The need for validation of ecological indices

Abstract Increased recognition of the need for ecosystem-based management has resulted in a growing body of research on the use of indicators to represent and track ecosystem status, particularly in marine environments. While multiple frameworks have been developed for selecting and evaluating indicators, certain types of indicators require additional consideration and validation. In particular, an index, which we define as an aggregation of two or more indicators, may have unique properties and behaviors that can make interpretation difficult, particularly in a management context. We assert that more rigorous validation and testing is required for indices, particularly those used to inform management decisions. To support this point we demonstrate the need for validation and then explore current development and validation processes for ecosystem indices. We also compare how other disciplines (e.g., medicine, economics) validate indices. Validating indices (and indicators) is particularly challenging because they are often developed without an explicit objective in mind. We suggest that exploring the sensitivity of an index to the assumptions made during its development be a pre-requisite to employing such an index.