Useless Arithmetic: Why Environmental Scientists Can't Predict the Future (review)

We all use models—simplified representations of things, processes, and systems—in ways that range from helping us make decisions ranging from day-to-day personal tasks (going to a grocery store requires a mental model of one's current location, the location of the store, and the route between the two) to enhancing scientific understanding of the most complex patterns and processes of the universe. Some of the models of natural systems that we use are rigorously specified with foundations in first principles, derivation with mathematical logic, and extensive testing. By using these models we not only constrain our faulty intuition, we can also find unexpected relationships, develop tests of our theories, and predict phenomena that are as yet unobserved. Other kinds of models that we use to design structures such as bridges, roads, buildings, automobiles, aircraft, and other human-scale artifacts are based on simple factors such as the strength of materials and the measurable dynamics of fluid flow (I'm sure that my aeronautical and hydraulic engineering friends will scream at me when I call their system "simple," but in fact it is when compared with a beach, a watershed, a natural community of organisms, or the earth's atmosphere. Simple does not mean "easy."). The best use of models in science is to help figure out what we don't know about the systems that we study. We also try to predict the future with models. Whether or not there will be rain tomorrow is a statement based on a variety of weather models.

Pilkey and Pilkey-Jarvis's evocatively but inaccurately titled book is a very harsh indictment of a particular use of quantitative mathematical models in a few, but important, natural systems. The particular use (and source of title inaccuracy) has to do with politically or financially driven results from quantitative forecasting of the consequences of some manipulations of natural systems. This use is the engineering misapplication of models and not of the utility of natural-science models. These models are called "quantitative mathematical models" throughout the book. The authors name names, expose motivations, and suggest alternatives. The audience for this book is "non-specialists who are interested in nature and in the politics of working with the earth" (page xiv). Most people I know who fit this label are environmentalists interested in understanding why our society is so bad at living with natural systems and what we can do about it. So the tone of the book is already biased, but nobody would say that it is intended to be an unbiased look at the ways that we plan and justify our interventions in natural systems.

Geographers in the southeastern U.S. are already familiar with Orrin Pilkey, a long-time professor of geology at Duke University and vocal critic of coastal engineering policies and practices. His documentation of the vicissitudes of beach replenishment, the very useful series of "Living with the Shore," and his more recent critiques of coastal business-as-usual are all well known to anyone interested in coasts and beaches. Eight of eleven (SEDAAG states plus Louisiana) south-eastern U.S. states have coastlines, so most of the readers of this journal probably have some level of interest in the topic. His co-author, Linda Pilkey-Jarvis, brings expertise in both environmental contamination problems and public administration. The pairing of a serious academic gadfly with an environmental scientist who works professionally in the public arena makes for a broad, and very useful, perspective.

The book is organized around seven case studies in which quantitative mathematical modeling has been at least misleading and at worst highly damaging. Each case is described in a chapter: marine fisheries, the proposed Yucca Mountain nuclear waste depository, the effects of rising sea level as a consequence of global climate change, the use of the Bruun Rule for...