Skip to main content
[Translate to English:]
[Translate to English:]
Institute of

OF Baltic Sea Fisheries


Agent-based modelling of Baltic cod

[Translate to English:] Abgemagerter Dorsch aus der östlichen Ostsee.
© D. Stepputtis
[Translate to English:] Abgemagerter Dorsch aus der östlichen Ostsee.

Multi-Level rule-based modelling of eastern Baltic Cod bioenergetics and species-habitat interactions

Despite extensive research in the past ICES scientist are currently not able to conclusively explain the ecological interrelationships which drive the current changes in the eastern Baltic cod stock. Equation-based macro-models do not deliver useful insights since they are not able to adequately depict the underlying physiological processes of cod. Therefore the aim of this project is the development of an agent-based multi-level model of the bioenergetics and behaviour of cod as well as of the hydrography of the Bornholm-Basin and the relevant aspects of its prey. Ultimately, the model developed in this project shall serve scientists from the fields of fishery biology and fish physiology as a simulation tool for academic research.

Background and Objective

Established equation-based macro-models currently do not deliver reliable predictions on stock biomass and on the state of the stock. Hypotheses on possible mechanisms which may explain, e.g. the increased occurrence of “meager-cod” (food shortage, hypoxia effects) and the lack of larger cod (fisheries, increase of natural mortality) can currently not be tested. At the time when changes in the stock occurred (approximately since 2010) no relevant experiments where conducted. The environmental conditions at that time cannot be replicated and the cod of that time have since died or are no longer in the condition they were then.

Since biological samples of ecological systems cannot be collected retrospectively, ecological hypotheses which pertain to the past can only be investigated with the aid of computer models. Such models however have to adequately and flexibly represent the dynamics of complex ecological systems at different scales, and yet remain clearly readable at the same time. The modelling language ML-Rules (originally developed for the modelling of cell-biological systems) appears promising in meeting these high demands and thus shall be used as a basis for our research.



The model will be developed as a set of submodels which depict separate aspects of the ecological system. Here, among others, habitat and behaviour of cod as well as its prey and bioenergetics in the Baltic Sea environment will be addressed. Due to the rule-based approach, each submodel can be developed and validated individually and the complete model can subsequently be used in its entirety.

  • Notably, the bioenergetics submodel, but also the other submodels, shall be designed and documented in such a manner that they can be used as research tools by respective domain scientists. It shall be checked if the submodels can be practically transferred to different systems (for example in case of the bioenergetics submodel to a different stock or species of fish).
  • Simulation experiment specification templates in SESSL (Simulation Experiment Specification via a Scala Layer) will be defined and stored to support the experimentation with the model, whether this is a calibration to new situations, or scans or calculating “what-if” scenarios.
  • Finally, a manual shall be developed which allows full understanding, self-sufficient augmentation, and further development as well as full use of the model without any assistance and thus make it versatile in use.


At the end of the project, an overall model consisting of functionally integrated sub-models is now available. The submodels include the facets of physiology, reproduction, behaviour, environment, prey and parasitation. The overall model thus allows the detailed investigation of the mechanisms already implemented, both in the form of individual facets and in their interaction and overall effect.  Based on the extensive and structured documentation, mechanisms can also be re-implemented according to current hypotheses and integrated into the existing model.

So far, the drivers oxygen deficiency, changed and reduced prey quality and increasing parasitation have been examined with the model, as well as their cumulative effects, also in connection with abiotic conditions in the context of annual migrations. The results demonstrate that these are not the sole drivers of the current development.

In terms of methodology, the project brought further developments in the area of comprehensive modelling of complex adaptive systems (CAS). In particular, the extension of the 'multi-faceted' modelling approach to include a functional white-box integration of different facets of individual processes is a promising tool for transparent and far-reaching research into this type of question. Through the systematic reuse of simulation experiments, the validation of already successfully implemented mechanisms could be maintained during the ongoing integration of new submodels with little effort.

The accompanying manual has been implemented according to the TRACE protocol, which is considered state-of-the-art for the documentation of ecological models. It is published in a Git repository together with the model, the various simulation experiments and example results.

Links and Downloads

Involved external Thünen-Partners

  • Universität Rostock
    (Rostock, Deutschland)


8.2016 - 12.2022

More Information

Project status: finished


  1. 0

    Pierce ME (2023) Integrating knowledge about complex adaptive systems - insights from modelling the Eastern Baltic cod. Rostock: Univ Rostock, 151 p, Rostock, Univ, Fak f Informatik und Elektrotechnik, Diss, 2022, DOI:10.18453/rosdok_id00004121

  2. 1

    Funk S, Funk N, Herrmann JP, Hinrichsen HH, Krumme U, Möllmann C, Temming A (2023) Tracing growth patterns in cod (Gadus morhua L.) using bioenergetic modelling. Ecol Evol 13(11):e10751, DOI:10.1002/ece3.10751

  3. 2

    Pierce ME, Krumme U, Uhrmacher AM (2018) Building simulation models of complex ecological systems by successive composition and reusing simulation experiments. In: Rabe M, Juan AA, Mustafee N, Skoogh A, Jain S, Johansson B (eds) Proceedings of the 2018 Winter Simulation Conference (WSC 2018), 9-12 Dec 2018, Göteborg, Sweden. pp 2363-2374, DOI:10.1109/WSC.2018.8632262

  4. 3

    Pierce ME, Warnke T, Krumme U, Helms T, Hammer C, Uhrmacher AM (2017) Developing and validating a multi-level ecological model of eastern Baltic cod (Gadus morhua) in the Bornholm Basin - a case for domain-specific languages. Ecol Model 361:49-65, DOI:10.1016/j.ecolmodel.2017.07.012

  5. 4

    Pierce ME, Warnke T, Helms T, Uhrmacher AM, Krumme U, Hammer C (2015) Individual-based cod simulation with ML-Rules. In: Yilmaz L (ed) Proceedings of the 2015 Winter Simulation Conference, December 6 - 9, 2015, Huntington Beach, CA. Piscataway, NJ: Omnipress, p 2

                                                                                                          zurück zum Arbeitsbereich

Scroll to top