Environmental Engineering Reference
In-Depth Information
With object-oriented modelling, however, it is usually possible to approximate
the actual structure to the ecological system quite well (see Chap. 12). Conse-
quently, the model results tend to be surprisingly robust when parameterization is
chosen according to biologically realistic values. The robustness of model results
can be related to the appropriateness of the model assumptions with regard to the
problem under investigation.
When one begins to work on an ecological model, one is seldom completely free
to choose its logical structure. The core task of the modelling effort is to concentrate
on the main driving forces of the complex ecological dynamics that one wants to
describe, and so one is compelled to include only the necessary structural compo-
nents. In other words: Modelling usually should follow Ockham's razor and one
should prefer the simplest explanation that conforms to the observations. This
implies that it may be necessary to change and adapt the structure of a model
when this is
ecologically required
(see Chap. 12). By a change which is
ecologi-
cally required
, we mean that either relevant changes in the modelled object
structure of biological entities (e.g. status or motivation of the individual) or the
modelled environmental structure (e.g. temporal and/or spatial) of the simulated
model universe have to be made. With equation-based models, which are mainly
parameter driven (see Chaps. 6 and 7), it is frequently more difficult to adapt the
model interaction structure, than with the object-oriented approach.
24.4 Across-Scale Modelling and Multi-level Modelling
How simple and straightforward the first ecological models of Verhulst in the
nineteenth century now seem - the increase of a population in time is proportional
to the population itself minus a constant times the square of the population size
and - voil
ยด
! - logistic growth emerges. Current progress in modelling is not just
a matter of adding building blocks to this ancestral model. The focus of scientific
progress in ecological modelling is on developing modelling systems that success-
fully can deal with the emergence of qualitatively new and unexpected properties
that arise from the interaction of particular components. This is a frequent result in
object-oriented modelling; e.g., model predators are allowed to randomly consume
prey, both populations proliferate - and wow! - the whole spatial area/community
enters into coherent oscillations. Though the emerging spatio-temporal pattern is
not a property of any single individual interaction, it emerges as a result of the
overall ongoing activity.
Starting from the interactions of individuals and their behaviours, through popu-
lation dynamics, community structure development, ecosystem and landscape
processes, and on upwards to the biosphere scale; each step is characterized by self-
organization effects that result in emergent properties (Jantsch 1980). The modelling
of self-organization processes, is, in our opinion, the supreme objective of eco-
logical modelling and allows to achieve a deeper understanding of how natural
processes interrelate (e.g. Kauffman 1993; Gell-Mann 1994). Approximating these
