Predicting Species Occurrences: Issues of Accuracy and Scale

Abstract

J. M. Scott, J. H. Heglund, M. L. Morrison, J. B. Haufler, M. G. Raphael, W. A. Wall, F. B. Samson (eds.). 2002. Predicting Species Occurrences: Issues of Accuracy and Scale. Island Press, Washington, D.C., 868 pp. ISBN 1-55963-787-0, price (hardcover), $95.00. In biodiversity conservation and management of natural resources, from policy decisions under the Convention for Biological Diversity to the implementation of national and local legislation, there is hardly a project where information on species occurrence is not needed. But animal and plant species are mobile, ranges shift over time, and individual animals move within their range and on a variety of scales. Representing complex ecological patterns such as species occurrences on a map inevitably presents a gross simplification of the real world, and models have always been the only tool available to researchers. The black blotches on the maps of many atlases of species distribution are merely the earliest and simplest models, worked out by experts on the basis of available data and their personal experience. In spite of their often-rough approximation, much conservation and management continues to be based on such simple models, which remain the only available means for representing the distribution of many species at a global scale. Species—habitat relationships were a great improvement for exploring the internal composition of species distributions, and they launched the new field of predicting species occurrences. The widespread application of computers offered a vast array of new tools, such as GIS and statistical packages and their ever-increasing ability to crunch large quantities of data, and remote sensing, which facilitated the collection of large amounts of environmental data. Ecologists did not miss these opportunities. However, we are still far from having solved all problems of predicting species occurrences. Spatial and temporal variations in the processes determining species occurrence still pose enormous challenges to modelers, and it is possible that accurate modeling will never be satisfactorily accomplished. In spite of the well-known axiom that ecological patterns should be studied at a variety of scales (Levin 1992), modelers often struggle for the magic model …