- Reviewed by
Having begun research on possible ecological effects of global warming in 1968 and developing one the first major methods to forecast such effects on forests, I looked forward to reviewing Saving a Million Species, because Lee Hannah, the book's editor, writes that the book's purpose is to determine "what is the extinction risk associated with climate change, and how many species may perish," and that the book "addresses this important question by synthesizing the [scientific] literature" (p.3). But of course an edited book with 20 chapters and 31 authors is a mixed bag.
The book is divided into six sections: Introduction; Refining First Estimates; Current Extinctions; Evidence from the Past; Predicting Future Extinctions; and Conservation Implications. The four chapters in "Evidence from the Past" are excellent, the data are recent, evidence is advancing rapidly, and the analysis is careful and objective, therefore helpful in thinking about current climate change. For example, in "Terrestrial Ecosystem Response to Climate Change during the Paleogene," authors William C. Clyde and Rebecca LeCain discuss the dramatic Paleocene-Eocene Thermal Maximum (PETM), a warming that occurred about 55 million years ago. They write: "Migration was an important coping mechanism for both mammal and plant taxa during the PETM. Mammals appear to have exhibited more community coherence . . . whereas plant taxa seem to have reacted more individualistically. This increased dispersal activity did lead to high extinction rates in most cases but not necessarily to lower diversity, given the offsetting effects of increased origination rates."
The book's first sections consider recent computer forecasting of possible biodiversity effects of global warming, dealing primarily with two of the four methods commonly in use today for these forecasts: climate envelope and species-area curve models, paying less attention to two other kinds of models: those in which individual species are the units of interest and those that directly integrate biodiversity within a general circulation (global climate) model.
One of the best of these chapters is "The Use and Misuse of Species-Area Relationships in Predicting Climate-Driven Extinctions." Authors John Harte and Justin Kitzes carefully analyze the theoretical limitations of this method, noting that "numerous conceptual and practical problems permeate this seemingly solid and straightforward approach." But they pay little attention to a primary failing of this method, which is that the data on which this method depends are crude at best and often lacking. For example, the best-known of the papers using this method, Thomas et al. (2004), concluded that the area of the boreal forest would decline 4%. But an empirical study of the biomass stored in the boreal forest of North America showed that botanical maps of the North American forests differed by a factor of two—200% in the area defined as boreal forest, eclipsing the forecast loss of 4%, and the estimate of biomass had a 95% confidence interval of more than 20% (Botkin and Simpson 1990). Therefore a forecast loss of 4% means little.
Climate envelope models, currently popular, are discussed in several chapters, without sufficient discussion of their fundamental limitations. These overlay the current distribution of a species or group of species on a current climate map (usually just temperature isotherms), typically represented as a box representing the biodiversity range. Then the box representing the biodiversity group is overlaid on a map of a forecasted change in climate. Among the imbedded assumptions are that the current biota distribution is in a steady state with the current climate, and that a single climate variable is sufficient to represent the entire causes of biota distribution.
The difficulties climate envelope models face are made clear by the history of Kirtland's warbler, the first songbird in the U.S. to be subjected to an annual monitoring that was a complete census. Ornithologists were shocked to discover that the population dropped by half in the first decade of this annual census. This warbler was known to nest in a very limited area in southern Michigan. Comparison of the nesting area...