干旱区生态环境知识资源中心

The increasing relevance and importance of the subject of landscape ecology to biogeography are introduced. Research into landscape and plant community boundaries, nevertheless, remains comparatively neglected. In particular, the nature of those boundaries in terms of the patterns of floristic change and related ecosystem properties constitutes a potentially significant new area of research for biogeographers. The term ’ecotone’ has traditionally been used to describe boundaries between plant communities and ecosystems at a range of scales. Various definitions are presented and the often confusing terminology surrounding the word ’ecotone’ is reviewed. Boundary types range from sharp, clearly defined boundaries (ecotones) between more highly modified plant communities and anthropogenically created land-use types at one extreme, to more gradual and diffuse boundaries (ecoclines) between natural and semi-natural plant communities at the other. It is proposed that the term ’transitional area’ is used to describe all types of vegetation boundary when working at the local/community scale.

There is little literature of direct significance to the subject of transitional areas. The concept can only be meaningfully discussed in the context of recent developments in the conceptualization of the plant community and these are summarized. The importance of mosaics within plant communities is described and the need to understand and recognize mosaics when studying transitional areas between plant communities is emphasized.

The range of research methods available to describe and analyse variations in patterns of floristics and associated environmental variables across transitional areas is then critically reviewed. The potential relevance of remote sensing and geographical information systems, network analysis and fractals is demonstrated. Ideas on possible adaptations of sampling strategies for the description of floristics and environmental/biotic factors to cater for boundary/transitional area situations are presented and the concept of a rectangular sampling area as an alternative to the more normal linear transect is introduced. The traditional approach to the description of vegetation change across boundaries using transects and similarity coefficients has now been superseded by new developments, notably moving-window analysis, the Mantel test, pattern analysis, semi-variograms, spectral analysis and analysis for spatial autocorrelation, and the scope of these methods is summarized.