The units of biodiversity

Biodiversity is a central concept in ecology and biology. Its underpinnings are multifaceted and complex and involve multiple spatiotemporal scales, and many ways of measuring relevant characteristics. Its comprehensive understanding requires a framework on which to organize concepts and associated metrics. The analysis of biodiversity is based on combinations of two types of units: study units (i.e., the inferential domain in time and space that characterizes sampling) and measurement units (i.e., metrics). We provide an integrated framework for the units of study derived from three aspects of organisms: their spatiotemporal relationships (geography), their evolutionary relationships (phylogeny), and their ecological relationships based on their requirements and effects (niche). We systematize the units of measurement based on four types of data (identity, abundance, phylogeny, traits), two properties of those data (magnitude and variability), and three approaches for their measurement (total, pairwise, nearest neighbor). Together, they define 14 basic elements that can be combined in many ways and be subject to various mathematical operations. The result is 130 different metrics, including those in the literature and those developed herein. We propose standardized symbols for these metrics and provide formulas using standard notations for their parameters. Importantly, we show how our framework can be used to align study units and measurement units with questions concerning the causes and consequences of biodiversity. We provide case studies on bats in Peru and trees in the eastern United States to ecological gradient theory, niche theory, and theory about relationships between biodiversity and productivity, and we discuss which metrics might be most appropriate in tests of island biogeography theory and the dilution effect of pathogen transmission. Our key recommendations are that researchers should: (1) harmonize study unit properties with explicitly defined questions, (2) couple metric properties with underlying processes, and (3) compare metrics with similar properties. By providing an overarching framework that clearly delineates units of study and units of measurement, we hope to ensure that appropriate data are applied to particular scientific questions, especially those of a comparative nature, thereby leading to robust conclusions of theoretical import or practical use in management or conservation.