Environmental Engineering Reference
In-Depth Information
5
Analyses of the current situation are useful, but they are not so good at helping
us to predict what the effect of an intervention might be, and particularly what the
magnitude of that effect might be. If we wish to manage cost-effectively, then it
is important to be able to weigh up the outcomes of different options. This requires
a mechanistic understanding of the system, not just correlations between variables,
which is what the statistical models discussed in Chapter 4 give us. In this chapter
we give some tips on building and using a mechanistic model of your system. This
usually involves expressing your understanding of the system in the form of
mathematical equations, although conceptual models can also be useful, either as
a first step or as an end in themselves. A good model is a major contribution to
sustainability analyses because it:
●
Forces you to be
explicit
about your assumptions concerning how the system
works, because you have to write them down in equation format.
●
Allows you to investigate the
logical consequences
of the assumptions you
have made and parameter values you have chosen for systems that are too
complex for intuitive assessment.
●
Allows you to investigate the consequences of making
different assumptions
about how the system works and of varying parameter values.
●
Gives
transparency
to decision-making processes because every step is explicit.
●
Allows you to give a quantitative assessment of
uncertainty
, and how it affects
predictions.
●
Enables you to generate
hypotheses
about the effects of management inter-
ventions, which can then be tested, both within the model, and with empir-
ical data. This forms the basis for adaptive management (Chapter 7).
●
In situations when real-life experiments are not possible (a common situation
for species of conservation concern) a simulation model allows
experimenta-
tion
in a 'virtual world'—see Chapter 7.
To be successful in modelling you don't necessarily need to have high-level math-
ematical skills. You do need to have patience, logic and a methodical approach, and
most importantly not to be satisfied with something that is nearly good enough.
You need to convince yourself at every step that you know exactly what is going on,
using graphs and other model outputs to check this. You also need time—learning
