Skip to content

An Integrated “End-To-End” Model For Climate-Fish Interactions

May 18th, 2012


Key Findings

  • This work develops a modeling framework that fully couples a one-dimensional physical mixed layer model, a biogeochemical model, and an upper trophic level fisheries model for the pelagic ecosystem.
  • Through a series of sensitivity analyses, we identify critical steps for addressing the challenges of developing models that resolve interactions spanning physics to fish in an integrated way.
  • The resulting model architecture reproduces both the annual mean biomass of 23 different fish and plankton groups in the Northeast subarctic Pacific, and seasonal physical and biogeochemical dynamics.
  • This model is being used to explore the combined impacts of climate and fishing in the Northeast Pacific Ocean. The methodologies are generalizable and will contribute to the development of holistic ecosystem-based modeling frameworks in other regions, or for global applications.

Kelly Kearney, Charles Stock, Kerim Aydin, Jorge Sarmiento. Journal: Ecological Modeling. DOI:10.1016/j.ecolmodel.2012.04.006

Summary

Climate impacts on marine ecosystems arise from a combination of direct influences of physical climate on organisms (e.g., temperature effects on metabolic process) and indirect effects controlled by interactions with directly affected organisms. Indirect influences may originate with primary producers (i.e., phytoplankton) and propagate upward from the bottom of the food web or with higher trophic levels (i.e., fish) and propagate downward. Elucidating and predicting the response of living marine resources to climate and fishing pressure thus requires movement toward models that resolve interactions spanning physics to fish in an integrated way.

There are a number of challenges for constructing integrated physics to fish models. Measurements of fish feeding, biomass and growth parameters are subject to considerable uncertainty. Planktonic ecosystem dynamics are very sensitive to mortality rates that are altered when fisheries food webs are explicitly modeled. Differences in modeling approach between the planktonic ecosystem and fisheries communities exacerbate these issues by introducing subtle but important dynamical differences in model formulation that can lead to significant changes in results. This is especially true for long simulations required for climate-ecosystem applications. Such difficulties have led most efforts to adopt a “one-way” coupling where physical and planktonic ecosystem simulations inform fisheries models, but feedbacks from fish to plankton are not resolved. Recent observations from heavily fished regions suggest that these latter feedbacks can be significant at continental-shelf scales.

This work develops a modeling framework that fully couples a one-dimensional physical mixed layer model, a biogeochemical model, and an upper trophic level fisheries model for the pelagic ecosystem. It clearly identifies critical steps for addressing the challenges described above through a series of sensitivity analyses. These steps include

  1. An analytical approach for identifying the primary food web linkages.
  2. Methodologies for bridging the vertically resolved perspective adopted for planktonic ecosystems and the vertically averaged perspective adopted in fisheries food web models.
  3. Methodologies for ensuring consistency in the representation of trophic interactions across fisheries and plankton group.
  4. An exploration of the impacts of non-predatory mortality on model behavior.

The resulting final model architecture is able to reproduce both the annual mean biomass of 23 different fish and plankton groups in the Northeast sub-artic Pacific while also capturing seasonal physical and biogeochemical dynamics (Fig. 1-2). An ensemble of plausible ecosystem model formulations was considered to assess model robustness and, in ongoing applications, assess uncertainty in results.

This model is presently being used to explore the combined impacts of climate and fishing in the Northeast Pacific Ocean. The methodologies, however, are generalizable and will contribute to the development of holistic ecosystem-based modeling frameworks in other regions or for global applications. The approach integrates and relies on diverse data sources. This is a strength within data-rich ecosystems, but a limitation for extrapolating to data poor regions. A key next step is thus incorporation of additional ecological and physiological considerations to allow the model to extrapolate to data poor environments.

Fig. 1: A schematic of the coupled planktonic ecosystem-fisheries foodweb model developed for the sub-arctic Pacific.
Fig. 1: A schematic of the coupled planktonic ecosystem-fisheries foodweb model developed for the sub-arctic Pacific.
Fig. 2: Vertically-integrated biomass, in g C m-2, for all living functional groups in the model over the final year of a 20 year simulation. The green bars to the left of each axis correspond to the yearly-mean observed biomass ranges.
Fig. 2: Vertically-integrated biomass, in g C m-2, for all living functional groups in the model over the final year of a 20 year simulation. The green bars to the left of each axis correspond to the yearly-mean observed biomass ranges.