Professional Documents
Culture Documents
PetersonCummingCarpenterinConBio2003 Scenarios
PetersonCummingCarpenterinConBio2003 Scenarios
net/publication/227600228
CITATIONS READS
1,414 4,746
3 authors:
Stephen R. Carpenter
University of Wisconsin–Madison
654 PUBLICATIONS 145,834 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
All content following this page was uploaded by Graeme S. Cumming on 26 February 2019.
Abstract: Conservation decisions about how, when, and where to act are typically based on our expectations
for the future. When the world is highly unpredictable and we are working from a limited range of expecta-
tions, however, our expectations will frequently be proved wrong. Scenario planning offers a framework for
developing more resilient conservation policies when faced with uncontrollable, irreducible uncertainty. A
scenario in this context is an account of a plausible future. Scenario planning consists of using a few con-
trasting scenarios to explore the uncertainty surrounding the future consequences of a decision. Ideally, sce-
narios should be constructed by a diverse group of people for a single, stated purpose. Scenario planning can
incorporate a variety of quantitative and qualitative information in the decision-making process. Often, con-
sideration of this diverse information in a systemic way leads to better decisions. Furthermore, the participa-
tion of a diverse group of people in a systemic process of collecting, discussing, and analyzing scenarios
builds shared understanding. The robustness provided by the consideration of multiple possible futures has
served several groups well; we present examples from business, government, and conservation planning that
illustrate the value of scenario planning. For conservation, major benefits of using scenario planning are (1)
increased understanding of key uncertainties, (2) incorporation of alternative perspectives into conservation
planning, and (3) greater resilience of decisions to surprise.
‡ email [email protected]
Paper submitted October 4, 2001; revised manuscript accepted June 20, 2002.
358
Conservation Biology
Volume 17, No. 2, April 2003
360 Scenarios as Tools for Conservation Peterson et al.
alternative models of how the world might work and are constructed to provide insight into drivers of change,
seek to develop policies that are robust to this uncer- reveal the implications of current trajectories, and illumi-
tainty. What distinguishes them is that management ex- nate options for action.
periments are built into the models. When experimental Unlike forecasts, scenarios stress irreducible uncertain-
manipulation is possible, traditional scientific approaches ties that are not controllable by the people making the
are effective at answering questions ( Medawar 1984 ). decisions. Although trends, expert predictions, visions of
Scenario planning is most useful when there is a high the future, and models are all parts of scenario-building
level of uncertainty about the system of interest and sys- exercises, they should not be mistaken for scenarios them-
tem manipulations are difficult or impossible. selves. Scenarios may encompass realistic projections of
current trends, qualitative predictions, and quantitative
models, but much of their value lies in incorporating both
The Possible: Scenarios
qualitative and quantitative understandings of the system
Scenarios were initially developed by Herbert Kahn in and in stimulating people to evaluate and reassess their
response to the difficulty of creating accurate forecasts beliefs about the system (Greeuw et al. 2000). Useful sce-
( Kahn & Wiener 1967; May 1996). Kahn worked at the narios incorporate imaginative speculation and a wide
RAND Corporation, an independent research institute range of possibilities; those based only on what we cur-
with close ties to the U.S. military. He produced fore- rently know about the system have limited power because
casts based on several constructed scenarios of the fu- they do not help scenario users plan for the unpredictable.
ture that differed in a few key assumptions ( Kahn &
Wiener 1967). This approach to scenario planning was
later elaborated upon at SRI International (May 1996), a Scenario Planning
U.S. research institute, and at Shell Oil ( Wack 1985a,
1985b, Schwartz 1991; Van der Heijden 1996). There are many different approaches to scenario plan-
A scenario describes a possible situation, but the term ning, but most of those presented here are derived from
has been used in a variety of ways. One common use of the qualitative approach of Shell/SRI International and
scenario refers to the expected continuation of the cur- share many features ( Wack 1985a; Millett 1988; Schoe-
rent situation—for example in the statement that “under maker 1991; Bunn & Salo 1993; Schoemaker 1995; May
the current scenario, we anticipate that the species will 1996; van der Heijden 1996; Bossel 1998; Ringland 1998;
become extinct in the next 10 years.” Another common Wollenberg et al. 2000). These approaches differ in em-
use of scenario derives from systems models. The re- phasis because of variation in the goals of those who have
sults of integrated computer simulation models depend created them.
on assumptions about the extrinsic drivers, parameters, Our approach to scenario planning is similar to these
and structure of the model. Variations in the assump- approaches but is also influenced by experience with
tions used to create such models are often described as adaptive environmental assessment and management
scenarios ( Meadows et al. 1992 ). For example, the dif- ( Holling 1978). We see scenario planning as consisting
ferences between the climatic scenarios of the Intergov- of six interacting stages that a small group of research
ernmental Panel on Climate Change ( IPCC ) are deter- scientists, managers, policymakers, and other stakehold-
mined by differences in assumptions about demography ers explore through a series of workshops. It begins with
and social, economic, technical, and environmental de- identification of a central issue or problem. This problem
velopment ( Nakicenovic & Swart 2000 ). Scenarios in is then used as a focusing device for assessment of the
this sense have been used in conservation for investigat- system; assessment is combined with the key problem
ing the outcomes of different ecological management re- to identify key alternatives. Alternatives are then devel-
gimes (Klenner et al. 2000), to consider questions such as oped into actual scenarios. Scenarios are tested in a vari-
how malaria might spread in a warmer world (Rogers & ety of ways before they are used to screen policy. Al-
Randolph 2000), or to select conservation sites based on though this overview presents scenario planning as a linear
projected changes in human settlement ( White et al. process, it is often more iterative: system assessment leads
1997 ). Although these approaches describe future situa- to redefinition of the central question, and testing can re-
tions as scenarios, scenario planning invests much more veal blind spots that require more assessment.
effort in constructing integrated and provocative alter-
native futures.
(1) Identification of a Focal Issue
We define a scenario as a structured account of a pos-
sible future. Scenarios describe futures that could be Scenario planning aims to enhance our ability to respond
rather than futures that will be ( van der Heijden 1996; quickly and effectively to a wide range of futures, avoid-
Raskin et al. 1998). In essence, scenarios are alternative, ing potential traps and benefiting from potential oppor-
dynamic stories that capture key ingredients of our un- tunities. Because the real world is highly complex and
certainty about the future of a study system. Scenarios the future contains an infinite number of possibilities,
Conservation Biology
Volume 17, No. 2, April 2003
Peterson et al. Scenarios as Tools for Conservation 361
scenario planning must be focused to be effective. Ex- to identify alternative ways that the system could evolve.
amining the future in light of a specific question ( e.g., These alternatives should be both plausible and relevant
how ecosystem services could change in Australian range- to the original question. Plausible alternatives should
lands over the next two decades) separates the relevant each represent a path shaped by the interaction of exist-
aspects of the future that are knowable from those that ing dynamics and possible future events.
are unknowable. Consideration of policy response op- An analysis of the role of uncertainties in the assess-
tions (e.g., what can be controlled within a region) dis- ment can be used to organize alternatives. Although
tinguishes aspects of the future that are controllable or some uncertainties arise from the unknown future be-
that can be influenced by the scenario maker from those havior of actors within the system, other uncertainties
that that are beyond control. are based on the unknown dynamics of nature and un-
The focal issue should emerge from negotiation known changes in system drivers. For example, the fu-
among participants in the planning process. For exam- ture of housing development within a watershed may be
ple, the initial question might be how to maintain a func- unknown but can be influenced by local policies. How-
tional longleaf pine forest in a protected area for the next ever, the future rainfall of that watershed is uncertain
20 years, or how to make an existing reserve network and cannot be influenced by local policies. Scenario
more robust. Frequently, problems are defined in a nar- planning usually focuses on uncontrollable uncertainties
row way that represents the interests of a dominant group because the attempts of stakeholders to address control-
of stakeholders. Involving a diverse group of stakeholders lable problems should be included within a scenario’s
can help to define an issue broadly ( Ney & Thompson dynamics.
2000). A more precise definition can be crafted after the A set of alternatives can be defined by choosing two
system of interest is further assessed. The selection of or three uncertain or uncontrollable driving forces. For
a focal issue or question implicitly defines a system of example, if two important uncertainties in a region are
interest. population growth and settlement pattern, they could
be used to define alternatives such as increased migra-
tion to cities, increased migration to rural areas, and gen-
(2) Assessment
eral population decline. The uncertainties chosen to de-
The focal issue should be used to organize an assessment fine the alternatives should have differences that are
of the people, institutions, ecosystems, and linkages among directly related to the defining question or issue. They
them that define a system. It is also important to identify should imaginatively but plausibly push the boundaries
external changes, either ecological or social, that influence of commonplace assumptions about the future. This set
the dynamics of a system. These could include processes of alternatives provides a framework around which sce-
as diverse as migration, the spread of invasive species, narios can be constructed.
climate change, or changes in tax laws. A key aspect of
assessment is identifying which uncertainties will have a (4) Building Scenarios
large impact on the focal issue. For example, climatic
change may be a key uncertainty in the planning of a A set of scenarios is built based on the understanding ac-
park system. A subset of these uncertainties will be used cumulated during the assessment process to flesh out a
to help define scenarios. few of the alternatives defined by the key uncertainties.
One of the most valuable uses of the assessment pro- A set of scenarios should usefully expand and challenge
cesses is to confront the initial question or focal issue current thinking about the system. The appropriate
with the complexity of the world. Does the question ad- number of scenarios is generally considered to be three
dress the underlying issue or only its symptoms? For ex- or four; two scenarios usually do not expand thinking
ample, managers could decide that determining the opti- enough, whereas more than four may confuse users and
mal configuration of parks is too narrow a question and limit their ability to explore uncertainty ( Wack 1985b;
focus instead on how their park management could pro- Schwartz 1991; van der Heijden 1996).
duce a regional landscape that supports biodiversity. An Scenarios convert the key alternatives into dynamic
assessment should consider the different world views stories by adding a credible series of external forces and
of the key actors because these views often significantly actors’ responses. Scenarios should become brief narra-
shape a group’s understanding of a system’s dynamics tives that link historical and present events with hypo-
(Ney & Thompson 2000). thetical future events. Within these storylines the inter-
nal assumptions of the scenario and the differences
between stories must be clearly visible. To be plausible,
(3) Identification of Alternatives
each scenario should be clearly anchored in the past,
Following determination of what is known and un- with the future emerging from the past and present in a
known about the internal and external forces that shape seamless way. There are a variety of approaches to as-
the system’s dynamics, the goal of scenario planners is sembling scenario stories, including the use of tables or
Conservation Biology
Volume 17, No. 2, April 2003
362 Scenarios as Tools for Conservation Peterson et al.
cubes to display critical uncertainties (Schoemaker 1991) use of the scenarios is to assess how existing polices
and a search for archetypal plots, such as “winners and would fare in different scenarios. Such an approach can
losers” or “victims become heroes” ( Schwartz 1991 ). identify weak policies and those that are more robust to
Whatever approach is taken, each story should track the uncertainty about the future. A slightly more sophisti-
key indicator variables that have been selected based on cated approach is to identify the properties of policies
the initial question ( e.g., percent of intact old-growth or actions that perform well in all the scenarios. For
forest, soil erosion ). To help communicate and discuss example, a park planning effort may use policies that
scenarios, it is useful to give each scenario a name that examine different development trajectories. Scenarios
evokes its main features. Successful scenarios are vivid could be used to help identify land-management strate-
and different, can be told easily, and plausibly capture gies that produce parks that are resilient in response to
future transformation. change. In this process, it is important to identify traps
and opportunities and aspects of the current situation that
could influence these scenario features. This process may
(5) Testing Scenarios
suggest novel policies, areas for research, and issues to
Once the scenarios have been developed, they should monitor.
be tested for consistency. The dynamics of scenarios Scenario planning that involves stakeholders can pro-
must be plausible; neither nature nor the actors involved vide a forum for policy creation and evaluation. Stake-
in the scenario should behave in implausible ways. In- holders who become involved in the scenario-planning
consistencies will quickly emerge as major obstacles to process are likely to find that some scenarios represent a
their usefulness for developing policy. Consistency can future that they would like to inhabit, whereas others
be tested by quantification, against stakeholder behav- are highly undesirable. This process of reflection can
ior, through expert opinion, and against other scenarios. stimulate people to think more broadly about the future
Quantifying the changes expected in the system al- and the forces that are creating it and to realize how
lows the plausibility of the scenarios to be tested in a their own actions can move the system toward a particu-
number of ways. Simulation models can be used to test lar kind of future. In this way, scenario planning allows
the dynamics of a scenario, but models are not the pri- people to step away from entrenched positions and iden-
mary tool for scenario development. The most impor- tify positive futures that they can work at creating. Policy
tant part of a scenario’s plausibility is likely to be the screening often identifies new questions, new variables,
behavior of actors. Many modeling exercises have been and new types of unknowns. These concerns can stimu-
criticized, correctly, by economists for assuming that peo- late either another iteration of the scenario-planning
ple are far more passive than they actually are ( Westley et process or another form of action.
al. 2002). The changes expected to occur within a sce- A successful scenario-planning effort should enhance
nario should be examined from the perspective of each the ability of people to cope with and take advantage of
key actor within it. If an actor is likely to respond to the future change. Decisions can be made, policies changed,
scenario in some way that is not part of the scenario, and management plans implemented to steer the system
that assumed inaction or action decreases the plausibil- toward a more desirable future. New research or moni-
ity of the scenario. Varying degrees of sophistication can toring activities may be initiated to increase understand-
be used in testing the plausibility of actor behavior. A ing of key uncertainties, and they may stimulate the for-
rough test would involve the scenario team adopting the mation of new coalitions of stakeholder groups.
world view of all actors to explore their behavior in a
scenario. A stronger test would be to involve various
groups of actors and stakeholders in the scenario pro-
cess to ensure that actor behavior is plausible ( Ney & Examples of Scenario Exercises
Thompson 2000). For example, whereas a scenario may
describe people moving away in response to an ecologi- Scenarios have been used in a variety of ways, but they
cal change, discussions with local residents may reveal have not been used much in conservation. Applications
that such emigration is actually quite unlikely. Testing of scenario planning can be organized by their use of
usually reveals problems with scenarios. Consequently, qualitative or quantitative methods and their approach
scenarios need to go through several iterations of refine- toward uncertainty. Most scenarios incorporate both
ment and testing before they can be used to evaluate qualitative and quantitative details, and the relative mix
policies. of these two aspects distinguishes different scenario ex-
ercises. Some scenarios are intended to facilitate the
management of uncertainty, whereas others are used to
(6) Policy Screening
discover it. We present three examples of how scenarios
Once a set of scenarios has been developed, it can be have been used to approach problems that were beyond
used to test, analyze, and create policies. The simplest the reach of traditional predictive methods.
Conservation Biology
Volume 17, No. 2, April 2003
Peterson et al. Scenarios as Tools for Conservation 363
Conservation Biology
Volume 17, No. 2, April 2003
364 Scenarios as Tools for Conservation Peterson et al.
investments of recent migrants to northern Wisconsin shareholder value) to develop strategy for an uncertain
and policy changes led to the establishment of a variety future. In contrast, Monte Fleur brought a group of dis-
of satellite offices in larger towns in the region. These connected people with divergent goals together to cre-
businesses and their workers were attracted by the low ate a shared understanding of the uncertainties surround-
cost of living and the availability of many outdoor recre- ing the transition to democracy. In northern Wisconsin, a
ation opportunities, but the expansion of towns pro- team of scientists created an initial set of scenarios to be-
duced water-quality problems along the Wisconsin River gin a scenario-planning process among a broad group of
and decreased fishing quality. In response to these prob- stakeholders. As we hope these examples make clear,
lems, town residents developed a town-centered system scenario planning can be modified in a multitude of ways
of regional environmental regulation. Some rural resi- to fit a particular context. We believe it can be a versatile
dents profited from these policies, whereas others were “soft” tool for conservation biology.
angered as the cost of water, wells, and fishing increased. There are not many examples of scenario planning that
In 2025, water quality and fishing near towns were rela- directly relate to biodiversity conservation. The Stockholm
tively poor but were improving. Due to strict regulation, Environment Institute’s global scenarios project ( Ham-
water and fish populations in rural lakes were high in mond 1998; Raskin et al. 1998 ) developed both global
quality, but many rural residents felt that town residents and regional scenarios based on extensive data and
had expropriated their ecosystems. quantitative simulation models. In other situations, sce-
In the “Lake Mosaic” scenario, the number of lake- narios have been based on stories of the future gener-
front second homes continued to increase, leaving few ated by small groups of rural villagers ( Wollenberg et al.
lakes undeveloped. Initial differences among lakes and 2000). The European Commission has funded a set of re-
their residents were amplified as new arrivals chose lo- gional scenario studies that are nested within a larger Eu-
cations that matched their interests. This process led to ropean development scenario study. This project com-
a social and ecological divergence among lakes oriented bines quantitative computer modeling with intensive
more toward motor sports and lakes that were quiet and stakeholder workshops (Rotmans et al. 2000). In the Neth-
less modified. Lake residents often organized themselves erlands, various forms of scenario planning involving plan-
into increasingly powerful associations. Some lake asso- ners, stakeholders, and scientists have been conducted
ciations restricted access to lakes and their fish popula- for a variety of rural development issues (Schoute et al.
tions, whereas other associations removed woody debris, 1995). However, the focus of all these scenarios is on so-
expanded lake access, and created artificial beaches. The cial rather than ecological dynamics.
focus on individual lakes led to conflict over issues that Despite their differences, the scenario exercises we
involved connections among lakes, such as road build- have presented share many features. In particular, each
ing, boating, fishing, and water quality. These conflicts considered a range of possible futures whose dynamics
left interlake issues largely unregulated, which allowed were analyzed in some detail. Desirable and undesirable
the expansion of many features of suburban develop- outcomes were described, and insights were gained into
ment, such as strip malls, that residents did not like. Al- the kinds of mechanisms that might produce them. In
though many residents were unhappy with these regional each instance, the purpose of the exercise was broader
changes, most believed their lakes were well managed. In than achieving a single, narrowly defined solution. At the
2025, water quality was generally good, except in some conclusion of most of these exercises, people took ac-
of the larger, shallower lakes near towns. Fishing quality tion using available levers—such as policy, organizational
was poor across the entire region, however, with the ex- priorities, and education—to shape their own future.
ception of some lakes that had been effectively privatized.
These scenarios illustrate that, due to the northern
Wisconsin landscape, the future of ecosystem services is Conclusion
likely to be heterogeneous. They were designed to start
a conversation among people in northern Wisconsin Conservation biology continually confronts situations in
about alternative futures and to begin a process of evalu- which decisions must be made in the face of uncertainty.
ating policies in terms of how they shape the ability of We suggest that the appropriate response to uncertainty
the northern highlands lake district to respond to poten- depends on the degree of uncertainty and the degree to
tial risks and to benefit from possible opportunities. which a system can be controlled. When control is diffi-
cult and uncertainty is high, scenario planning is an ef-
Implications of Scenario Examples fective way of coping. In other situations, hedging, adap-
tive management, and optimal management may be
These three scenario exercises differ in their approaches more appropriate (Fig. 1).
to uncertainty and in how they were constructed. In the Scenario planning, although potentially rewarding, risks
Shell Oil case, scenario planning was used to allow a falling into the same traps as other planning or modeling
well-defined actor (Shell) with a clear goal (to maximize exercises. However, a variety of factors, such as overly
Conservation Biology
Volume 17, No. 2, April 2003
Peterson et al. Scenarios as Tools for Conservation 365
Conservation Biology
Volume 17, No. 2, April 2003
366 Scenarios as Tools for Conservation Peterson et al.
Rogers, D. J., and S. E. Randolph. 2000. The global spread of malaria in van der Heijden, K. 1996. Scenarios: the art of strategic conversation.
a future, warmer world. Science 289:1763–1766. Wiley, New York.
Rotmans, J., M. van Asselt, C. Anastasi, S. Greeuw, J. Mellors, S. Peters, Wack, P. 1985a. Scenarios: uncharted waters ahead. Harvard Business
D. Rothman, and N. Rijkens. 2000. Visions for a sustainable Eu- Review 63:72–89.
rope. Futures 32:809–831. Wack, P. 1985b. Scenarios: shooting the rapids. Harvard Business Re-
Sarewitz, D., R. A. Pielke, and R. Byerly. 2000. Prediction: science, view 63:139–150.
decision making, and the future of nature. Island Press, Washing- Walters, C. J. 1986. Adaptive management of renewable resources.
ton, D.C. Macmillan, New York.
Schoemaker, P. J. H. 1991. When and how to use scenario planning: a Westley, F., S. R. Carpenter, W. A. Brock, C. S. Holling, and L. Gunder-
heuristic approach with illustration. Journal of Forecasting 10:549– son. 2002. Why systems are not just social and ecological systems.
564. Pages 103–120 in L. Gunderson and C. S. Holling, editors. Panar-
Schoemaker, P. J. H. 1995. Scenario planning: a tool for strategic think- chy: understanding transformations in human and natural systems.
ing. Sloan Management Review 36:25–40. Island Press, Washington, D.C.
Schoute, J. F. T., P. A. Finke, F. R. Veeneklass, and H. P. Wolfert, edi- White, D., P. G. Minotti, M. J. Barczak, J. C. Sifneos, K. E. Freemark,
tors. 1995. Scenario studies for the rural environment. Kluwer Aca- M. V. Santelmann, C. F. Steinitz, A. R. Kiester, and E. M. Preston.
demic Publishers, Dordrecht, The Netherlands. 1997. Assessing risks to biodiversity from future landscape change.
Schwartz, P. 1991. The art of the long view: paths to strategic insight Conservation Biology 11:349–360.
for yourself and your company. Doubleday, New York. Wollenberg, E., D. Edmunds, and L. Buck. 2000. Using scenarios to
Scott, J. C. 1998. Seeing like a state: how certain schemes to improve make decisions about the future: anticipatory learning for the adap-
the human condition have failed. Yale University Press, New Ha- tive co-management of community forests. Landscape and Urban
ven, Connecticut. Planning 47:65–77.
Conservation Biology
Volume 17, No. 2, April 2003