Statistical analysis of ecological communities: progress, status, and future directions

Abstract

In the first half of the twentieth century, plant ecology was primarily a descriptive science. The traditional methods used to classify plant communities involved subjective decisions that made them non-repeatable. Although a few earlier papers had experimented with more formal statistical techniques, it was not until the 1950s that work by several research groups and individuals laid the groundwork for a quantitative revolution in the analysis of vegetation. A major contributor was David W. Goodall, whose series of papers under the main title of ‘‘Objective Methods for the Classification of Vegetation’’ pioneered the use of positive correlation among species to classify vegetation into homogeneous groups (Goodall 1953a), showed how fidelity (the degree to which a species is a good indicator of a proposed vegetation type) could be quantified (Goodall 1953b), demonstrated the use of discriminant functions to allocate new vegetation samples to previously defined community groups (Goodall 1953b), coined the term ‘‘ordination’’ for techniques that place vegetation samples in a coordinate space rather than dividing them into groups and first applied principal components analysis to ordinate vegetation data (Goodall 1954), and performed one of the first analyses of spatial variation in plant community composition (Goodall 1961). In addition to these, David Goodall wrote numerous other insightful reviews and research papers on the sampling, description and statistical analysis of plant communities that were enormously influential in the development of what came to be known as ‘‘quantitative ecology’’. Among his other contributions were clarification of the issues involved in the ‘‘community vs individualistic continuum’’ controversy (Goodall 1963), plotless methods for testing interspecific association (Goodall 1965), methods for detecting outliers or deviant observations (Goodall 1966a, 1969), the development of a probabilistic similarity index (Goodall 1966b), a technique for the analysis of spatial pattern in vegetation (Goodall 1974), evaluation of methods for testing dispersion patterns (Goodall and West 1979), and the development of a ‘‘model-based’’ ordination method (Johnson and Goodall 1980; Goodall and Johnson 1982, 1987). This special issue of Plant Ecology honors the work of David Goodall, who turned 100 on 4 April, 2014 and is still active as an honorary research associate at the Centre for Ecosystem Management at Edith Cowan University, Australia. We invited a range of Communicated by Neal Enright.