Vorhersagen der räumlichen Verteilung mariner Arten und Identifizierung prioritärer Schutzgebiete – Vergleich und Kopplung von Nahrungsnetzmodellen mit hierarchischen Bayes-Modellen
Zusammenfassung der Projektergebnisse
A prominent tool to mitigate anthropogenic pressures and to protect marine biodiversity and ecosystem functioning is the implementation of marine protected areas. For a successful identification and prioritization of suitable conservation areas, it is crucial to understand the spatial distribution of biodiversity, target species and essential habitats. Within this project, we tested and developed approaches for the parameterization of the spatial-temporal capabilities of Ecopath with Ecosim (EwE) to predict marine species distribution in the data-limited and small-scale bay system Chwaka Bay (Zanzibar, Tanzania) in the Western Indian Ocean. The central aim of the project was to investigate the complementarity and applicability of Bayesian hierarchical species distribution (B- HSD) modelling and the new Habitat Foraging Capacity (E-HFC) module of EwE for the prediction of species distribution and the identification of priority areas for conservation. As a first step, we tested six data-poor approaches for the parameterization of the predatorprey interactions (vulnerability parameterization) in EwE to simulate alternative gear regulation strategies in Chwaka Bay. The simulated changes in the biomass of functional groups and associated fishers’ profits – following gear regulations – varied among vulnerability settings, but the qualitative results (direction of change) were highly consistent. Our analysis illustrates that an unfitted EwE model can be informative for management strategy exploration. However, it also shows that a sensitivity analysis of the vulnerabilities should always be performed to better quantify parameter uncertainty and increase the robustness of simulations. We compared the spatial distribution of four Chwaka Bay target species predicted by (1) B-HSD models, (2) an E-HFC model informed by literature and general speciesenvironmental relationships from online databases (uninformed E-HFC model) and (2) an E-HFC model informed by the B-HSD models (informed E-HFC). Model predictions were evaluated using participatory maps developed with local fishers during a participatory workshop. We found that the species distributions predicted by the B-HSD models, the fishers, and the uninformed and informed E-HFC models varied strongly, illustrating high uncertainty associated with the model predictions. Overall, the species distribution predicted by the uninformed E-HFC model was less accurate, according to a ranking by fishers. This shows the difficulties to use the E-HFC model under limited information on species distribution drivers. Using the species distribution maps of the B-HSD model to inform the E-HFC model considerably improved the predictions of two of the four species, suggesting a way to better parameterize the E-HFC model. This work provides approaches that facilitate more robust use of the spatial-temporal capabilities of EwE in data-limited case studies.
Projektbezogene Publikationen (Auswahl)
-
(2021). Supporting Spatial Management of Data-Poor, Small-Scale Fisheries With a Bayesian Approach. Front. Mar. Sci. 8:621961
Rehren, Jennifer; Pennino, Maria Grazia; Coll, Marta; Jiddawi, Narriman & Muhando, Christopher
-
Making use of the spatial-temporal capabilities of EwE in small-scale, tropical fisheries settings – approaches to overcome data limitations. German Research Foundation Scholarship Report. 2021
Rehren, J., Coll, M., Pennino, M., Jiddawi, N.
