Begging the question: Spatial Pattern Analysis in Plant Ecology, by M.R.T. Dale

Abstract

Cambridge University Press, Cambridge Studies in Ecology, 1999. £45.00/$69.95 hbk (x + 326 pages)ISBN 0 521 45227 9I approached this book with some interest as an opportunity to learn what progress had been made since I had last taken a serious interest in spatial pattern in plant communities as a PhD student1xAustin, M.P. J. Ecol. 1968; 56: 197–218CrossrefSee all References. Following the ideas of A.S. Watt2xWatt, A.S. J. Ecol. 1947; 35: 1–22CrossrefSee all References on pattern, Greig-Smith3xGreig-Smith, P. J. Ecol. 1979; 67: 755–779CrossrefSee all References had pioneered the quantitative analysis of spatial pattern in plant communities using block size analysis of variance. Kershaw4xQuantitative and Dynamic Ecology. Kershaw, K.A. and Looney, J.H.H. See all References then extended the approach to covariance analysis. Strong nonrandom pattern was revealed in many apparently homogeneous grassland communities. The two questions at the time were: did pattern decline from early to late succession and what scales of pattern were there and why? It was soon clear that nonrandom patterns persisted during succession. Three scales of pattern were often found reflecting the ‘fairy ring’ growth of the dominant species and its competitive impact on subordinate species. These observations were consistent with Watt’s description of pioneer, building, mature and senescent phases of growth in patch forming plants. My own research confirmed these results but also convinced me that much of the variation I observed was due to the rather trivial fact that transects intersect circular patches at different angles.How have things changed? The book by Dale has a strong methodological emphasis. There are numerous new methods with odd names and abbreviations, such as three-term local quadrat variance (3TLQV). There are methods for one- or two-dimensional analysis, species covariance analysis and multispecies pattern analysis. There is recognition that the size and pattern of gaps need to be distinguished from that of patches. The difficulty of providing unambiguous interpretations of the peaks in the graphs of variance block size remains. The author recommends using more than one method on each data set. New relevant topics are raised. How do we test whether species limits along environmental gradients tend to co-occur? An example of this is whether species zonation occurs in the intertidal zone. Another important issue is how we separate spatial and environmental influences on the distribution of species5xBorcard, D., Legendre, P., and Drapeau, P. Ecology. 1992; 73: 1045–1055CrossrefSee all References. Issues are raised, technical details of possible methods are given and examples of calculations provided, but rarely is there any follow through. The progress in ecological understanding achieved by these methods is ignored. One is left with the question has there been any progress?The phenomenon of vegetation groves in arid landscapes is illustrated in the first chapter on ‘concepts of spatial pattern’. The ‘broussee tigree’ in Niger or the Mulga groves in Australia are obvious extreme examples of spatial organization in plant communities. Are there other less obvious examples of maximizing water-use through spatial patterning waiting to be detected using quantitative pattern-analysis? Do these patterns tell us something about the organization of ecosystems in relation to climate? The author quotes data and results from a study of pattern based on a chronosequence of moraines, but there is hardly any discussion of the results in terms of the role of nucleation in succession6xYarranton, G.A. and Morrison, R.G. J. Ecol. 1974; 62: 417–427CrossrefSee all References or their relevance to any theory of primary succession. Shipley and Keddy7xShipley, B. and Keddy, P.A. Vegetatio. 1987; 69: 47–55Crossref | Scopus (102)See all References are mentioned for using species limits to test differences between the community and continuum concepts of vegetation organization. A whole series of tests are outlined for addressing such theoretical questions.Will the researcher interested in these questions realize this book might have some useful methods for him? Will the student realize the relevance of the methods to these questions? The author lists some ecological questions at the end of the last chapter. They seem something of an afterthought. Some of the questions would be better answered by experiment than spatial pattern analysis. The author has concentrated on the methodology to the exclusion of the plant ecology.As a result of reading this book, I think spatial pattern analysis has become a methodology in search of questions. New ecologically relevant questions need to be found. This book is for the specialist analyst who already has a question or the enthusiast for block size analysis of variance.