| Modeling in Natural Resource Management: Development, Interpretation, and Application. Edited by Tanya M. Shenk and Alan B. Franklin. 2001. Island Press, Washington, DC. 223p. US$28.00 paper. ISBN 1-55963-740-4. |
| It would be difficult to disagree with the proposition that effective natural resource management is a complex undertaking. Inter alia, this undertaking involves a careful delineation of an objective function, the identification of one or more state variables, a chronicling of the instruments available to a natural resource manager for control purposes, and an adequate understanding of the ways in which one or more kinds of uncertainty affect the management undertaking. What role can formal modeling play in shedding light on the difficult task of natural resource management? This question has increasingly been asked by both researchers and practitioners. In addition, this is also the central question that is addressed in this edited book. Specifically, the 12 chapters of this book present overviews of various models, they develop and interpret individual models, and finally, they apply models to specific management contexts. In Chapter 2, James Nichols presents his views on the functions served by models. In his view, there are 3 classes of model use. Specifically, all models are either theoretical, empirical, or decision-theoretical. It is noted that theoretical models can be very useful in decision-theoretic settings in natural resource management because "models with substantial mechanistic differences may lead to very similar management policies" (p. 15). Further, the author points out that although this point is not often recognized, "modeling exercises that are "theoretical" in the sense of involving no confrontation with data can be extremely useful" (p. 16). Although this chapter provides a helpful 3 part classification of models, it occasionally makes statements that are too insouciant. For instance, on one occasion, readers are asked to "exploit optimal stochastic control methodologies to assist in the design of studies directed at discriminating among multiple hypotheses" (p. 22). This is all well and good but the devil here is truly in the details. Which methods does one exploit? What is the best way to model one or more than one kind of stochasticity? Finally, how does one determine whether the multiple hypotheses have been framed appropriately? Unfortunately, these sorts of questions receive inadequate attention in this chapter. Marc Mangel, Oyvind Fiksen, and Jarl Giske discuss the role and the properties of theoretical and statistical models in chapter 4. In contrast with the approach presented in Chapter 2, these authors propose a somewhat different 3 part classification of models. In particular, we are informed that "it is possible to classify models broadly as statistical, theoretical, or logical" (p. 57). The authors rightly caution the practitioner that theoretical models should not be evaluated solely on how well they fit the data. As they point out, theoretical models may perform well when judged on the basis of criteria "such as elegance, internal logic, and explanatory power" (p. 60). This chapter concludes by providing a useful 5 part compendium of pitfalls to avoid when connecting models to data. Specifically, one should (i) avoid too many uncertain parameters, (ii) compare multiple models with data, (iii) be thinking about alternate models, (iv) not go places where there are no data, and (v) not confuse statistical and theoretical models. Aspects of optimal decision making are nicely discussed by Michael Conroy and Clinton Moore in Chapter 6. This chapter begins by correctly pointing out that renewable resource management "typically involves, at its core, an optimization problem of intrinsically dynamic systems" (p. 91). Although this is certainly true, one wishes the authors had pointed out that the renewable resources that are routinely the object of managerial interest are not only dynamic but also stochastic in nature. Following this introductory discussion, the authors go on to make 2 useful points. First, they note that although dynamic programming is a useful technique to use to solve optimal control problems, one must understand that owing to the so called "curse of dimensionality," this technique "works only for systems in which there are just a few state variables and decisions" (p. 93). As such, on occasion, a researcher will be forced to use simulation models. Although these models may be used to find optimal solutions to a specific problem, "they provide no guarantee that a solution is optimal" (p. 93). Second, for many natural resource management problems, useful insights will only be gained by using the idea of adaptation in concert with optimization. How do animals select resources? This interesting question forms the subject matter of the terse Chapter 9 by Lyman McDonald and Bryan Manly. To answer this question, one must model effectively (of course!) with resource selection functions. But what constitutes effective modeling? The authors provide four utile rules of thumb. First, one should clearly state the underlying assumptions, the relevant scale, the animals under study, the time period under consideration, and the study area. Second, one should critically assess the underlying assumptions with the collected data. Third, one should use flexible regression models or models that are justified by the sampling design. Finally, the analyst should assume "responsibility for predictions based on the resource selection function" (p. 144). Early on in Chapter 1, Tanya Shenk and Alan Franklin—the editors of this book—indicate that one of their key objectives is to "present basic principles for understanding and evaluating models" (p. 2). By presenting such principles, the editors hope to "demystify models" (p. 2). In all fairness, it should be noted that the editors are only partially successful in accomplishing the above objective. In part, this is because the individual chapters all present somewhat different perspectives on what constitutes effective modeling in natural resource management. In addition, some of the chapters are fairly general in their discussion of modeling. In contrast, other chapters provide too elaborate an account of the details of specific problems. Further, even though there exists a vast literature in economics on natural resource management, one would not know that by perusing this book. Finally, the last chapter of this book correctly recognizes that "agricultural and environmental policy often is driven, not by the science of wildlife ecologists and managers, but rather by politics and economics" (p. 204). This recognition notwithstanding, there is not a single chapter in this book that is written by an economist, or for that matter, by a political scientist. Given these drawbacks, it is not possible to praise this book unequivocally. Even so, let me conclude this review by pointing out that this book does provide a helpful overview of some of the basic principles for comprehending and assessing models that are frequently used in natural resource management.—Amitrajeet A. Batabyal, Rochester Institute of Technology, Rochester, New York. |