Integration of indicator outcomes for assessing good environmental status: A model-based approach

Integrated assessments are required to track ecosystem status, any change thereof and progress due to conservation measures. The European Union Marine Strategy Framework Directive (MSFD), which aims among others at maintaining biodiversity, requires regular quantitative assessments built from monitoring data and indicator developments to decide on the environmental status of each element in a marine ecosystem. Beyond the daunting tasks incurred with large-scale monitoring and timely analyses, collapsing the multivariate information from indicators on several species into a single number reflecting an overall environmental status is the challenge associated with integration. The ‘One Out, All Out’ (OOAO) integration rule is currently in use in part due its simplicity and ease of use. Its main drawback is its built-in sensitivity to error in binary assessment outcomes: statistical noise will dominate integration with an increasing number of species or indicators. We developed a model-based approach to integration for MSFD assessments within the statistical framework of Signal Detection Theory. We benchmarked the model performance against the OOAO integration rule in a simulation study. The model-based integration scheme redresses known deficiencies with the OOAO rule and clarifies statistical assumptions behind integration rules, unifying them across a continuum. We illustrated our approach with a case study on the ecosystem component ’marine mammals’ in the Northeast Atlantic. We believe our model-based approach provides a more comprehensive integrated output of ecosystem assessments that aligns better with decision-making under uncertainty in evidence-based policy. • Assessments under ecosystem-based approaches use indicators on a panel of species. • Multivariate assessment outcomes must be collapsed into a single, integrated value. • One-Out, All-Out (OOAO) is an overly conservative integration rule. • A model-based approach for integration is developed with a marine mammal case study. • Model-based integration gives a more nuanced assessment of Good Environmental Status.