Journal cover Journal topic
Geoscientific Model Development An interactive open-access journal of the European Geosciences Union
Geosci. Model Dev., 9, 947-964, 2016
https://doi.org/10.5194/gmd-9-947-2016
© Author(s) 2016. This work is distributed under
the Creative Commons Attribution 3.0 License.
Development and technical paper
04 Mar 2016
Couplerlib: a metadata-driven library for the integration of multiple models of higher and lower trophic level marine systems with inexact functional group matching
Jonathan Beecham1, Jorn Bruggeman2, John Aldridge1, and Steven Mackinson1 1Cefas Lowestoft Laboratory, Pakefield Road, Lowestoft, Suffolk, NR33 0HT, UK
2Bolding & Burchard ApS. Strandgyden 25, 5466 Asperup, Denmark
Abstract. End-to-end modelling is a rapidly developing strategy for modelling in marine systems science and management. However, problems remain in the area of data matching and sub-model compatibility. A mechanism and novel interfacing system (Couplerlib) is presented whereby a physical–biogeochemical model (General Ocean Turbulence Model–European Regional Seas Ecosystem Model, GOTM–ERSEM) that predicts dynamics of the lower trophic level (LTL) organisms in marine ecosystems is coupled to a dynamic ecosystem model (Ecosim), which predicts food-web interactions among higher trophic level (HTL) organisms. Coupling is achieved by means of a bespoke interface, which handles the system incompatibilities between the models and a more generic Couplerlib library, which uses metadata descriptions in extensible mark-up language (XML) to marshal data between groups, paying attention to functional group mappings and compatibility of units between models. In addition, within Couplerlib, models can be coupled across networks by means of socket mechanisms.

As a demonstration of this approach, a food-web model (Ecopath with Ecosim, EwE) and a physical–biogeochemical model (GOTM–ERSEM) representing the North Sea ecosystem were joined with Couplerlib. The output from GOTM–ERSEM varies between years, depending on oceanographic and meteorological conditions. Although inter-annual variability was clearly present, there was always the tendency for an annual cycle consisting of a peak of diatoms in spring, followed by (less nutritious) flagellates and dinoflagellates through the summer, resulting in an early summer peak in the mesozooplankton biomass. Pelagic productivity, predicted by the LTL model, was highly seasonal with little winter food for the higher trophic levels. The Ecosim model was originally based on the assumption of constant annual inputs of energy and, consequently, when coupled, pelagic species suffered population losses over the winter months. By contrast, benthic populations were more stable (although the benthic linkage modelled was purely at the detritus level, so this stability reflects the stability of the Ecosim model). The coupled model was used to examine long-term effects of environmental change, and showed the system to be nutrient limited and relatively unaffected by forecast climate change, especially in the benthos. The stability of an Ecosim formulation for large higher tropic level food webs is discussed and it is concluded that this kind of coupled model formulation is better for examining the effects of long-term environmental change than short-term perturbations.


Citation: Beecham, J., Bruggeman, J., Aldridge, J., and Mackinson, S.: Couplerlib: a metadata-driven library for the integration of multiple models of higher and lower trophic level marine systems with inexact functional group matching, Geosci. Model Dev., 9, 947-964, https://doi.org/10.5194/gmd-9-947-2016, 2016.
Publications Copernicus
Download
Short summary
This paper is a description of how very different higher and lower trophic level models (Ecopath with Ecosim) and ERSEM, respectively, can be coupled together using a metadata coupling system together with a number of examples of short- and long-range projections for end to end modelling.
This paper is a description of how very different higher and lower trophic level models (Ecopath...
Share