeli-hood that the system will change in a spe-cific direction after perturbation (Fig. 1.5f).The stronger and more persistent the direc-tional changes in exogenous control vari-ables, the more likely it is that new stateswill differ from those that we have knownin the past. This represents our concept ofsystem response to a directionally changingenvironment.

Issues of Scale: Exogenous, Slow,and Fast Variables

Changes in the state of a system depend onvariables that change slowly but strongly influ-ence internal dynamics. Socialecological sys-tems respond to a spectrum of controls thatoperate across a range of temporal and spatial

scales. These can be roughly grouped as exoge-nous controls, slow variables, and fast variables(Fig. 1.3). We describe these first for ecologicalsubsystems, then consider their social counter-parts.

Exogenous controls are factors such asregional climate or biota that strongly shapethe properties of continents and nations. They


remain relatively constant over long time peri-ods (e.g., a century) and across broad regionsand are not strongly influenced by short-term,small-scale dynamics of a single forest stand orlake. At the scale of an ecosystem or watershed,there are a few critical slow variables, i.e., vari-ables that strongly influence socialecologicalsystems but remain relatively constant overyears to decades despite interannual variationin weather, grazing, and other factors, becausethey are buffered by stabilizing feedbacks thatprevent rapid change (Chapin et al. 1996,Carpenter and Turner 2000). Soil organic mat-ter, for example, retains pulses of nutrients fromautumn leaf fall, crop residues, or windstorms;retains water and nutrients; and releases theseresources which are then absorbed by plants.The quantity of soil organic matter is bufferedby feedbacks related to plant growth and lit-ter production. Critical slow variables includepresence of particular functional types of plantsand animals (e.g., evergreen trees or herbivo-rous mammals); disturbance regime (propertiessuch as frequency, severity, and size that char-acterize typical disturbances); and the capacityof soils or sediments to supply water and nutri-ents. Slow variables in ecosystems, in turn, gov-ern fast variables at the same spatial scale (e.g.,deer or aphid density, individual fire events)that respond sensitively to daily, seasonal, andinterannual variation in weather and other fac-tors. When aggregated to regional or globalscales, changes that occur in ecosystems, forexample, those mediated by human activities,can modify the environment to such an extentthat even regional controls such as climateand regional biota that were once consideredconstant parameters are now directionallychanging at decade-to-century time scales(Foley et al. 2005). Regardless of the causes,persistent directional changes in broad regionalcontrols, such as climate and biodiversity,inevitably cause directional changes in crit-ical slow variables and therefore the struc-ture and dynamics of ecosystems, including thefast variables. The exogenous and slow vari-ables are critical to long-term sustainability,although most management and public atten-tion focus on fast variables, whose dynamics aremore visible.

Analogous to the ecological subsystem, thesocial subsystem can be viewed as composed ofexogenous controls, critical slow variables, andfast variables (Straussfogel 1997). These consistof vertically nested relationships, ranging fromglobal to local, and linked by cross-scale inter-actions (Ostrom 1999a, Young 2002b, Adgeret al. 2005). At the sub-global scale a predomi-nant history, culture, economy, and governancesystem often characterize broad regions ornation states such as Europe or sub-SaharanAfrica (Chase-Dunn 2000). These exogenoussocial controls tend to be less sensitive tointerannual variation in stock-market pricesand technological change than are the internaldynamics of local socialecological systems;the exogenous controls constrain local options.This asymmetry between regional and localcontrols occurs in part because of asymmetricpower relationships between national and localentities and in part because changes in a smalllocality must be very strong to substantiallymodify the dynamics of large regions. Regionalcontrols sometimes persist for a long time andchange primarily in response to changes thatare global in extent (e.g., globalization of mar-kets and finance institutions), but at other timeschange can occur quickly, as with the collapseof the Soviet Union in the 1990s or the global-ization of markets and information (Young etal. 2006). As in the biophysical system, a fewslow variables (e.g., wealth and infrastructure;property-and-use rights; and cultural ties to theland) are constrained by regional controls andinteract with one another to shape fast variableslike community income or population density.Both slow and fast social variables can havemajor effects on ecological processes (Costanzaand Folke 1996, Holling and Sanderson1996).

Systems differ in their sensitivity to differ-ent types of changes or the range of conditionsover which the change occurs. The !Kung Sanof the Kalahari Desert will be much more sen-sitive than people of a rainforest to a 10-cmincrease in annual rainfall because it repre-sents a doubling of rainfall rather than a 5%increase. Regions also differ in their sensi-tivity to introduction of new biota (sprucebark beetle, zebra mussel, or West Nile virus),


new economic pressures (development of aqua-culture, shifting of car manufacture to Asia,collapse of the stock market), or new culturalvalues. There are typically relatively few (oftenonly three to five) slow variables that are crit-ical in understanding the current dynamics ofa specific system (Carpenter et al. 2002), somanagement designed to reduce sensitivity todirectional changes in slow variables is not animpossible task. The identity of critical con-trol variables may change over time, however,requiring continual reassessment of our under-standing of the socialecological system. Thekey challenge, requiring collaborative researchby managers and natural and social scientists, isto identify the critical slow variables and theirlikely changes over time.

Incorporating Scale, Human Agency,and Uncertainty into Dynamic Systems

Cross-scale linkages are processes that con-nect the dynamics of a system to events occur-ring at other times or places (see Chapter 5).Changes in the human population of a region,for example, may be influenced by the wealthand labor needs of individual families (finescale), by national policies related to birth con-trol (focal scale), and by global inequalitiesin living standards that influence immigration(large scale). Events that occur at each scaletypically influence events at other scales. Theuniversal importance of cross-scale linkages insocialecological systems makes it important tostudy them at multiple temporal and spatialscales, because different insights and answersemerge at each scale (Berkes et al. 2003).

Legacies are past events that have largeeffects on subsequent dynamics of socialecological systems. This generates a pathdependence that links current dynamics topast events and lays the foundation for futurechanges (North 1990). Legacies include theimpact of plowing on soils of a regeneratingforest, the impact of the Depression in the1930s on economic decisions made by house-holds 40 years later, and the continuation ofsubsistence activities by indigenous people whomove from villages to cities. Because of path

dependence, the current dynamics of a systemalways depend on both current conditions andthe history of prior events. Consequently, dif-ferent trajectories can occur at different timesor places, even if the initial conditions werethe same. Path dependence is absolutely crit-ical to management, because it implies thathuman actions taken today, whether construc-tive or destructive, can influence the future stateof the system. Good management can make adifference!

Human agency (the capacity of humans tomake choices that affect the system) is oneof the most important sources of path depen-dence. Human decisions depend on both pastevents (legacy effects) and the plans that peoplemake for the future (reflexive behavior). Thestrong path dependence of socialecologicalsystems is typical of a general class of systemsknown as complex adaptive systems. These aresystems whose components interact in waysthat cause the system to adjust (i.e., adapt)in response to changes in conditions. This isnot black magic, but a consequence of inter-actions and feedbacks. Some of the most fre-quent failures in resource management occurbecause managers and resource users fail tounderstand the principles by which complexadaptive systems function. It is therefore impor-tant to understand their dynamics. Understand-ing these dynamics also provides insights intoways that managers can achieve desirable out-comes in a system that is responding simultane-ously to management actions and to persistentdirectional changes in exogenous controls.

Whenever system components with differ-ent properties interact spontaneously with oneanother, some components persist and oth-ers disappear (i.e., the system adapts; Levin1999; Box 1.2). In socialecological systems,for example, organisms compete or eat oneanother, causing some species to become morecommon and others to disappear. Similarly,purchasing or competitive relationships amongbusinesses cause some firms to persist and oth-ers to fail. Those components that interactthrough stabilizing feedbacks are most likelyto persist. This self-organization of compo-nents linked by stabilizing feedbacks occursspontaneously without any grand design. It


causes complex adaptive systems to be rela-tively stable (tend to maintain their proper-ties over time; DeAngelis and Post 1991, Levin1999). This self-regulation simplifies manage-ment challenges in many respects. A complexadaptive system like a forest, for example, tendsto take care of itself. This differs from adesigned structure like a car, whose compo-nents do not interact spontaneously and wheremaintenance must be continually applied just tokeep the car in the same condition (Levin 1999).

If conditions change enough to alter theinteractions among system components, the sys-tem adapts to the new conditions, hence theterm complex adaptive system (Levin 1998).The new balance of system components, in turn,alters the way in which the system respondsto perturbations (path dependence), creatingalternative stable states, each of which couldexist in a given environment (see Chapter 5).Given that exogenous variables are alwayschanging on all time scales, socialecologicalsystems are constantly adjusting and chang-ing. Consequently, it is virtually impossible tomanage a complex adaptive system to attainconstant performance, such as the constant pro-duction of a given timber species. System prop-erties are most likely to change if there aredirectional changes in exogenous controls. Thestronger and more persistent the directionalchanges in control variables, the more likely itis that a threshold will be exceeded, leading toa new state.

If a threshold is exceeded, and the systemchanges radically, new interactions and feed-backs assume greater importance, and somecomponents of the previous system may dis-appear. If a region shifts from a mining to atourist economy, for example, the communitymay become more concerned about fundingfor education and regulations that assure cleanwater. The regime shifts that occur as the sys-tem changes state also depend on the past stateof the system (path dependence). The presenceof a charismatic leader or nongovernmentalorganization (NGO), for example, can be crit-ical in determining whether large cattle ranchesare converted to conservation easements orsubdivisions when rising land values and taxesmake ranching unprofitable.

These simple generalizations about complexadaptive systems have profound implicationsfor resource stewardship: (1) Social and ecolog-ical components of a socialecological systemalways interact and cannot be managed in iso-lation from one another. (2) Changes in socialor ecological controls inevitably alter socialecological systems regardless of managementefforts to prevent change. (3) Historical eventsand human actions, including management, canstrongly influence the pathway of change. (4)The thresholds and nonlinear dynamics asso-ciated with path dependence, compounded bylack of information and human volition, con-strain our capacity to predict future change.Resource management and policy decisionsmust, therefore, always be made in an envi-ronment of uncertainty (Ludwig et al. 1993,Carpenter et al. 2006a).

Adaptive Cycles

The long-term stability of systems depends onchanges that occur during critical phases ofcycles of long-term change. All systems expe-rience disturbances such as fire, war, reces-sion, change in leadership philosophy, or clo-sure of manufacturing plants that cause largerapid changes in key system properties. Suchdisturbances have qualitatively different effectson socialecological systems than do short-term variability and gradual change. Adap-tive cycles provide a framework for describ-ing the role of disturbance in socialecologicalsystems (Holling 1986). They are cycles of sys-tem disruption, reorganization, and renewal. Inan adaptive cycle, a system can be disruptedby disturbance and either regenerate to a sim-ilar state or be transformed to some new state(Fig. 1.6a; Holling 1986, Walker et al. 2004).Adaptive cycles exhibit several recognizablephases. The cycle may be initiated by a distur-bance such as a stand-replacing wildfire thatcauses a rapid change in most properties ofthe system. Trees die, productivity decreases,runoff to streams increases, and public faithin fire management is shattered. This releasephase occurs in hours to days and radicallyreduces the structural complexity of the system.Other factors that might trigger release include


Connectednessweak strong

Co

nservation

Release

Renewal

Growth

Cap

ital

littl

elo

ts

a. Adaptive cycle

b. Panarchy

Intermediatesize and speed

Large andslow

Small andfast

Revo

ltRe

volt

Reme

mber

Figure 1.6. (a) Adaptive cycle and (b) cross-scalelinkages among adaptive cycles (panarchy) in asocialecological system. At any given scale, a systemoften goes through adaptive cycles of release (col-lapse), renewal (reorganization), growth, and conser-vation (steady state). These adaptive cycles of changecan occur at multiple levels of organizations, suchas individuals, communities, watersheds, and regions.These adaptive cycles interact forming a panarchy.For example, dynamics at larger scales (e.g., migra-tion dynamics or wealth) provide legacies, context,and constraints that shape patterns of renewal (sys-tem memory). Dynamics at finer scales (e.g., insectpopulation dynamics, household structure) may trig-ger release (revolt; e.g., insect outbreak). Redrawnfrom Holling and Gunderson (2002) and Hollinget al. (2002b).

threshold response to phosphorus loading ofa lake, collapse of the local or regional econ-omy, or a transition from traditional to inten-sive agriculture. Following release, there is a rel-atively brief (months to years) renewal phase.For example, after forest disturbance, seedlingsestablish and new policies for managing theforest may be adopted. Many things can hap-

pen during renewal: The species and policiesthat establish might be similar to those presentbefore the fire. It is also a time, however, whenthere is relatively little resistance to the estab-lishment of a new suite of species or poli-cies that emerge from the surrounding land-scape (see Fig. 2.4). These innovations may leadto a system that is quite different from theprefire system, i.e., a regime shift. After thisbrief window of opportunity for change, theforest goes through a growth phase over sev-eral decades, when environmental resources areincorporated into living organisms, and policiesbecome regularized. The nature of the regener-ating forest system is largely determined by thespecies and regulations that established duringrenewal. During the growth phase, the forest isrelatively insensitive to potential agents of dis-turbance. The high moisture content and lowbiomass of early successional trees, for exam-ple, make regenerating forests relatively non-flammable. Constant changes in the nature ofthe forest cause both managers and the publicto accept changing conditions and regulationsas a reasonable pattern. As the forest developsinto the steady-state conservation phase, theinteractions among components of the systembecome more specialized and complex. Lightand nutrients decline in availability, for exam-ple, leading to specialization among plants touse different light environments and differentfungal associations (mycorrhizae) to acquirenutrients. Similarly, in the policy realm, therelatively constant state of the forest leads tomanagement rules that are aimed at main-taining this constancy to provide predictablepatterns of recreation, hunting, and forest har-vest. Due to the increased interconnectednessamong these social and ecological variables, theforest becomes more vulnerable to any factorthat might disrupt this balance, including fire,drought, changes in management goals, or ashift in the local economy. Large changes in anyof these factors could trigger a new release inthe adaptive cycle.

Many human organizations also exhibitcyclic patterns of change. A business or NGO,for example, may be founded in response toa perceived opportunity for profit or socialreform. If successful, it grows amidst constant


adjustment to changes in personnel and activ-ities. Eventually it reaches a relatively sta-ble size, at which time the internal structureand operating procedures are regularized, mak-ing it less flexible to respond to changes inthe economic or social climate. When condi-tions change, the business or NGO may eitherenter a new period of adjustment (growth) ordecline (release), followed by potential renewalor collapse.

Perhaps the most surprising thing aboutadaptive cycles is that the sequence of phases(release, renewal, growth, and conservation)can be used as a way of thinking aboutmany types of social-ecological systems, includ-ing lakes, businesses, governments, nationaleconomies, and cultures, although the sequenceof phases is not always the same (Gundersonet al. 1995). Clearly the specific mechanismsunderlying cycles in these different systemsmust be quite different. One of the unsolvedchallenges in understanding socialecologicalsystems is to determine the general systemproperties and mechanisms that underlie theapparent similarities in cyclic patterns of dif-ferent types of systems and to clarify the dif-ferences. The specific mechanisms of adap-tive cycles in different types of systems aredescribed in many of the following chapters.

One of the most important managementlessons to emerge from studies of adaptivecycles is that socialecological systems aretypically most vulnerable (likely to change toa new state in response to a stress or distur-bance) and create their own vulnerabilitiesin the conservation phase, where they typi-cally spend most of their time. In this stage,managers frequently seek to reduce fluctu-ations in ecological processes and preventsmall disturbances in order to increase theefficiency of achieving management goals(e.g., the amount of timber to be harvested;number of houses that can be built; the budgetto pay salaries of personnel), increasing thelikelihood that even larger disturbances willoccur (Holling and Meffe 1996, Walker andSalt 2006). Flood control, for example, reducesflood frequency, which encourages infrastruc-ture development in floodplains where it isvulnerable to the large flood that will eventu-

ally occur. Prevention of small insect outbreaksincreases the likelihood of larger outbreaks.Management that encourages small-scaledisturbances and innovation during the conser-vation phase reduces the vulnerability to largerdisruptions (Holling et al. 1998, Carpenter andGunderson 2001, Holling et al. 2002a). Thespecific mechanisms that link stability in theconservation phase to triggers for disruptionare described in later chapters.

Release and crisis provide important oppor-tunities for change (Gunderson and Holling2002, Berkes et al. 2003; Fig. 1.5b). Someof these changes may be undesirable (inva-sion of an exotic species, dramatic shift inpolitical regimes that decrease social equity),whereas others may be desirable (implemen-tation of innovative policies that are moreresponsive to change). Recognition of thesechanging properties of a system through thelens of an adaptive cycle suggests that effec-tive long-term management and policy-makingmust be highly flexible and adaptive, looking forwindows of opportunity for constructive policyshifts.

Most socialecological systems are spatiallyheterogeneous and consist of mosaics of subsys-tems that are at different stages of their adap-tive cycles. Interactions and feedbacks amongthese adaptive cycles operating at differenttemporal and spatial scales account for theoverall dynamics of the system (termed panar-chy; Fig. 1.6b; Holling et al. 2002b). A forest, forexample, may consist of different-aged standsat different stages of regeneration from loggingor wildfire. In this case, the system as a wholemay be at steady state (a steady-state mosaic)even though individual stands are at differentstages in their cycles (Turner et al. 2001). In gen-eral, there are different benefits to be gainedat different phases of the cycle, so policies thatpermit or foster certain disturbances may beappropriate. Many families contain individualsat various stages of birth, maturation, and deathand benefit from the resulting diversity of skills,perspectives, and opportunities. Similarly, in ahealthy economy new firms may establish atthe same time that other less-efficient firms goout of business. Maintenance of natural cyclesof fire or insect outbreak produces wildlife


habitat in the early growth phase and preventsexcessive fuel accumulation that might other-wise trigger more catastrophic fires. Perhaps themost dangerous management strategy would beto prevent disturbance uniformly throughout aregion until all subunits reach a similar state ofmaturity, making it more likely that the entiresystem will change synchronously.

Sustainability in a DirectionallyChanging World

Conceptual Frameworkfor Sustainability Science

A systems perspective provides a logical frame-work for managing changes in socialecologicalsystems. To summarize briefly the previous sec-tions, the dynamic interactions of ecologicaland social processes that characterize most oftodays urgent problems necessitate a socialecological framework for planning and stew-ardship. Any sustainable solution to a resourceissue must be compatible with current socialand ecological conditions and their likely futurechanges. A resource policy that is not eco-logically, economically, and culturally sustain-able is unlikely to be successful. Sustainableresource stewardship must therefore be mul-tifaceted, recognizing the interactions amongecological, economic, and cultural variables andthe important roles that past history and futureevents play in determining outcomes in specificsituations. In addition, systems undergo cyclicchanges in their sensitivity to external perturba-tions, so management solutions that may havebeen successful at one time and place may ormay not work under other circumstances.

The complexity of these dynamics helpsframe the types of stewardship approaches thatare most likely to be successful. It is unlikelythat a rigid set of rules will lead to success-ful stewardship because key decisions mustfrequently be made under conditions of nov-elty and uncertainty. Moreover, under currentrapid rates of global environmental and socialchanges, the current environment for decision-making is increasingly different from past

conditions that may be familiar to managers orthe future conditions that must be accommo-dated. The more rapidly the world changes, theless likely that rigid management approacheswill be successful. By considering the systemproperties presented above, however, we candevelop resilience-based approaches that sub-stantially reduce the risk of undesirable socialecological outcomes and increase the likelihoodof making good use of unforeseen opportuni-ties. This requires managing for general sys-tem properties rather than for narrowly definedproduction goals. In this section, we present aframework for this approach that is describedin detail in subsequent chapters.

Sustaining the desirable features of our cur-rent world for future generations is an impor-tant societal goal. The challenge of doing soin the face of persistent directional trends inunderlying controls has led to an emerging sci-ence of sustainability (Clark andDickson 2003).Sustainability has been adopted as a centralgoal of many local, national, and internationalplanning efforts, but it is often unclear exactlywhat it is or how to achieve it. In this bookwe use the United Nations Environment Pro-gramme (UNEP) definition of sustainability:the use of the environment and resources tomeet the needs of the present without com-promising the ability of future generations tomeet their own needs (WCED 1987). Accord-ing to this definition, sustainability requiresthat people be able to meet their own needs,i.e., to sustain human well-being (that is, thebasic material needs for a good life, freedomand choice, good social relations, and personalsecurity) now and in the future (Dasgupta2001; see Chapter 3). Since sustainability andwell-being are value-based concepts, there areoften conflicting visions about what shouldbe sustained and how sustainability should beachieved. Thus the assessment of sustainabil-ity is as much a political as a scientific pro-cess and requires careful attention to whosevisions of sustainability are being addressed(Shindler and Cramer 1999). Nonetheless, anyvision of sustainability ultimately depends onthe life-support capacity of the environmentand the generation of ecosystem services(see Chapter 2).


Types and Substitutability of Capital

Sustainability requires that the productive baserequired to support well-being bemaintained orincreased over time. Well-being can be definedin economic terms as the present value offuture utility, i.e., the capacity of individualsor society to meet their own needs (Dasguptaand Maler 2000, Dasgupta 2001). Well-beingalso has important social and cultural dimen-sions (see Chapter 3), but the economic def-inition enables us to frame sustainability ina systems context. Sustainability requires thatthe total capital, or productive base (assets) ofthe system, be sustained. This capital has nat-ural, built (manufactured), human, and socialcomponents (Arrow et al. 2004). Natural cap-ital consists of both nonrenewable resources(e.g., oil reserves) and renewable ecosystemresources (e.g., plants, animals, and water) thatsupport the production of goods and serviceson which society depends. Built capital consistsof the physical means of production beyondthat which occurs in nature (e.g., tools, clothing,shelter, dams, and factories). Human capital isthe capacity of people to accomplish their goals;it can be increased through various forms oflearning. Together, these forms of capital con-stitute the inclusive wealth of the system, i.e.,the productive base (assets) available to soci-ety. Although not included in the formal defini-tion of inclusive wealth, social capital is anotherkey societal asset. It is the capacity of groupsof people to act collectively to solve problems(Coleman 1990). Components of each of theseforms of capital change over time. Naturalcapital, for example, can increase throughimproved management of ecosystems, includ-ing restoration or renewal of degraded ecosys-tems or establishment of networks of marine-protected areas; built capital through invest-ment in bridges or schools; human capitalthrough education and training; and social cap-ital through development of new partnershipsto solve problems. Increases in this productivebase constitute genuine investment. Investmentis the increase in the quantity of an asset timesits value. Sustainability requires that genuineinvestment be positive, i.e., that the productivebase (genuine wealth) not decline over time

(Arrow et al. 2004). This provides an objectivecriterion for assessing whether management issustainable.

To some extent, different forms of capital cansubstitute for one another, for example, naturalwetlands can serve water purification functionsthat might otherwise require the constructionof expensive water treatment facilities. Well-informed leadership may be able to implementmore cost-effective solutions to a given prob-lem (a substitution of human for economic cap-ital). However, there are limits to the extentto which different forms of capital can be sub-stituted (Folke et al. 1994). Water and food,for example, are essential for survival, and noother forms of capital can completely substi-tute for them (see Chapter 12). They there-fore have extremely high value to society whenthey become scarce. Declines in the trust thatsociety has in its leadership; sense of culturalidentity; the capacity of agricultural soils toretain sufficient water to support production;or the presence of species that pollinate criti-cal crops, for example, cannot be readily com-pensated by substituting other forms of capital.Losses of many forms of human, social, and nat-ural capital are especially problematic becauseof the impossibility or extremely high costsof providing appropriate substitutes (Folkeet al. 1994, Daily 1997). We therefore focusparticular attention on ways to sustain thesecomponents of capital, without which futuregenerations cannot meet their needs (Arrowet al. 2004).

Well-informed managers often have guide-lines for sustainably managing the componentsof inclusive wealth. For example, harvestingrates of renewable natural resources shouldnot exceed regeneration rates; waste emissionsshould not exceed the assimilative capacityof the environment; nonrenewable resourcesshould not be exploited at a rate that exceedsthe creation of renewable substitutes; edu-cation and training should provide opportu-nities for disadvantaged segments of society(Barbier 1987, Costanza and Daly 1992, Folkeet al. 1994).

The concept of maintaining positive genuineinvestment as a basis for sustainability is impor-tant because it recognizes that the capital assets


of socialecological systems inevitably changeover time and that people differ through timeand across space in the value that they placeon different forms of capital. If the productivebase of a system is sustained, future generationscan make their own choices about how best tomeet their needs. This defines criteria for decid-ing whether certain practices are sustainable ina changing world. There are substantial chal-lenges in measuring changes in various forms ofcapital, in terms of both their quantity and theirvalue to society (see Chapter 3). Nonetheless,the best current estimates suggest that manu-factured and human capital have increased inthe last 50 years in most countries but that nat-ural capital has declined as a result of deple-tion of renewable and nonrenewable resourcesand through pollution and loss of the functionalbenefits of biodiversity (Arrow et al. 2004). Insome countries, especially some of the poorerdeveloping nations, the loss of natural capitalis larger than increases in manufactured andhuman capital, indicating a clearly unsustain-able pathway of development (MEA 2005d).Some argue that there have also been substan-tial decreases in social capital as a result ofmodernization and urban life (Putnam 2000).

Managing Change in Ways that FosterSustainability

Managing for sustainability requires atten-tion to changes typical of complex adaptivesystems. In the previous section we definedcriteria to assess sustainability. These crite-ria are of little use if the system to whichthey are applied changes radically. Now wemust place sustainability in the context of the

directional changes in factors that govern theproperties of most socialecological systems.Three broad categories of outcome are possi-ble: (1) persistence of the fundamental prop-erties of the current system through adapta-tion, (2) transformation of the system to afundamentally different, potentially more desir-able state, or (3) passive changes (often degra-dation to a less-favorable state) of the sys-tem as a result of failure of the system toadapt or transform. Intermediate outcomesare also possible, if some components (e.g.,ecological subsystems, institutions, or socialunits) of the system persist, others transform,and others degrade (Turner et al. 2003). Sus-tainability implies the persistence of the fun-damental properties of the system or of activetransformation through deliberate substitutionof different forms of capital to meet societysneeds in new ways. In contrast, degradationimplies the loss of inclusive wealth and there-fore the potential to achieve sustainability.

How can we manage the dynamics of changeto improve the chances for persistence ortransformation? Four general approaches havebeen identified as ways to foster sustainabil-ity under conditions of directional change:(1) reduced vulnerability, (2) enhanced adap-tive capacity, (3) increased resilience, and(4) enhanced transformability. Each of theseapproaches emphasizes a different set ofprocesses by which sustainability is fostered(Table 1.2, Fig. 1.7). Vulnerability addresses thenature of stresses that cause change, the sensi-tivity of the system to these changes, and theadaptive capacity to adjust to change. Adap-tive capacity addresses the capacity of actorsor groups of actors to adjust so as to minimizethe negative impacts of changes. Resilience

Table 1.2. Assumptions of frameworks addressing long-term human well-being. Modified from Chapin et al.(2006a).

Assumed change in Nature of mechanisms Other approachesFramework exogenous controls emphasized often incorporated

Vulnerability Known System exposure and sensitivity todrivers; equity

Adaptive capacity, resilience

Adaptive capacity Known or unknown Learning and innovation NoneResilience Known or unknown Within-system feedbacks and

adaptive governanceAdaptive capacity,

transformabilityTransformability Directional Learn from crisis Adaptive capacity, resilience


Learning, coping,

innovating, adapting

Drivers System dynamics Outcomes

Externaldrivers

Imp

acts

Inte

ract

ion

s

Sensitivity

Natural &human capital

Biotic &social

interactions

Vulnerability

Adaptability

Resilience

Transformability

Persistence

Activelynavigated

transformation

Unintendedtransformation

Exp

osu

re

Figure 1.7. Conceptual framework linking humanadaptive capacity, vulnerability, resilience, and trans-formability. See text for definition of terms. Thesystem (e.g., household, community, nation, etc.)responds to a suite of interacting drivers (stresses,events, shocks) to produce one of three potentialoutcomes: (1) persistence of the existing systemthrough resilience; (2) actively navigated transfor-mation to a new, potentially more beneficial statethrough transformability; or (3) unintended trans-formation to a new state (often degraded) due tovulnerability and the failure to adapt or transform.These three outcomes are not mutually exclusive,because some components (e.g., ecological subsys-tems, institutions, or social units) of the system maypersist, others transform, and others degrade. Thesensitivity of the system to perturbations dependson its exposure (intensity, frequency, and duration)to each perturbation, the interactions among dis-tinct perturbations, and critical properties of the sys-tem. The system response to the resulting impacts

depends on its adaptive capacity (i.e., its capacity tolearn, cope, innovate, and adapt). Adaptive capac-ity, in turn, depends on the amount and diversityof social, economic, physical, and natural capitaland on the social networks, institutions, and entitle-ments that influence how this capital is distributedand used. System response also depends on effec-tiveness of cross-scale linkages to changes occurringat other temporal and spatial scales. Those compo-nents of the system characterized by strong stabi-lizing feedbacks and adaptive capacity are likely tobe resilient and persist. Alternatively, if the existingconditions are viewed as untenable, a high adaptivecapacity can contribute to actively navigated trans-formation, the capacity to change to a new, poten-tially more beneficial state of the system or sub-system. If adaptive capacity of some componentsis insufficient to cope with the impacts of stresses,they are vulnerable to unintended transformationto a new state that often reflects degradation inconditions.

incorporates adaptive capacity but also entailsadditional system-level attributes of socialecological systems that provide flexibility toadjust to change. Transformability addressesactive steps that might be taken to change thesystem to a different, potentially more desirablestate. Although anthropologists, ecologists, and

geographers developed these approaches some-what independently (Janssen et al. 2006), theyare becoming increasingly integrated (Berkeset al. 2003, Turner et al. 2003, Young et al.2006). This integration of ideas provides pol-icy makers and managers with an increasinglysophisticated and flexible tool kit to address


the challenges of sustainability in a directionallychanging world. We apply the term resilience-based ecosystem stewardship to this entiresuite of approaches to sustainability, becauseof its emphasis on sustaining functional proper-ties of socialecological systems over the longterm despite perturbation and change. Theseissues represent the core challenges of man-aging socialecological systems sustainably. Wenow briefly outline this suite of approaches.

Vulnerability

Vulnerability is the degree to which a sys-tem is likely to experience harm due to expo-sure to a specified hazard or stress (Turneret al. 2003, Adger 2006). Vulnerability theoryis rooted in socioeconomic studies of impactsof events (e.g., floods or wars) or stresses (e.g.,chronic food insecurity) on social systems buthas been broadened to address responses ofentire socialecological systems. Vulnerabilityanalysis deliberately addresses human valuessuch as equity and well-being. Vulnerability toa given stress can be reduced by (1) reducingexposure to the stress (mitigation); (2) reduc-ing sensitivity of the system to stress by sustain-ing natural capital and the components of well-being, especially for the disadvantaged; and/or(3) increasing adaptive capacity and resilience(see below) to cope with stress (Table 1.3;Turner et al. 2003). The incorporation ofadaptive capacity and resilience as integralcomponents of the vulnerability framework(Turner et al. 2003, Ford and Smit 2004) illus-trates the integration of different approaches tosustainability science.

Exposure to a stress can be reduced byminimizing its intensity, frequency, duration,or extent. Prevention of pollution or banningof toxic pesticides, for example, reduces thevulnerability of people who would otherwisebe exposed to these hazards. Mitigation(reduced exposure) is especially challengingwhen the stress is the cumulative effect of pro-cesses occurring at scales that are larger thanthe system being managed. Anthropogeniccontributions to climate warming through theburning of fossil fuels, for example, is globally

Table 1.3. Principal sustainability approaches andmechanisms. Adapted fromLevin (1999), Folke et al.(2003), Turner et al. (2003), Chapin et al. (2006a),Walker et al. (2006).

VulnerabilityReduce exposure to hazards or stressesReduce sensitivity to stressesSustain natural capitalMaintain components of well-beingPay particular attention to vulnerability of the

disadvantagedEnhance adaptive capacity and resilience (see below)

Adaptive capacityFoster biological, economic, and cultural diversityFoster social learningExperiment and innovate to test understandingSelect, communicate, and implement appropriate

solutions.ResilienceEnhance adaptive capacity (see above)Sustain legacies that provide seeds for renewalFoster a balance between stabilizing feedbacks and

creative renewalAdapt governance to changing conditions

TransformabilityEnhance diversity, adaptation, and resilienceIdentify potential future options and pathways to

get thereEnhance capacity to learn from crisisCreate and navigate thresholds for transformation

dispersed, so it cannot be reversed by actionstaken solely by those regions that experiencegreatest impacts of climatic change (McCarthyet al. 2005). Other globally or regionally dis-persed stresses include inadequate suppliesof clean water and uncertain availability ofnutritious food (Steffen et al. 2004, Kaspersonet al. 2005).

Sensitivity to a stress can be reduced in atleast three ways: (1) sustaining the slow eco-logical variables that determine natural capital;(2) maintaining key components of well-being;and (3) paying particular attention to theneeds of the disadvantaged segments of soci-ety, who are generally most vulnerable. Thepoor or disadvantaged, for example, are espe-cially vulnerable to food shortages or eco-nomic downturns, and people living in flood-plains or the wildlandurban interface areespecially vulnerable to flooding or wild-fire, respectively. An understanding of the


causes of differential vulnerability can lead tostrategies for targeted interventions to reduceoverall vulnerability of the socialecologicalsystem.

The causes of differential vulnerability areoften deeply rooted in the slow variables thatgovern the internal dynamics of society, suchas power relationships or distribution of land-use rights among segments of society (seeChapter 3). Conventional vulnerability anal-ysis assumes that the stresses are known orpredictable (i.e., either steady state or chang-ing in a predictable fashion). However, long-term reductions in vulnerability often requireattention to adaptive capacity and resilience atmultiple scales in addition to targeted effortsto reduce exposure and sensitivity to knownstresses.

Adaptive Capacity

Adaptive capacity (or adaptability) is thecapacity of actors, both individuals and groups,to respond to, create, and shape variabilityand change in the state of the system (Folkeet al. 2003, Walker et al. 2004, Adger et al.2005). Although the actors in socialecologicalsystems include all organisms, we focus par-ticularly on people in addressing the role ofadaptive capacity in socialecological change,because human actors base their actions notonly on their past experience but also on theircapacity to plan for the future (reflexive action).This contrasts with evolution, which shapes theproperties of organisms based entirely on theirgenetic responses to past events. Evolutionhas no forward-looking component. Adaptivecapacity depends on (1) biological, economic,and cultural diversity that provides the buildingblocks for adjusting to change; (2) the capacityof individuals and groups to learn how theirsystem works and how and why it is changing;(3) experimentation and innovation to testthat understanding; and (4) capacity to governeffectively by selecting, communicating,and implementing appropriate solutions(Table 1.3) We discuss the social and culturalbases of adaptive capacity in Chapters 3 and

4 and here focus on its relationship to systemproperties.

Sources of biological, economic, and cul-tural diversity provide the raw material onwhich adaptation can act (Elmqvist et al. 2003,Norberg et al. 2008). In this way it defines theoptions available for adaptation. People canaugment this range of options through learning,experimentation, and innovation. This capac-ity to create new options is strongly influencedby peoples access to built, natural, human,and social capital. Societies with little accessto capital are constrained in their capacity toadapt. People threatened with starvation, forexample, may degrade natural capital by over-grazing to meet their immediate food needs,thereby reducing their potential to cope withdrought or future food shortage. Rich coun-tries, on the other hand, have greater capac-ity to engineer solutions to cope with floods,droughts, and disease outbreaks. Natural cap-ital also contributes in important ways toadaptive capacity, although its role is oftenunrecognized until it has been degraded. Sys-tems that have experienced severe soil erosion,for example, have fewer options with whichto experiment and innovate during times ofdrought, and highly engineered systems thathave lost their capacity to store floodwatershave fewer options to adapt in response tofloods. The role of human capital in adaptivecapacity is especially important. It is muchmorethan formal education. It depends on an under-standing of how the system responds to change,which often comes from experience and localknowledge of past responses to extreme eventsor stresses. As the world changes, and new haz-ards and stresses emerge, this understandingmay be insufficient. Willingness to innovate andexperiment to test what has been learned and toexplore new approaches is crucial to adaptivecapacity.

Social capital through networking to select,communicate, and implement potential solu-tions is another key component of adaptivecapacity. Leadership, for example, is often crit-ical in building trust, making sense of complexsituations, managing conflict, linking actors, ini-tiating partnerships among groups, compilingand generating knowledge, mobilizing broad


support for change, and developing and com-municating visions for change (Folke et al. 2005;see Chapter 5). It takes more than leaders,however, for society to adapt to change. Socialnetworks are critical in effectively mobilizingresources at times of crisis (e.g., war or floods)and in providing a safety net for vulnerable seg-ments of society (see Chapters 4 and 5).

In the context of sustainability, adaptivecapacity represents the capacity of a socialecological system to make appropriate substi-tutions among forms of capital to maintain orenhance inclusive wealth. In this way the sys-tem retains the potential for future generationsto meet their needs.

Resilience

Resilience is the capacity of a socialecologicalsystem to absorb a spectrum of shocks orperturbations and to sustain and develop itsfundamental function, structure, identity, andfeedbacks through either recovery or reor-ganization in a new context (Holling 1973,Gunderson and Holling 2002, Walker et al.2004, Folke 2006). The unique contribution ofresilience theory is the recognition and identifi-cation of several possible system properties thatfoster renewal and reorganization after pertur-bations (Holling 1973). Resilience depends on(1) adaptive capacity (see above); (2) biophysi-cal and social legacies that contribute to diver-sity and provide proven pathways for rebuild-ing; (3) the capacity of people to plan for thelong term within the context of uncertainty andchange; (4) a balance between stabilizing feed-backs that buffer the system against stressesand disturbance and innovation that createsopportunities for change; and (5) the capacityto adjust governance structures to meet chang-ing needs (Holling and Gunderson 2002, Folkeet al. 2003, Walker et al. 2006; Table 1.3). Lossof resilience pushes a system closer to its lim-its. When resilience has been eroded, a distur-bance, like a disease, storm, or stock marketfluctuation, that previously shook and revital-ized the resilient system, might now push thefragile system over a threshold into an alter-native state (a regime shift) with a new trajec-

tory of change. Such system changes radicallyalter the flow of ecosystem services (Chapter 2)and associated livelihoods and well-being ofpeople and societies. Clearly, resilience is anessential feature of resource stewardship underconditions of uncertainty and change, so thisapproach to resource management is evenmoreimportant today than it has been in the past.

We have already discussed the role of sta-bilizing feedbacks in buffering systems fromchange and the role of adaptive capacity incoping with the impacts of those changes thatoccur. Sources of diversity, which is essen-tial for adaptation, are especially important inthe focal system and surrounding landscape attimes of crisis, i.e., during the renewal phaseof adaptive cycles, when there is less resis-tance to establishment of new entities. Fosteringsmall-scale variability and change logicallycontributes to resilience because it maintainswithin the system those components that arewell adapted to each phase of the adaptivecycleranging from the renewal to the con-servation phase. This reduces the likelihoodthat the inevitable disturbances will have catas-trophic effects. Conversely, preventing small-scale disturbances such as insect outbreaks orfires tends to eliminate disturbance-adaptedcomponents, thereby reducing the capacity ofthe system to cope with disturbance.

Biophysical and social legacies contribute toresilience through their contribution to diver-sity. Legacies provide species, conditions, andperspectives that may not be widely repre-sented in the current system. A buried seedpool or stems that resprout after fire, for exam-ple, give rise to a suite of early successionalspecies that are well adapted to postdisturbanceconditions but may be uncommon in the matureforest. Similarly, the stories and memories ofelders and the written history of past eventsoften provide insight into ways in which peoplecoped with past crises as well as ideas for futureoptions that might not otherwise be considered.This often occurs by drawing on social memory,the social legacies of knowing how to do thingsunder different circumstances. A key challengeis how to foster and maintain social memory attimes of gradual change, so it is available whena crisis occurs.


One of the key contributions of resiliencetheory to resource stewardship is the recogni-tion that complex adaptive systems are con-stantly changing in ways that cannot be fullypredicted or controlled, so decisions mustalways be made in an environment of uncer-tainty. Research and awareness of processesoccurring at a wide range of scales (e.g.,the dynamics of potential pest populations orbehavior of global markets) can reduce uncer-tainty (Adger et al. 2005, Berkes et al. 2005),but managing for flexibility to respond to unan-ticipated changes is essential. This contrastswith steady-state management approaches thatseek to reduce variability and change as a wayto facilitate efficient harvest of a given resourcesuch as fish or trees (Table 1.1).

Transformability and Regime Shifts

Transformability is the capacity to reconcep-tualize and create a fundamentally new sys-tem with different characteristics (Walker et al.2004; see Chapter 5). There will always be acreative tension between resilience (fixing thecurrent system) and transformation (seeking anew, potentially more desirable state) becauseactors in the system usually disagree aboutwhen to fix things and when to cut losses andmove to a new alternative structure (Walkeret al. 2004). Actively navigated transformationsrequire a paradigm shift that reconceptualizesthe nature of the system. During transforma-tion, people recognize (or hypothesize) a fun-damentally different set of critical slow vari-ables, internal feedbacks, and societal goals.Unintended transformations can also occur insituations where management efforts have pre-vented adjustment of the system to changingconditions, resulting in a fundamentally differ-ent system (often degraded) characterized bydifferent critical slow variables and feedbacks.The dividing line between persistence of a givensystem and transformation to a new state issometimes fuzzy. Total system collapse seldomoccurs (Turner andMcCandless 2004, Diamond2005). Nonetheless, actively navigated trans-formations of important components of a sys-tem are frequent (e.g., from an extractive to

a tourism-based economy). In general, diver-sity, adaptive capacity, and other componentsof resilience enhance transformability becausethey provide the seeds for a new beginning andthe adaptive capacity to take advantage of theseseeds.

Transformations are often triggered by crisis,so the capacity to plan for and recognize oppor-tunities associated with crisis contributes totransformability (Gunderson and Holling 2002,Berkes et al. 2003). Crisis is a time when soci-ety, by definition, agrees that some componentsof the present system are dysfunctional. Duringcrisis, society is more likely to consider novelalternatives. It is also a time when, if novel solu-tions are not seized, the system can becomeentrenched in the very policies that led to crisis,increasing the likelihood of unintended trans-formations. Climate-induced increases in wild-fires in the western USA, for example, threatenhomes that have been built in the wildlandurban interface. One potential transformationwould be policies that cease assuming publicresponsibility for private homes built in remotefire-prone areas and instead encourage moredense development of areas that could be pro-tected from fire and served by public trans-portation. This would reduce the need and costof wildfire suppression, increase the economicefficiency of public transportation, and reducethe use of fossil fuels. Alternatively, currentpolicies of fire suppression and dispersed res-idential development in forested lands mightpersist and magnify the risk of catastrophicloss of life and property as climate warmingincreases wildfire risk and fire suppression leadsto further fuel accumulation.

Sometimes systems exhibit abrupt transi-tions (regime shifts) to alternate states becauseof threshold responses to persistent changes inone or more slow variables. Continued phos-phorus inputs to clearwater lakes, for example,may lead to abrupt transitions to a turbid-wateralgal-dominated regime (Carpenter 2003). Sim-ilarly, persistent overgrazing can cause shrubencroachment and transition from grasslandto shrubland (Walker et al. 2004). Regimeshifts are large changes in ecosystems thatinclude both changes in stability domains ofa given system (e.g., clearwaterturbid-water


transitions; Fig. 1.7d) and system transforma-tions (Carpenter 2003, Groffman et al. 2006).

Challenges to Sustainability

The major challenges to sustainability varytemporally and regionally. Issues of sustainabil-ity are often prominent in developing nations,especially where substantial poverty, inade-quate educational opportunities, and insuffi-cient health care limit well-being (Kaspersonet al. 2005). These situations sometimes coin-cide with a high potential for environmentaldegradation, for example, soil erosion and con-tamination of water supplies, as people try tomeet their immediate survival needs under cir-cumstances of inadequate social and economicinfrastructure. Sustainable development seeksto improve well-being, while at the same timeprotecting the natural resources on which soci-ety depends (WCED 1987). In other words,it seeks directional changes in some under-lying controls, but not others. Questions areoften raised about whether sustainable devel-opment can indeed be achieved, given its twingoals of actively promoting economic devel-opment while sustaining natural capital. Thefeasibility of sustainable development dependson the multiple effects of development on sys-tem properties and the extent to which thesenew system properties can be sustained overthe long term. In other words, how does devel-opment influence the slow variables that gov-ern the properties of socialecological systemsand how can t

Documents

1_A Framework for Understanding (2)