This report is a deliverable of Work Package 7 (WP7 – Synthesis & training development) of the FP7 MareFrame research project. The report comprises two main parts. The first part presents a protocol for the comparison and evaluation of the ecosystem models that have been developed and applied within the MareFrame project. This is to inform on their general suitability to provide support to an Ecosystem Approach to Fisheries Management (EAFM) (task 7.2.1). The second part of the report reviews and evaluates the progress on the ecosystem models developed for the various case studies (task 7.2.2). This work is based on application of the protocol developed in the first part, with the addition of a procedure to assess progress in the application of the models. The timing of this deliverable (M36) was specified to provide a suitable checkpoint to assess status, as well as to provide specific recommendations on corrective measures where needed, so as to ensure that the ecosystem modelling- related objectives of the MareFrame project will be achieved.
This report should be viewed in the context of the subsequent report D7.5, which will evaluate the Decision Support Framework (DSF) developed within the project and applied in the various case studies. Within the MareFrame project, use of the term DSF encompasses the ecosystem models, decision tools and co-creation process developed to facilitate progress towards an EAFM. The scope of this report is limited to address the comparison and use of the ecosystem models that comprise the empirical basis for the DSF, and hence considers only scenarios relevant for the evaluation of management alternatives. Furthermore, this report is tightly linked to activities in other parts of the project. In particular, procedures for model comparison will be developed in concert with activities in WP4 (especially with deliverable D4.2), which has established common reporting procedures for model outputs. The outcomes of the comparative analysis will inform WP5 and WP6 in particular.
Ecosystem models show considerable variability in their output for all the case studies, which is a consequence of the high structural uncertainty inherent in ecosystem models. Some of these differences arise from the range in scope (from tactical to strategic) covered by the different models examined, as well as from the different extents to which they focused on securing good fits to the data available. In general, comparative approaches are recommended as the way forward, both to quantify structural uncertainty and to find results which are robust to model formulation.
All of the MareFrame case studies adopted the co-creation approach. This led to confirmation of the high potential which ecosystem models possess to highlight the trade-offs to which fisheries management needs to give attention. However, in several cases the approach also served to emphasise the limits of the current models and the difficulties in implementing them to address some of the specific ‘co-created management objectives‘.
A general feature of the ecosystem models considered is that increased model complexity comes at the expense of precision and ability to fit available data. The inclusion of more species, trophic layers and processes often requires more assumptions, readily finds itself compromised by paucity of data, and can lead to difficulties in achieving statistically appropriate fits to data. Nevertheless, the management questions posed, as well as the development of the associated decision support tools, for the different case studies were found to require models which addressed certain aspects of this increased ecosystem complexity. These aspects go far beyond what traditionally needs to be considered for single species stock assessments.
Regardless of the fact that a management question may require the provision of analyses to inform a short term tactical decision, it is the long-term implications of that decision for the ecosystem are of most interest in an EAFM context. In several of the case studies, rather than indications of which model outperformed the others, what emerged was the need for complementarity. Management of fisheries requires explicit recognition of the complexity of individual fish populations in terms of their abundance and demographic structure, but this does impose strong limitations in the context of an EAFM unless this is limited to a handful of the most important targeted species.
There is a particularly strong need for methodologies which synthesise the considerable quantity of outputs generated from multiple ecosystem models. This is especially the case in a framework which involves close interactions with stakeholders, as applies in all the MareFrame case studies. The DST has an important role to play in this. In the spirit underlying MareFrame, the DST should be able to incorporate output from multiple models as well as take account of some other sources of uncertainty, such as arise from stochastic aspects and from environmental variability. Caution is recommended in the combination of multi-model outputs into integrated statistics, and simple model averaging should not replace an in-depth understanding of these uncertainties so that they can be accorded appropriate relative weightings in advising decisions.
More info D7.3 MAREFRAME_Simulation based training tool for a DSS for EAFM D7.2 MAREFRAME_Analysed case studies with respect to project progress evaluation.