Developing and validating a multi-level ecological model of eastern Baltic cod (Gadus morhua) in the Bornholm Basin - A case for domain-specific languages

Recent changes in environmental and ecological conditions are a major reason that established stock assessment macro-models no longer provide reliable estimates on the stock status of eastern Baltic cod (Gadus morhua). To account for a complex ecological system, an individual-based modelling approach is needed that turns the attention from the overall population to the individual fish, its metabolic processes, behaviour, and interaction with environmental conditions.The multi-level rule-based modelling language ML-Rules was used to develop the first iteration of an individual-based model. The different stratification layers of the estuarine environment are modelled as explicit entities with varying properties and are populated by individual cod and its prey. The cod model focuses on growth and mortality dependent on abiotic factors and behaviour within the stratified environment of the Bornholm Basin.Despite the complexity of the formalised interactions, the resulting model specification is succinct and compact. Its structure and declarative nature facilitates reuse and subsequent extension. To allow easy replication of simulation experiments, we used the simulation experiment specification language SESSL for calibration and validation. The extendibility of the model and replicability of simulation experiments, as provided by the domain specific languages ML-Rules and SESSL respectively, will be key assets in future research, as the model provides a first, but essential, step toward studying the impact of specific environmental changes and certain behaviour pattern on the cod population in the Bornholm Basin.