Stop using dichotomous terms to reference observations of scale-dependent habitat selection

Abstract

Many studies dichotomize habitat selection into “local” or “landscape” effects, with little explanation regarding what each represents. Ambiguous use of these terms across studies may confound observations of selection operating across different spatiotemporal resolutions and impede synthesis of scale-dependent habitat selection. To examine the consistency of use of the terms “local” and “landscape” level habitat selection and evaluate potential implications for conservation science. We reviewed 136 multi-level studies observing local and/or landscape effects. From each study, we identified which of Johnson’s (Ecology 61:65–71, 1980) selection orders was observed at the local and landscape scale. We assessed consistency in the selection order observed at each scale. At the local scale, 54% of studies observed 3rd order selection and 38% of studies observed 4th order selection. At the landscape scale, 64% of studies observed 2nd order selection, while only 8% of studies observed 1st order selection. Overlap also occurred; 2nd order selection was observed at the local scale in 8% of studies, while 3rd and 4th order selection was observed at the landscape scale in 24 and 3% of studies, respectively. Selection orders observed at local and landscape scales are inconsistent across studies, making comparisons and synthesis difficult. This inconsistency makes the terms “local” and “landscape” meaningless, and may lead to misidentification of limiting factor(s) most important for conservation efforts. Dichotomous terms should no longer be used in reference to levels of selection, but in reference to specific explanatory variables whose characteristic scale(s) best fit the appropriate definition.