Environmental Engineering Reference
In-Depth Information
We presented our categorization system of ecological models without ignoring
that a strict categorization might result not only in logical problems (see Chaps. 2
and 4) but could also lead to inadequate practical results. As far as we see this issue,
sharp boundaries between model types are not reasonable. For intellectual inspira-
tion such categories are only important insofar as they help to identify and
understand real world processes.
24.3 Structurally and Functionally Realistic Model
Construction Tends to Generate Stable Model
Behaviour
From the experience of many years of applying models, using different techniques
and approaches, we arrived at an interesting conclusion. The more abstract a
model is, i.e. the more it ignores the structural heterogeneities of the system being
modelled, the more effort is required to parameterize it to meet a specific situation.
On the other hand, if the model adequately represents the structural characteristics of
the system being modelled, parameterization frequently tends to be not such a
difficult matter. That is, model behaviour is not too sensitive to parameter choice,
and it is relatively easy to decide on the range of parameters that are biologically
reasonable. With the appropriate caution, this observation can be turned around:
if we observe crucial sensitivities in a model, it may be a hint that it might be
structurally inappropriate for the way the ecological system actually operates. Why
this can often be the case is explained below.
Organisms that cannot complete their whole life cycle within the environment that
is available to themwill quickly vanish from the community of a particular ecosystem.
Only organisms that are compatible with a particular environmental setting, and the
uncertainty and variability it involves, can persist in the particular system. The
considerable losses of biodiversity that accompany environmental alterations indicate
that the persistence of organisms is not a trivial issue. It also implies that existence
requires an environment that meets stringent conditions. In particular, environmental
properties are needed that can maintain sufficient population densities as well as
prevent mass proliferations and the resulting overall system instabilities that might
lead to collapses. This can be considered as a standard situation, fromwhich of course,
exceptions and deviations exist.
Now, modelling such a situation may be difficult. Maintaining a model system in
reasonably stable conditions often requires very careful fine-tuning of the model to
meet those observed conditions. It is quite likely that in order to make the model
mimic those observed conditions, one needs to apply mechanisms in the model that
differ from the way the biotic system actually functions. With differential equation
based models it is a frequent experience that parameter fine-tuning is required to an
extent that is biologically unreasonable (for example, if making the model system
behaviour properly depends on the fifth decimal of a parameter).
