The European regional seas ecosystem model, a complex marine ecosystem model

Abstract

This paper presents an overview of the concept, structure and implementation of the European Regional Seas Ecosystem Model (ERSEM). The model dynamically simulates the biogeochemical seasonal cycling of carbon, nitrogen, phosphorus and silicon in the pelagic and benthic food webs of the North Sea, and is forced by irradiance, temperature and transport processes.

The model has a coarse spatial resolution into ten boxes, the ICES boxes, of which the five deepest have been resolved into surface (0 to 30 m) and deep (30 m to bottom) boxes.

At the open boundaries, time series are prescribed for dissolved and particulate nutrients. River loads of nutrients for the rivers discharging into the North Sea are prescribed at monthly intervals. A general circulation model has been used to aggregate the exchange volumes across the box boundaries into daily in- and outflows. From these, the horizontal transports of dissolved and suspended constituents are calculated. Vertical transport is in the form of sinking and sedimentation for particulates and in the form of turbulent diffusion for dissolved constituents.

The physical model contains all information specific to the area to be modelled, whereas the biological/chemical submodels have been constructed not to be site-specific.

The biological variables are represented as functional groups expressed in units of organic carbon and the chemical variables as the internal pools in the biological variables and as the dissolved inorganic pools in water and sediment, expressed in units of N, P and Si.

The model runs in a software environment (SESAME) developed for enabling the development of large and complex models in a modular way by a consortium of institutes, each focusing on different. aspects of the ecosystem, translating these into modules within the model. With the exception of fish populations, where size- and age-structure are explicity represented, all the other biological components have been modelled as unstructured populations aggregated into functional groups. This approach is shown to be appropriate for taxa having short generation times in relation to the annual cycle and for taxa which do not span more than one trophic level during their lifetime.