Adaptive Management of Renewable Resources

Adaptive Management of Renewable Resources

An overview of an IlASA book written by Carl Walters and published by Macmillan Publishing Company EXECUTIVE REPORT 12 A y gust 1986

Ezecutive Reports b r i n g t o g e t h e r t h e flndings of r e s e a r c h d o n e at IIASA a n d e l s e w h e r e and summarize them f o r a wide r e a d e r s h i p . This o v e r v i e w d o e s n o t n e o e s s a r i l y r e p r e s e n t t h e views of t h e sponsoring o r g a n i z a t i o n s or of individual workshop p a r t i c i p a n t s . Copies of t h i s Exeoutive R e p o r t oan b e obtained from t h e Publioations Department, I n t e r n a t i o n a l I n s t i t u t e f o r Applied Systems Analysis, Schlossplatz 1, A-2361 Laxenburg, Austria.

Copyright O 1986

S e c t i o n s of t h i s publication may b e r e p r o d u c e d i n magazines a n d n e w s p a p e r s with acknowledgment to t h e I n t e r n a t i o n a l I n s t i t u t e f o r Applied Systems Analysis. P l e a s e send t w o tear s h e e t s of a n y p r i n t e d r e f e r e n o e to t h i s report to t h e Publications Department, IIASA, A-2361 Laxenburg, Austria.

FORE WORD

This Executive Report reviews t h e book A d a p t i v e Manage- m e n t of R e n e w a b l e R e s o u r c e s by Professor Carl Walters. I hazard a prediction. It will become a classic for t h e science and management of renewable resources. As such it will stand with t h e earlier classics in t h e field

-

Holt's (1957) "On t h e dynamics of exploited fish popula- tions ", Ricker's (1958) "Handbook of computations for bio- logical statis tics of fish populations ", and Ivlev's (1961) E x p e r i m e n t a l E c o l o g ~ of t h e F e e d i n g F i s h . All these con- cern fisheries ecology, economics, and management. In t h e field of fisheries, basic empirical and theoretical science, mathematics, and hard management practice have been com- bined more effectively than for any other renewable resource. Professor Walters extends that base into exam- ples that cover a full range of living resources-forests, wildlife, and range resources.

One could call this a book in applied ecology, but that would be wrong. It is basically a book on human behavior and management science. The system that Professor Walters defines is one that includes t h e fish, t h e fishermen who har- vest them, and t h e bureaucrats who attempt t o monitor and manage both. As a consequence, its central theme is on human learning of t h e laws that determine how a partially observed system functions.

We do not learn from a system that is constant. This is not serious if the system is known, is static, and presents no surprises. But resource systems a r e exactly the opposite.

They are known only very partially, which will always be so;

they a r e dynamic and they produce endless surprises -from the collapse of fisheries to the reemergence of other ecosystems. And the act of management and harvesting changes the fundamental structure of the resource itself.

Age structure changes; genetic stocks change; interacting species disappear and new ones emerge; climate and ocean conditions themselves become modified by human act ions producing unexpected resource consequences.

The approach Professor Walters presents is rooted in t h e reality of this change and of the inherent unknowability of the evolving character of t h e system. Hence management has to be adaptive. And it has to be actively so. In this way management designs become explicit experiments to manipu- late systems into regimes of behavior that are most condu- cive to learning. It combines, therefore, an equal emphasis on producing economic return and social persistence.

This body of work owes much of its character and uniqueness to an important s e t of conjunctions that occurred in t h e very early days of the International Institute for Applied Systems Analysis (IIASA). In 1974-75 Professor Walters was Deputy Leader of IIASA's Ecology Project. He found, during the same period, a happy intersection of opportunity. His experience in systems ecology and fisheries management began to move in major new directions opened by Tjalling Koopmans' kind of economics, George Dan tzig's optimization studies , and Howard Raif f a's decision theory. It is an example of the power of intersecting t h e different experiences and strengths of individuals of uni- formly outstanding competence.

The book owes its sweep in part to those connections.

If that was all, however, it might be of only theoretical interest. But Professor Walters has turned the book into one of profound applied consequence by testing and applying t h e ideas within the hard reality of resource industries and resource management agencies.

It is that combination of empirical scholarship, of theory and application, that, in my view, will make this book a classic.

C.S. Holling Institute of Animal Resource Ecology Vancouver, BC, Canada

References

Beverton, R.J. and Holt, S . (1957) On t h e dynamics of exploited fish populations. f i s h . Invest., London, S e r . 2 , 19.

Ivlev, V.S. (1961), E z p e r i m e n t a t EcoLogy oj' t h e Feeding oj' f i s h (Yale University P r e s s , New Haven).

Ricker, W.E. (1958) Handbook of computations f o r biological statistics of fish populations. h t t . Fish. Res. Board C a n . , 119.

CONTENTS

Foreword 1 Introduction

2 Adaptive Management and Uncertainty 3 Model Building and Parameters

4 Feedback 5 Conclusions

8 Contents of Adaptive Management of Renewable Resources

7 The Author

iii 1

3 9 13 18

8 Purchasing Adaptive Management of Renewable Resources

Renewable natural resources provide important contribu- tions t o food, fiber, and recreation in many p a r t s of t h e world. The economies of some regions a r e heavily dependent on fisheries and forestry, and consumptive use of wildlife (hunting) is a traditional recreational pastime across Europe and North America. The management of renewable resources usually involves public agencies t h a t a r e responsible for harvest regulation, and of t e n production enhancement, s o as t o provide sustainable yields into t h e long-term f u t u r e (resource husbandry). The t r a c k record of such agencies has been spotty: many resources have been mined t o low lev- els before effective harvest regulation could b e developed, while o t h e r s have been managed s o conservatively a s t o miss major harvesting opportunities.

Three key features of renewable resources have made them difficult t o manage. First, sustainable production depends on leaving behind a "capital" stock a f t e r each har- vesting, and t h e r e a r e definite limits t o t h e production r a t e s t h a t this stock can maintain. Second, harvesting is normally undertaken by a community o r industry of harvesters whose activities (investment, searching, e t c . ) a r e not completely monitored o r regulated, s o t h a t dynamic responses, such a s overcapitalization of fishing fleets, a r e common. Third, t h e biological relations hips b e t ween managed stock size and

production r a t e s arises through a complex interplay between t h e organisms and t h e i r surrounding ecosystem; for any particular population, this relationship cannot b e predicted in advance from ecological principles and must ^, instead, b e learned through actual management experience.

ADAPTIVE MANAGEMENT AND UNCERTAINTY

Most management agencies maintain monitorirlg and research activities that a r e aimed at understanding the stock- production relationship. However, research activities are often not closely integrated with management decision mak- ing, and scientists have traditionally recommended conserva- tive harvest policies so as to protect the population until b e t t e r biological understanding can be accumulated. A fun- damental presumption in such recommendations is that the ecological basis for production can be researched on a piecemeal, experimen tal-componen t s basis, and the results eventually synthesized into an overall understanding of how the resource behaves. However, various at tempts to conduct such syntheses, in the form of predictive mathematical models of resource behavior, have not been notably success- ful; the modeling exercises have revealed large gaps in understanding of various processes that a r e difficult to study in the field o r laboratory, and predictions of optimum stock sizes of ten involve gross extrapolations beyond the range of recent historical or experimental experience (Fig- u r e 2.1).

Frustration with the linkage between science and management has led to the concept that management should be viewed as an adaptive process, in which regulatory and enhancement actions a r e treated as deliberate experiments

Figure 2.1- Relationship between number of sockeye salmon al- lowed t o spawn in t h e F r a s e r River, BC, and number of resulting offspring measured as r e c r u i t s t o t h e fishery f o u r y e a r s l a t e r . Data a r e f o r 1939-73,-omitting e v e r y f o u r t h (cycle) y e a r begin- ning in 1942. The c u r v e s 71 and

v 2

with uncertain outcomes. This concept goes f a r beyond t h e traditional notion that uncertainties imply risks that should be accounted for through cautious decision making; risky choices a r e also seen in adaptive management as opportuni- ties to learn more about system potentials, and hence to have positive value in reducing t h e legacy of uncertainty that will be faced by future decision makers. Basic research is seen not as taking a lead in developing t h e understanding needed for making predictions, but rather as a means to b e t t e r understand the response patterns revealed by management (in hindsight) and as an exploratory investment that might uncover new policy instruments and options.

It would b e easy enough t o design a blind process of trial-and-error management t h a t would b e adaptive in t h e evolutionary sense t h a t major mistakes would t e n d not t o b e repeated. But such a process would b e unnecessarily waste- ful: b y analysis of historical experience in relation t o eco- logical t h e o r y and constraints, i t should b e possible t o design much more intelligent, d i r e c t e d s e a r c h e s for produc- tive and sustainable harvest policies. Thus, adaptive management is seen a s involving t h r e e essential tasks. First, it involves s t r u c t u r e d synthesis and analysis, through attempts t o build predictive models, of major processes and uncertainties; t h e objective h e r e is not t o build a single best prediction o r t o define a single best policy choice, but is instead t o identify a strategic range of alternative hypotheses t h a t a r e consistent with historical experience, but t h a t imply different responses (opportunities for improved harvest) outside t h e range of t h a t experience.

Second, adaptive management involves t h e use of formal optimization techniques t o s e a r c h f o r optimum policies t h a t account not only for existing uncertainties, but also for t h e effects t h a t c u r r e n t decisions will have on t h e uncertainties t h a t f u t u r e decision makers will face. (In o t h e r words, t h e adaptive manager attempts t o model not only t h e managed system, but also t h e d a t a gathering and learning process about t h a t system.) Third, adaptive management involves t h e design and implementation of improved monitoring programs for detecting system responses more quickly, along with t h e design of more flexible harvesting industries t h a t can respond t o unexpected changes quickly without undue economic o r social hardship.

A c e n t r a l controve,-s y in adaptive management concerns t h e question of whether i t is worthwhile t o engage in delib- erate and perhaps risky experiments involving substantial changes in harvesting r a t e s , thus allowing measurement of production r a t e s across a range of stock sizes. This involves two distinct issues, t h e f i r s t of which is not biological. To conduct variable harvest experiments means e i t h e r giving up harvests today in favor of possibly higher harvests in t h e future, o r else taking more today while risking losses in t h e

Table 2.1. Conventional v e r s u s a d a p t i v e a t t i t u d e s a b o u t t h e ob- j e c t i v e s of formal policy analysis (Table 22.1 in A d a p t i v e M a n a g e m e n t of Renewable Resources).

Conventional

(1) S e e k p r e c i s e p r e d i c t i o n s

(2) Build p r e d i c t i o n from detailed understanding (3) Promote s c i e n t i f i c

consensus

(4) Minimize conflict among a c t o r s

(5) Emphasize s h o r t - t e r m o b j e c t i v e s

(6) P r e s u m e c e r t a i n t y i n seeking b e s t a c t i o n (7) Define b e s t a c t i o n

from set of obvious a l t e r n a t i v e s

(8) S e e k p r o d u c t i v e equilibrium

Adaptive

( l a ) Uncover r a n g e of possibilities

(Za) P r e d i c t from e x p e - r i e n c e with a g g r e g a t e r e s p o n s e s

(3a) E m b r a c e a l t e r n a t i v e s (4a) Highlight difficult

t r a d e - o f f s

(5a) Promote long-term o b j e c t i v e s

(6a) Evaluate f u t u r e f e e d b a c k a n d learning

(7a) S e e k imaginative new options (8a) E x p e c t a n d p r o f i t

from c h a n g e

f u t u r e if s t o c k s are d e p l e t e d . This trade-off b e t w e e n p r e s e n t a n d f u t u r e values is seldom c l e a r - c u t , a n d t h e r e is seldom c o n s e n s u s among management a c t o r s ( h a r v e s t e r s v e r s u s c o n s e r v a t i o n i s t s , e t c . ) a b o u t t h e b e s t p o i n t t o aim f o r in t h e trade-off; a d a p t i v e management is u n n e c e s s a r y o r i r r e l e v a n t in situations w h e r e f u t u r e h a r v e s t s c a r r y l i t t l e weight in r e l a t i o n t o t h e p r e s e n t .

Beyond t h e fundamental i s s u e of values, t h e r e is a t e c h n i c a l issue t h a t modeling a n d optimization c a n h e l p t o resolve: t h i s is t h e issue of passive v e r s u s a c t i v e a d a p t a t i o n . A t r a d i t i o n a l p r e s c r i p t i o n from model b u i l d e r s h a s b e e n t h a t o n e should buiid t h e b e s t possible p r e d i c t i v e model, t h e n a c t as though t h i s model were c o r r e c t u n t i l e v i d e n c e t o t h e con- t r a r y becomes available. This passively a d a p t i v e a p p r o a c h t o management c a n work q u i t e well in c o n t e x t s w h e r e e v e n t h e nominal b e s t decision would b e informative, b u t i t c a n

Table 2.2 Conventional v e r s u s adaptive t a c t i c s f o r policy development and presentation (Tarble 11.2 in Adaptive Manage-

ment of R m s w a b l e Resources).

Conventional

(1) Committee meetings and hearings (2) Technical r e p o r t s

and p a p e r s

(3) Detailed f a c t s and figures t o

back arguments (4) Exhaustive

presentation of quantitative options (5) Dispassionate view (6) P r e t e n s e of s u p e r i o r

knowledge o r insight

Adaptive

( l a ) S t r u c t u r e d workshops (2a) Slide shows and

computer games (3a) Compressed v e r b a l

and visual arguments (4a) Definition of few

s t r a t e g i c a l t e r n a t i v e s (5a) Personal enthusiasm (6a) Invitation t o and

assistance with a l t e r n a t i v e assessments

result in managed stocks being locked into unproductive equilibria a t f a r from t h e best levels (see f i g u r e 2.1). A key problem for t h e adaptive manager is t o recognize when such an unproductive and uninformative equilibrium exists o r is likely t o develop; given t h a t recognition, formal optimization methods can be used t o compare passive adaptation with more daring options t h a t involve probing changes in harvest r a t e s .

Policy analysis for adaptive management involves some quite different attitudes than a r e conventionally held by scientists and analysts in t h e renewable resources fields (Table 2.1). The conventional attitudes (and goals of analysis) have arisen from t h e presumption t h a t biological uncertainties a r e small and can b e resolved through careful modeling; in such cases i t might, indeed, b e best to delib- e r a t e l y seek stable and productive equilibrium in resource stocks. The adaptive analysts attitudes given in Table 2.1 reflect a much more humble, if not pessimistic, viewpoint about t h e magnitude of uncertainties and t h e importance of seeking imaginative new ways t o deal with t h e s e uncertain- ties. Along with changes in attitudes, policy analysis f o r

adaptive management should involve some changes in tactics for policy development and communication (Table 2.2); t h e s e changes again reflect a more humble perspective about t h e need t o involve a variety of a c t o r s and ideas in policy formu- lation and decision making. In s h o r t , by explicitly revealing uncertainties and difficult choices related t o risks and time preferences, t h e adaptive analyst must discard any cloak of authority t h a t might b e fashioned from t h e conventional trappings (massive r e p o r t s , c h a r t s , etc.) of policy analysis.

Model building for renewable resource management has often been pursued under t h e assumption t h a t bigger is always b e t t e r , with t h e key t o successful prediction being more precise and detailed calculations. Adaptive policy design seldom involves very complicated models, for some very good reasons. First, with a bit of careful analysis it is often pos- sible t o show t h a t t h e details simply do not matter, a t least in comparison t o broader uncertainties about what factors to model in t h e f i r s t place. A good example of this problem occurred with t h e Peru anchoveta (Figure 3.1), t h e world's largest fishery; advisers t o t h e Peruvian government ago- nized in great detail over t h e ecology of t h e fish and i t s relation t o t h e El NiEo oceanographic phenomenon, but t h e y did not make an effective c a s e for t h e broader need t o regu- late t h e fishing industry s o t h a t recovery would be possible if a collapse did occur. Second, with a limited data base and a s model complexity increases i t becomes progressively more difficult t o estimate each model parameter with any statisti- cal precision; on t h e o t h e r hand, sensitivity of t h e model predictions t o each parameter does not necessarily decrease as t h e number of parameters increases. Third, and perhaps most important, models should b e understandable if t h e y a r e t o b e of value in stimulating imaginative s e a r c h e s for b e t t e r policy options and in clarifying possible outcomes in debates

t h a t involve a c t o r s with conflicting objectives. Particularly in conflict situations, complex models a r e more likely t o c r e a t e f u r t h e r confusion a n d d i s t r u s t , r a t h e r t h a n t o pro- mote t h e kind of mutual understanding t h a t is important t o cooperative problem solving in t h e f a c e of u n c e r t a i n t y .

1955 '59 '63 '67 '7 1 '75 '79

-

Estimated

Figure 3.1. Development of the Peruvian anchoveta fishery.

The sharp collapse in 1972-73 was apparently associated with a major oceanographic change known as El Ni"no. (Figure 2.1 in Adaptive Management of Renewable Resources.)

T h e biological a n d physical environments f o r renewable r e s o u r c e production a r e often changing in time, d u e both t o human influences on ecosystems a n d t o natural "climate"

changes on various time scales. Thus, i t is unwise t o assume c o n s t a n t p a r a m e t e r values f o r a n y r e s o u r c e production model a n d t o t r u s t t h a t o l d e r historical d a t a a n d e x p e r i e n c e are relevant t o t h e prediction of f u t u r e responses. F u r t h e r , i t is generally not possible t o a n t i c i p a t e t h e p a r a m e t e r changes b y using more detailed models t h a t s p e l l out t h e c a u s e s of change; usually, t h e effects of s e v e r a l possible causes a r e "confounded" in t h e historical d a t a s o t h a t t h e c o r r e c t one(s) cannot b e determined with a n y confidence a n d , in a n y case, t h e c o r r e c t causal a g e n t is likely t o b e

unpredictable in its behavior. A basic consequence of slow and unpredictable changes in production relationships is that uncertainty about the relationships will grow over time if t h e system is not disturbed regularly so as to sample a range of stock sizes. This means that management choices

Change in second policy variable

\

Figure 3.2. Koonce's donut. Changes in policy variables must be reasonably large to allow learning about policy effects, but very large changes imply unacceptable risks. ( f i g u r e 7.6 in A d a p t i v e Management o,fRenewable Resources.)

generate a donut-shaped pattern of possible outcomes regarding uncertainty (Figure 3.2). If management policies a r e held steady and the stock size remains near its histori- cal average, t h e manager is operating in a donut hole of growing uncertainty. Moderate disturbances and policy changes will result in enough informative variation to stay in a domain of decreasing uncertainty (the donut itself). Large

a n d indefensibly r i s k y d i s t u r b a n c e s define t h e outside of t h e donut. Thus, t h e donut r e p r e s e n t s a compromise o r bal- a n c e d level of variation where t h e manager a n d t h e harvest- ing industry can d e t e c t a n d profit from change; a major chal- lenge f o r t h e adaptive manager is t o define where t h i s domain lies in terms of t h e p r a c t i c a l policy instruments a t his o r h e r disposal and t h e objectives a n d constraints defined b y t h e harvesting i n d u s t r y and o t h e r a c t o r s involved in decision making.

Some management agencies attempt t o induce informative variation by making small policy changes (tinkering) o r b y not trying t o control stock sizes too precisely s o t h a t t h e effects of random, natural variations (dithering) a r e not fully dampened through responsive changes in harvest r a t e s . One objective in t h e development of adaptive management t h e o r y has been t o determine, by using formal optimization tech- niques, whether t h e tinkering approach is, in fact, any b e t t e r than purely passive adaptation o r t h e more extreme approach of making e i t h e r large changes o r no changes a t all. The optimization results available t o d a t e all point t o t h e same conclusion, namely t h a t tinkering (and related incre- mental approaches t o management) is not a wise approach.

Small changes have practically no value in resolving major uncertainties (effects a r e too small t o d e t e c t against t h e background noise caused b y o t h e r factors), y e t cause annoy- a n c e (or even severe hardship) for t h e harvesting industry.

In terms of harvest r a t e variation, long-term harvests a r e likely t o b e maximized b y following e i t h e r a passive adaptive approach (no deliberate changes) o r else making large and very informative experimental changes ( f i g u r e 4.1). In s h o r t , tinkering is not a good compromise when faced with a h a r d choice between doing nothing (living with uncertainty) and doing a really substantial experiment.

case A: G = ul") case B: G> ^{U ( O )}

Harvest rate, ut

Figure 4.1. Examples of how the long-term value of harvests can b e broken down into components as functions of harvest r a t e . The total value i s

ficl).

u ( O ) . In case B , even using u ( O ) i s informative since i t i s f a r from the historical average C. (Part of f i g u r e 8.20 in Adaptive Management of RenewabLe Resources.)

There a r e a t least two ways to avoid hard choices between passive and active adaptive policies. One is to make use of spatial structure within t h e managed system; most renewable resources are aggregates of smaller "replicate"

substocks that a r e likely to be informative about one another (display similar responses to disturbance). Provided that t h e replicates do not each have a "dependent economic community" (harvesters, processors, resort owners, etc.) that cannot easily move its activities to other replicates, there can be considerable flexibility to experiment with harvest rate trade-offs between replicates (increase harvest in some, reduce in others by moving harvesting effort) without significantly changing t h e overall performance (yields, employment generated, etc.) of the managed system.

Beyond offering opportunities for economic trade-offs between replicate substocks, spatially structured systems offer t h e possibility of scientific control (in t h e experimen- tal sense) of t h e effects of large-scale environmental factors that may simultaneously affect several replicates, but be

confounded within each replicate with t h e effects of local biological and policy changes.

A second way t o avoid hard choices is t o invest in b e t t e r monitoring programs (so that smaller changes can be detected) and in socioeconomic programs t h a t will confer greater flexibility t o respond when experiments s t a r t t o show unfavorable results. Often, high harvest rates and pro- duction enhancement programs a r e allowed t o continue long a f t e r their deleterious effects have become obvious, simply because cutting back on them would cause immediate and po- litically unacceptable hardships for t he harvesting industry.

Socioeconomic programs that might prevent this pathological dependence include license limitation (to prevent t h e number of harvesters from becoming too large in t h e first place), subsidies for retraining and investment in other industries, and insurance schemes t o tax t h e industry during good times s o as t o provide financial assistance during bad times

.

The most risky "experiments" in renewable resource management have involved populations t h a t a r e subject to increasing natural difficulties as stock sizes decline. For example, lake trout in t h e Laurentian Great Lakes of North America a r e preyed upon by a parasitic fish, t h e sea lam- prey (Figure 4.2). When trout a r e abundant, t h e number killed by lamprey is small compared t o t h e trout population size and t h e r e can be a stable "balance" or equilibrium. If trout harvest rates increase and their abundance declines, t h e number killed by lamprey does not decline proportion- ally (lamprey a r e efficient a t finding trout even when t h e trout a r e scarce), so t h e lamprey kill becomes progressively more important and can cause t h e trout population t o sud- denly crash t o a very low level. One management strategy in such situations is t o keep harvest rates very low, s o that t h e

"cliff edge" for sudden collapse is not approached. How- ever, trout yields a r e higher near t h e cliff edge and t h e edge moves in time (changing ecological parameters), so t h a t i t is difficult t o find a consensus on just how low a harvest r a t e is safe enough. An adaptive strategy, called a "surfing policy", would be t o let t h e harvest rates increase until a

A Harvesting effort, u Harvesting effort, u Option A : (stable, conservative) Option B : pathological surfing

A A

m N

. - _V1

Y U

0

-0 r X X

m

E

Option C: productive surfing

Harvesting effort, u

3 b

.- _L 0

Figure 4.2 T h r e e policy options f o r regulation of harvesting ef- f o r t on l a k e t r o u t in t h e G r e a t Lakes. In option A, e f f o r t i s k e p t low and steady. In option B, e f f o r t i s allowed to i n c r e a s e until a major collapse o c c u r s , and t h e n t h e r e i s a long r e c o v e r y period.

In option C, e f f o r t a l s o i n c r e a s e s until collapse s t a r t s , but detec- tion a n d r e s p o n s e to t h e collapse i s much f a s t e r . B a n d C are

"surfing" policies. ( f i g u r e 322.1 in A d a p t i v e Management of R m a b L e Resources.)

2 .- 0,

u >

W 0 I b 0 b

b 0

collapse begins, then cut back quickly so as to allow recovery. The success of such a policy depends critically on two factors noted above:

(1) How early t h e collapse is detected (quality of the moni- toring system).

(2) The flexibility of t h e management system t o quickly cut back on harvests.

In t h e lake trout example, flexibility is t h e key limiting fac- tor: collapses can be quickly detected with existing monitor- ing programs, but harvest r a t e reductions a r e highly politi- cal issues (a large tourism industry depends partly on t h e trout fishery) requiring perhaps years (and very clear evi- dence of collapse) t o debate and implement. If greater flexi- bility could be achieved, trout yields under a surfing policy would be cyclic (collapse-recovery-collapse.. .), but would be higher on average than is now considered safe.

- CONCL USIONS

There is still much to learn about adaptive management, par- ticularly in terms of how to design imaginative policies that make use of spatial replication and permit more flexible responses to natural and man-made surprises. The key prob- lem now is not how to gather more data or construct more models in the hope of making more accurate predictions, but rather to develop a broader consensus about what the major uncertainties a r e and about the crucial role of ongoing management decisions in providing the experiments needed to resolve these uncertainties. When we begin to more widely embrace uncertainties and hard decision choices, rather than to pretend that future study will do t h e job, human ingenuity will be quick to find the imaginative options and wise compromises that are so badly needed.

CONTENTS OF ADAPTIVE MANAGEMENT OF RENEWABLE R E S O U R C E S

Introduction

Objectives, Constraints. and Problem Bounding A Process for Model Building

Models of Renewable Resource Systems Simple Balance Models in Applied Population Dynamics

Embracing Uncertainty The Dynamics of Uncertainty Feedback Policy Design Actively Adaptive Policies

Adaptive Policies for Replicated Systems Adaptive Policy Design for Complex Problems

- T H E A U T H O R

Carl Walters is a Professor a t t h e I n s t i t u t e of Animal R e s o u r c e Ecology, The University of British Columbia, Van- couver, Canada. Professor Walters w a s Deputy Leader of IIASA's Ecology Project in 1974-75 a n d in 1982-83 h e w a s Leader of t h e Project on Adaptive Policy Design f o r t h e Sus- tainable Use of Resources. He has b e e n heavily involved in t h e development of r a p i d techniques f o r teaching systems analysis a n d mathematical modeling t o biologists a n d r e s o u r c e managers, employing problem-oriented works hops a n d seminars. He also u n d e r t a k e s basic r e s e a r c h on prob- lems of adaptive control, as applied t o sequential decision making in renewable r e s o u r c e s , a n d maintains a field r e s e a r c h program in aquatic ecology.

PURCHASING ADAPTIVE MANAGEME17N7' -

OF' RENEWABLE R E S O U R C E S

A d a p t i v e Management of Renewable Resources by Carl Walters has been published by t h e Macmillan Publishing Com- pany. Copies of t h e book (ISBN 0-02-947970-3) a r e available through your local bookseller o r directly from Macmillan Publishing Company, 866 Third Avenue, N e w York, NY10022.