Environmental Engineering Reference
In-Depth Information
model could predict their offtake rates based upon the model's input para-
meters such as an animal population's size and movement behaviour.
●
It is also possible to validate the model
qualitatively
, by showing that
the broad system behaviour is as has been observed previously in this or other
systems. This can be done through generating hypotheses based on the model,
which can then be tested. Qualitative approaches can be useful in distin-
guishing between models with different structural assumptions; do some
models produce behaviour that is more in line with what we know about this
system than the others?
This qualitative validation can lead into a research agenda, which is one of the
strengths of a modelling approach. For example, if you adjust the price per unit of
prey in our hunter model example, you will find that the number of hunters and
the prey population size adjust automatically such that the average hunter cost
becomes the same as the price. This is an outcome of the model rather than an
assumption, but it fits the predictions of the theory of open access resource
harvesting developed using analytical models (Clark 1990). The prediction could
then be tested in the real world by assessing whether or not hunters in open access
systems are, on average, making profits over and above their costs, and whether or
not an increase in the price hunters gain from their prey leads not to more profits,
but to more hunting.
Alvard (1993) took this hypothesis-testing approach when he used data to test
whether Amazonian hunters acted as 'ecologically noble savages' conserving their
resources, or whether they acted like 'rational economic man', hunting according
to some measure of short-term profitability. Both hypotheses predicted that prey
populations would be depleted near the village, and that hunters would instead
hunt in areas further away where prey were more abundant. But he predicted that
if someone with a conservation motive came across a prey item on returning home
through a depleted area they would leave it, while a rationally economic hunter
would kill it. His empirical work showed that the latter was what occurred. Note
that this test was based on a conceptual model, rather than a mathematical one.
Scenario exploration is when you can finally use your robust and validated
model to address the issues of concern to you, and possibly to make management
recommendations. In Chapter 6 we explore various management interventions,
such as imposing harvesting quotas, setting aside no-take areas, or increasing the
opportunity cost of hunting by providing an alternative livelihood (Box 5.3).
Scenario exploration can include many of the techniques from sensitivity
analysis, such as varying a parameter value over a range (such as the harvest rate or
the price per prey item). It can also include elements of the analyses of model
uncertainty—what if all vital rates are correlated, with good and bad years, for
example, or what if there is a biological or anthropogenic Allee effect, such that
population growth rates increase with population size at low levels (Stephens
et al
.
