20: Biophysical Models–An Evolving Tool in Marine Ecological Research


Alejandro Gallego

Abstract (from the book)

Although they have been in use for some time, biophysical models are still a relatively new tool in the study of the ecology of marine zoo- and ichthyo-plankton. As the range of specific applications has expanded so has their level of complexity and sophistication. From simple particle-tracking models simulating the transport of zero-drag, neutrally buoyant particles, the field has evolved towards the development of true biophysical models where the “particles” represent biological entities with increasingly sophisticated submodels simulating their development, survival and behaviour. Here I present the results of a modelling experiment to illustrate the effects of increasing model complexity on the trajectory and final distribution of “particles” (e.g. representing early life stages of marine fish). The outcomes are widely applicable and demonstrate the importance of selecting the appropriate level of complexity required for the specific research objectives.

Summarising questions
What are “bio-physical models”, in the context of marine research?

Generally these are Lagrangian Individual-Centred Models, although Eulerian approaches and those that do not focus on the individual are also found in marine ecological literature, as well as a number of cases of hybrid coupled Eulerian-Lagrangian models. The definitions of the Eulerian and Lagrangian approaches used throughout this chapter follow Turchin (1998; Quantitative analysis of movement: measuring and modelling population redistribution in plants and animals (Sinauer Associates, Sunderland, MA)).

To what type of ecological questions are they applied?

These models are relevant to most processes influenced by temporal and/or spatial variability, where no (or at least non-directed) horizontal movements can be assumed (e.g. in the case of planktonic organisms) and where the history of individuals matters.

Why use models, instead of e.g. direct observations?

Physical-biological interactions in the aquatic environment can take place at very fine spatial and temporal scales. Direct observation (in the laboratory and in the field) of these processes is often difficult and costly, although recent technological advances (video plankton recorders, VPR; optical plankton counters, OPC; holographic photography, multi-frequency acoustics, etc.) have shown great potential and are making a useful contribution. “Traditional” field methods (e.g. net-based sampling) offer limited spatial and temporal coverage and resolution, and rely on the time and increasingly rare expertise of highly specialised taxonomic analysts. Bio-physical models can get round some of these problems and complement experimental and observational work.

What are the basics of a bio-physical model?

An accurate representation of flowfields, generally (but not exclusively) the output of a hydrodynamic model (HDM), and a particle-tracking algorithm for the advection/diffusion of the particles representing biological entities. Of course, it can get considerably more complex than that.

How complex can it get?

A whole range of biological “modules” can be included, to provide further biological “realism” to the way particles represent biological entities. The model can simulate the growth of individuals (e.g. as a function of environmental conditions, which may be provided by additional output from the HDM, such as temperature fields, or a function of food availability, which may be provided by the – generally offline – output of a lower trophic level model). Mortality can also be implemented (e.g. from a simple function, like size-dependent mortality, to the explicit representation of predation mortality forced by a higher lower trophic level model). The particles can also be ascribed “behaviours” such as vertical or horizontal migration in response to internal or environmental cues, of varying degrees of complexity. And so on!

What should be watch out for?

With increasing computing power, increasingly complex models are becoming more common. However, note that complex models require a larger number of parameters that may not be readily known. In addition to parameter uncertainty, parameter variability has to be considered. Also, discrepancies with “reality” may still be due to unknown or not-modelled processes. Complex models are also more computationally expensive. On the other hand, too simple, non-mechanistic models may not be applicable beyond their immediate spatial and temporal domain, and may not provide sufficient insight into the relevant underlying processes.

What is the right balance?

Models should be “as simple as possible but as complex as necessary”. A number of tools such as sensitivity analysis, ensemble methods, etc. can be used to estimate the adequate level of complexity. It is important to evaluate necessary complexity in light of the model objectives and the observed biological patterns that it aims to reproduce.

Further Reading

For a comprehensive document on bio-physical modelling methodology, see North, E. W., Gallego, A., Petitgas, P. (eds.). 2009. Manual of recommended practices for modelling physical – biological interactions during fish early life. ICES Cooperative Research Report No. 295. 111 pp. (http://www.ices.dk/pubs/crr/crr295/CRR%20295.pdf)

A Theme Section on bio-physical models in Marine Ecology Progress Series (Advances in modelling physical-biological interactions in fish early life history, edited by Gallego A, North EW, Petitgas P and Browman HI, (2007) MEPS 347:121-306) can be found in http://www.int-res.com/articles/theme/m347_TS.pdf


http://northweb.hpl.umces.edu/LTRANS.htm. Dr. Elizabeth North provides FORTRAN 90 open source code of an off-line particle-tracking model that uses the stored flowfields of a 3D hydrodynamic model. Originally designed to simulate the transport of oyster larvae, it can be easily adapted to simulate other planktonic organisms.