Environmental Engineering Reference
In-Depth Information
ontogenetically-varying behaviour was relatively simplistic. With the development
of our ability to model the physical and biological environment (e.g. light, tempe-
rature, salinity, predator and prey fields, etc.) at increasingly finer resolution, we
must continue to test the sensitivity of our results to increasingly complex models to
decide on the right level of complexity.
Biophysical models for the study of marine ecological issues are becoming more
common and more sophisticated (Miller 2007). This, in part, has paralleled the
developments in computing power and sometimes it is unclear whether increased
model complexity is necessary, rather than giving in to the temptation to virtually
reproduce nature. Higher complexity does not always result in a better model
because of parameter uncertainty and variability - often functions and parameters
used in a model are taken from different species, developmental stages, environ-
ments, etc. The output of more complex models is more difficult to interpret and
“external” issues such as higher computational costs complicate necessary steps
like a comprehensive sensitivity analysis. Biophysical modelling research would
benefit from the type of analysis to investigate the effect of varying degrees of
complexity that has been carried out for trophic ecosystem models (e.g. Anderson
2005). Methods such as using the output of a more complex model as a baseline to
evaluate the performance of simpler alternatives (e.g. Fulton et al. 2003; Raick et al.
2006) have provided a very useful insight and some general guidelines (Fulton et al.
2003). In the absence of a more comprehensive exercise, the obvious conclusion, as
we wrote (Gallego et al. 2007) is that biophysical “models should be as simple as
possible but as complex as necessary” and “the level of complexity should be
adjusted to the model objectives and the observed biological patterns that it aims to
reproduce”.
