Integrating Edge Detection and Dynamic Modeling in Quantitative Analyses of Ecological Boundaries

D variously in terms of edges, ecotones, borders, and boundaries, the generalized concept of ecological boundaries has received considerable research attention over the past decade (e.g., Chen et al. 1992, Hansen et al. 1992, Gosz 1993, Risser 1995, Saunders et al. 1999). For the purposes of this article, we define boundary as a zone between contrasting habitat patches that delimits the spatial heterogeneity of a landscape (Cadenasso et al. 2003). Boundaries are specific landscape features that can possess diverse structural properties (e.g., open or closed, crisp and sharp or gradual and fuzzy, straight or sinuous; see Strayer et al. 2003). We use the term edge to refer to a sharp boundary, whereas we use the term boundary to encompass both sharp boundaries (edges) and those with more gradual structure. Increasingly, scientists’ attention has turned from describing boundaries to investigating their functional importance (Didham et al. 1996, Fagan et al. 1999, Laurance et al. 2001), as when edges influence material flows across a landscape (Wiens 1992), provide or limit wildlife habitat (Woodroffe and Ginsberg 1998), or serve as a barrier to movement (e.g., Bider 1968). Human activities strongly influence the extent and type of boundaries found on Earth. These anthropogenic impacts are increasingly being seen in a dynamic context, and this new perspective has led ecologists to at least two key insights. First, boundaries have both local and long-distance influences, affecting the structure of adjacent habitats and the processes operating in these habitats and in habitats remote from the boundaries themselves (Laurance et al. 1997, Woodroffe and Ginsberg 1998, Saunders et al. 1999, Cadenasso and Pickett 2000). Second, boundaries are themselves dynamic entities, with the potential to vary both spatially and temporally (Wiens et al. 1985, Williams-Linera 1990, Forman 1995), as when the structure of boundaries changes during succession (e.g., Matlack 1994). Given the complexities inherent in this dynamic view of ecological boundaries, quantitative frameworks can significantly enhance our understanding of boundary structures, functions, and consequences. Historically, studies of ecological boundaries have typically adopted either pattern-based or process-based perspectives, though the interrelationships between structure and function