Domains of scale in forest-landscape metrics: Implications for species-habitat modeling

摘要

Observational scale defines the field-of-view used to quantify any set of data, and thus has profound implications on the development and interpretation of species-habitat models. However, most multiscale studies choose observational scales using criteria unrelated to how metrics quantify along the scale continuum: scale choice is either arbitrary or via orders of resource selection irrespective of potential among-scale differences in mean or variation. Here, I use GIS to examine these issues for 9 forest-landscape metrics across 15 observational extents (while holding grain constant) and demonstrate how associated averages and variation change markedly and unpredictably across observational scale continua, emphasizing that a priori selection of 2 "different" scales cannot be done intuitively, arbitrarily, or based on orders of resource selection. Evaluation of the scale-domain continuum is a critically absent step in the building and interpretation of all multi-scale ecological models.