Landscape ecology and biogeography: Rethinking landscape metrics in a post-FRAGSTATS landscape

Landscape pattern indicators or 'metrics' provide simple measures of landscape structure that can be easily calculated with readily available data and software. Unfortunately, the ecological relevance of many metrics (i.e. the relationship between metric values and the real-world ecological processes that they are meant to serve as proxies for) is often unproven and questionable, and concerns are regularly voiced that such metrics fail to capture important aspects of landscape function. In this paper, I provide a review of landscape measures that may better link landscape pattern and function, ranging from approaches that extend existing metrics by incorporating a more functional component (e.g. core area measures, least cost distances) to those rooted in graph, network, and electrical circuit theory. While more 'functional' approaches are becoming increasingly popular, the selection of appropriate landscape metrics in many applications involves tradeoffs regarding data requirements, ease of calculation, functional basis, and simplicity of interpretation by a range of specialist and non-specialist stakeholders. Regardless, there continues to be a need for landscape metrics because they are seen by many land managers and stakeholders as simple, intuitive tools for assessing and monitoring changes in landscape pattern and, by extension, the effects on underlying ecological processes. Future needs include: (1) the development of more user-friendly landscape analysis software that can simplify graph-based analyses and visualization; and (2) studies that clarify the strengths and weaknesses of different approaches, including the potential limitations and biases in graph and network-based measures.