Learning to Manage While Managing to Learn - resources on ecology and computer models

The Ecological Detective Confronting Models with Data Ray Hilborn and Marc Mangel 1997; 330 pp. $24.95 Princeton University Press

An excellent accessible introduction to ecological problem solving. --GP

"Each of the case studies we use to illustrate a particular point is a bona fide research study conducted by one of us. Even so, some readers of drafts accused us of the unpleasant and unprofessional, but too common (especially in evolutionary biology), behavior of setting up "strawpersons" just to knock them down ... This charge is unfair. These apparently ridiculous models were in fact proposed and used by pretty smart people. Why? Because they had no alternative models. Our view is that the confrontation between more than one model arbitrated by the data underlies science. If there is only one model, it will be used, whether the questions concern management (as in the Serengeti example) or basic science (as in the insect oviposition example). Without multiple models, there is no alternative.

Adaptive Management of Renewable Resources C. J. Walters 1986

Out of Print, but available from Fisheries Center at the University of British Columbia, www.fisheries.ubc.ca /Books/books.htm

This book is a classic. It focuses on the organizational technical and modeling aspects of adaptive management. The book's early chapters are very accessible while the later ones provide a wealth of technical detail on mathematical approaches to management under uncertainty. --GP

"When you reject the extreme stances and recognize modeling as a very human way of groping for understanding, it should be obvious who will benefit most from it: those who engage in it directly A great deal of money has been wasted by government agencies on contracts to model builders, in the hope that grand predictive models will be produced and then used by the agency The modelers certainly learned a lot from these efforts, and have produced many lovely (and largely ignored) reports detailing formulations, predictions, and uncertainties. A few of the models have seen some use, but mostly as interactive computer games that are not taken seriously, or as generators of thick printouts to impress audiences who will never read them.

Compass and Gyroscope Integrating Science and Politics for the Environment K. Lee 1995 (reprint ed.); 243 pp. $19.95 Island Press

In elegantly written introduction to adaptive management that emphasizes social learning and politics. --GP

"I have come to think of science and democracy as compass and gyroscope--navigational aids in the quest for sustainability Science linked to human purpose is a compass: a way to gauge directions when sailing beyond the maps. Democracy, with its contentious stability is a gyroscope: a way to maintain our bearing though turbulent seas. Compass and gyroscope do not assure safe passage through rough, uncharted waters, but the prudent voyager uses all instruments available, profiting from their individual virtues.

Conservation Ecology www.consecol.org Free

This is a five-year-old, peer-reviewed, Internet-based ecology journal that is the leading journal on theory and practice of adaptive management. Its editor-in-chief is C.S. Holling, one of the founders of adaptive management. --GP

Nature continually changes. Yet people and animals depend upon a reliable supply of clean air, fresh water, and fertile soil. How can we maintain the ecologies that we need and love amid continual, unpredictable change?

Ecological management offers an answer, but is difficult. Like walking in the dark, ecological management is an activity full of uncertainty. To make matters worse, the continual change of nature makes you unsure of where you stand, while nature's unpredictability means that those snuffling sounds you hear in the distance could be coming towards you.

In nature, uncertainty is inescapable. While collecting ecological data can help reduce uncertainty, it can't eliminate it all. Consider Lake Mendota, Wisconsin. Scientists have extensively studied this lake for over a hundred years. They understand a lot about the lake--the dynamics of fish, what controls aquatic weeds, and when it will turn green. However other dynamics remain uncertain. Lake Mendota is currently soaking up a fair amount of fertilizer from the farms in the lake's watershed. One of the world's top lake ecologists, Prof. Steve Carpenter, recently attempted to predict if the lake would be irreversibly changed by these inputs of nutrients. But even with excellent data and solid ecological theory, the complexity of nature prevented him from being able to predict whether changes will be irreversible.

Ecologists have nevertheless been getting better at sketching the big picture, and figuring out effective ways to cope with this uncertainty. There are a variety of ways to manage unpredictable complexity.

People can simply ignore it, and try to bull their way through. They can seek solace in simplistic explanations that claim certainty. In ecological management, this approach manifests itself as top-down, command-and-control practice.

Command-and-control management assumes that effective policies have been carefully planned to solve problems, and therefore that technocrats know best. When the policy fails, people become the scapegoats: the failure is all due to human inability or unwillingness to follow, enforce, or implement the commands.

However, as has been repeatedly demonstrated, in places ranging from Maoist China to the capitalist USA, command-and-control management typically breeds surprising ecological disasters. Natural resources collapse amidst social and economic conflict. We might put it this way: while walking in the dark, just confidently stride toward where you think you want to go. Morale is high, and once in a while this approach works well. Usually, it results in a trip to the emergency room.

Ecological policies are not "solutions" because we do not (and never will) know enough to "solve" an ecosystem. All management policies implicitly embody someone's concept of how an ecosystem works. When the inner workings of an ecosystem are frankly unknown and unknowable, then many alternate conceptions of an ecosystem are plausible. Many models are equally consistent with existing data.

Policies should properly be viewed as questions rather than answers. The most important thing is not selecting the "correct" policy and imposing it, but managing in a way that allows people to learn from managing.

This approach is called adaptive management. Adaptive management consists largely of creating effective ways to learn. Its three main tools are simulation modeling, modeling workshops, and management experiments.

Effectively applying these tools requires a mix of social and technical skills that are best carried out by teams.

Modeling

Computer models integrate existing experience and scientific information. A model creates a simplified version of the real world. This cartoon world provides a flexible and forgiving arena for experimentation. This can reveal some practical possibilities that are by no means obvious from the situation on the ground. Of course, computer models do not "solve" management problems, but they reduce uncertainty and suggest new approaches. They can illuminate the chief, credible, alternate ways of conceptualizing a problem.

Computer models are most useful when they are built with specific questions in mind. New questions then require new computer models. To encourage question asking, adaptive management focuses not just on the model itself, but on the process of model building. The creation of a new model means conceptualizing the system under study. This is a powerful way to improve management, because a new model suggests new policies.

Modeling Workshops

Intensive modeling workshops bring together a small group of research scientists, managers, policy makers, and other stakeholders. Over a series of meetings, they develop a model. A limit of two or i three days seems to work best; it keeps discussion and activity focused. The group goal of producing a sensible model encourages participants to synthesize their knowledge and experience.

The aim of the model workshop is to identify the areas of ignorance as well as knowledge. A good modeling workshop identifies issues of consensus, as well as disagreement. A good modeling workshop also has a strong focus on a specific problem; this winnows down the nearly infinite set of things that are uncertain or unknown. An effective workshop reveals the uncertainties most directly relevant to the issue at hand.

Ideally, a series of workshops should produce a rough computer model, identify the key uncertainties in this model, and create ways of addressing these uncertainties. Running a workshop to achieve these goals is a difficult task. It requires a mix of organization, creativity, political savvy, verbal lucidity, and a broad understanding of both human and ecological aspects of the system. Good workshops can produce both ecological understanding and social flexibility.

Management Experiments

Frequently, a modeling workshop identifies major gaps in knowledge. These gaps concern large activities (such as the impacts of dam removal) that produce whole-ecosystem effects. Small-scale scientific experiments cannot evaluate these actions. Often the quickest, most effective way to fill the gaps is through focused, large-scale management experiments.

Management interventions can be used as tools to probe the functioning of an ecosystem. Interventions are treated as experimental manipulations. They are designed to test key hypotheses.

For example, prescribed fire is used to maintain forest communities--but there is uncertainty over what fire frequency to use. Traditional management would estimate and apply the "best" fire frequency across the board. Adaptive management would apply a range of different fire frequencies to different portions of the landscape.

Introducing this variation into the system increases the costs of management. But it also teaches managers about the genuine effects of fire frequency in the real world. This learning is the path to improvement. The experimental use of multiple fire frequencies allows the effects of weather and other chance events to be uncovered in a way that is impossible with uniform command-and-control.

Ecological management is a process grounded in the local. It depends upon local constraints, the present state of local institutions, and the personalities of key people. Experimental management is difficult to achieve in a political context where people claim or expect certainty.

Ecological problems are often politically polarizing. Conflicts divide people into camps. Each camp feels that it would be politically disadvantageous, a sign of weakness or lack of resolve, to publicly acknowledge any uncertainty in their view of the system. Therefore, adaptive management workshops do not just build models. They use those models to create new networks of political actors, who share some common understanding and experience. Scientists, managers, and other stakeholders must develop experimental institutions and effective ways of working together across institutional boundaries. These networks then have enough strength and resiliency to implement the plan they have developed.

Conclusions

Wandering around in the dark is best done carefully. It requires knowledge of where you are, and a plan to reach some eventual goal. When we develop a model, we find ourselves forced to think in depth about what we need to know. Workshops help people test their own knowledge and develop better, shared models of the world.

Management experiments test these models against the world. Testing your path may bruise your shins, but it should prevent you from falling off a cliff.

Access to Tools

A diversity of software is used for modeling. As a working ecologist, I often begin with diagrams. These diagrams can be frequently revised, and may eventually lead to a quick-and-dirty computer model.

When starting off, I favor Microsoft Excel, or the more sophisticated Matlab mathematics software. While Excel allows quick and sketchy work to get up to speed, Matlab allows far more sophisticated and complicated models to be constructed, analyzed, and displayed. Matlab and Mathematica are similar analysis programs. Detailed local models often have to be specially programmed using C++ or Visual Basic. --GP

Matlab
Windows, Linux, and Unix versions
The MathWorks, Inc.
508/647-7000, www.mathworks.com

A student version is usually available for about $100. The commercial version costs about $2,000.

Microsoft Excel
Windows and Macintosh versions
$499 as part of Microsoft Office
At computer stores or www.microsoft.com

Excel is an easy-to-use, accessible calculation tool. Most Windows users have it on their computers already. Simple models can be made quickly with spreadsheets, giving you a feel for how mathematical relationships become rich and complex.

Garry Peterson ia an ecologist with the Center for Limnology at the University of Wisconsin in Madison. He is a member of the Resilience Alliance and the Balaton group, two international networks of people engaged in the theory and practice of building sustainable societies.

COPYRIGHT 2001 Point Foundation
COPYRIGHT 2001 Gale Group