Quantitative Evaluation of Marine Ecosystem Indicator Performance Using Food Web Models

Successful ecosystem-based management requires the selection and use of informative indicators of ecosystem status. We analyzed seven marine food web models to evaluate the performance of candidate indicators of ecosystem structure and function. The basic approach involved simulating fishing perturbations to each model, measuring the response of ecosystem attributes and candidate indicators to the perturbations, and testing the ability of the indicators to track changes in the values of the attributes. We focused on 22 ecosystem attributes, encompassing structural and functional properties that are relevant to a number of stakeholder groups but are typically difficult to measure directly (for example, food web structure, energy recycling). We tested for correlations between the attributes and 27 empirically tractable candidate indicators (for example, foraging guild biomasses, ratios of community-level groups) within each of the models and quantified consistency in indicator performance across the models. Our analysis suggests that no single indicator is sufficient to describe all of the ecosystem attributes, but at the same time highlights broad, catch-all indicators (for example, detritivores, jellyfish) and distinguishes the strongest attribute-indicator relationships. Ecosystem indicators consisting of lower-trophic level, higher-productivity functional groups tended to perform particularly well. We also identified indicators that showed strong or weak associations with different attributes, but together captured changes in nearly all of them. Examples of such complementary indicators include phytoplankton, zooplanktivorous fish, piscivorous fish, and trophic level of the catch. Quantitative approaches such as this one will enable managers to make informed decisions about ecosystem-scale monitoring in the oceans.