Abstract
CONTEXT: Theoretical and methodological developments in the field of fragmentation-biodiversity research continue to rely on the central concept of the habitat patch where patch size and number are considered particularly relevant to spatially structured ecological communities. However, although great interest has been shown in the effects of habitat fragmentation, appropriate methods for the spatial delineation of habitat have not received equal attention. In this paper, we argue that existing methods are not consistent with a functional definition of habitat as they fail to address key methodological challenges. These relate to the need to acknowledge a) the contribution of multiple resource types to habitat, b) the influence of neighbouring land cover types and c) the continuity-contiguity problem (the tendency of habitat to exhibit properties of gradation and aggregation). OBJECTIVES: In this second of two papers on this topic, we present an application of a new methodological framework outlined by Dennis et al. (this issue) that offers a route to a more functional definition and delineation of habitat through the use of spatial kernels and the generation of Type 1 and 2 fuzzy sets from landscape classification algorithms. We present a demonstration of the framework applied to a real-world landscape, in which we illustrate the impact of adopting alternative perspectives with respect to habitat delineation on the ecological process of habitat connectivity. METHODS: We demonstrate the functional delineation of habitat for a focal generic woodland species (FGWS) in a real-world landscape classified through the application of a fuzzy Random Forest classifier. We employ nesting, foraging and dispersal parameters relevant to the FGWS to achieve a functional estimate of habitat. We test the influence of habitat fragmentation (number of patches controlling for total habitat amount) on potential functional connectivity for the FGWS based on contiguous (emphasising aggregation and homogeneity), continuous (emphasizing gradation) and functional (integrating multiple resource types and neighbourhood effects) habitat perspectives. RESULTS: Our results indicate large discrepancies between the three perspectives on habitat delineation across key fragmentation-relevant metrics (total area, number of patches and potential functional connectivity). Importantly, a functional habitat perspective supports markedly different conclusions (compared to contiguous and continuous perspectives) with respect to the relationship between fragmentation (number of patches) and connectivity, and estimates of the contribution of individual habitat patches to landscape-scale connectivity. CONCLUSION: The functional habitat perspective, operationalized by harnessing uncertainty in landscape classification and employing spatial kernels to parameterise neighbourhood effects based on species-specific parameters, achieves a functional delineation of habitat. Our study suggests that such a view has major implications for our understanding of habitat fragmentation because it requires that the latter also be assigned a functional definition. The framework centred on functional habitat delineation is generalizable to a wide range of landscape contexts and advances current methods in spatial ecology. This opens up opportunities for inquiry and the development of new theoretical positions within the fragmentation-biodiversity debate. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1007/s10980-025-02290-y.