Factors limiting our understanding of ecological scale

Multi-scale studies ostensibly allow us to form generalizations regarding the importance of scale in understanding ecosystem function, and in the application of the same ecological principles across a series of spatial domains. Achieving such generalizations, however, requires consistency among multi-scale studies not only in across-scale sample design, but also in basic rationales used in the choice of observational scale, including both grain and extent. To examine the current state of this science, here we review 79 multi-scale wildlife-habitat studies published since 1993. We summarize rationales used in scale choice and also review key differences in scale-specific experimental design among studies. We found on average that 70% of the observational scales employed in wildlife-habitat research were chosen arbitrarily with no biological connection to the system of study, and with no consideration regarding domains of scale for either dependent or independent variables. Further, we found it common to change either both grain and extent, or the entire suite of independent variables across scales, making cross-scale extrapolations and generalizations impossible. We discuss these sampling limitations by clarifying the differences between multi-scale versus multi-design studies, including the distinction between spatial versus scalar observations, and how these may differ from the commonly cited “orders of resource selection”. We conclude by reviewing both existing and suggested alternatives to reduce the arbitrary nature of observational-scale choice prevalent in today's literature.