Pierce, Maria E. (2022) Integrating Knowledge about Complex Adaptive Systems - Insights from Modeling the Eastern Baltic Cod. PhD thesis, Institute for Visual and Analytic Computing, University of Rostock.
Full text not available from this repository.Abstract
Currently, the Eastern Baltic cod (EBC) is in continuing decline regarding both the productivity of the stock and the condition of its individuals. Supporting management efforts to assist the stock in its recovery will require a functional understanding of the new dynamics of the EBC stock and the Balticecosystem. However, aquatic environments are challenging to research because they are not directly accessible to humans, encompass many scientific disciplines, and are complex adaptive systems (CAS). This thesis explores how modelling and simulation methods can be applied and adapted to meet the specific needs of fisheries biologies’ current challenges regarding the management of the EBC and potentially other stocks facing similar challenges. To effectively incorporate modelling and simulation into the workflow of departmental research, specific key requirements and matching solutions were identified: functionally integrating knowledge across scientific domains, which can be achieved through multi-faceted modelling; providing the required accessibility, which can be supported by suitable domain-specific languages (DSLs) and thorough documentation; modelling CAS to a high degree of verisimilitude of functionality, which requires a careful selection of type of model based on the respective system of interest and research question; and continued validity, which can be streamlined by iterative validation. As the general approach, multi-faceted modelling was adapted to the given requirements by using white-box integration of submodels to provide insight into mechanisms and interaction of processes in the modelled system. These models were specified in the DSL ML-Rules, a language for multi-level modelling and simulation of cell-biological systems, which was previously tested for its suitability for ecological models. The systematic reuse of simulation experiments during the iterative integration of models helped to maintain the continued validity of the model. Additionally, a focus was placed on documentation to further ensure accessibility and invite scrutinisation of the simulation study. Both structured natural language documentation using the TRACE protocol and formal provenance documentation based on the provenance data model (ProvDM) were applied here. Provenance metadata was then used to link the two documentations, which provides increased navigability by supporting queries of the provenance graph. The simulation study currently covers the facets of physiology, reproduction, behaviour, environment, prey and parasitation and iteratively worked towards understanding the decline in growth and condition. Although the processes formalised to this point have not uncovered the mechanisms of the decline, several suspected causes could be dismissed. Notwithstanding the domain results, the study illustrates the merits and viability of the multi-faceted approach for the functional integration of knowledge about CAS.
| Item Type: | Thesis (PhD) |
|---|---|
| Projects: | Agent-based modelling of Baltic cod |
