Environmental Engineering Reference
In-Depth Information
Often the graphics and sensitivity analyses that the models produce automatically
are high quality and more extensive than you might be able to produce yourself.
The weaknesses include that the modelling package is only able to answer a
limited range of questions, which may not be the ones that you are interested in.
For example, all the high profile conservation packages focus on the biological side
of things. There are some quite specific forestry and fisheries packages that focus
on the manager's perspective, although one, ParFish, is a tool for setting up and
managing participatory fisheries, so has a broader perspective. There is a discon-
nect between toolkits for modelling population dynamics and decision support
tools, which don't necessarily include a proper mechanistic model of the system.
This lacuna is where many of the questions we are addressing in this topic lie. Even
if you want to produce a standard population model, the specifics of your system
may not fit well with the options that the software provides, for example, the
functional forms for density dependence. The model may require information that
you don't have (for example, on frequency of catastrophes or genetic diversity), and
guessing at values may produce misleading outputs.
In the next Section (5.3) we outline how to build your own model. A self-built
model will produce a deeper understanding of your particular system, and model-
ling is a transferable skill that it is worth investing time in obtaining. Whether or
not you use a package, the steps outlined in Section 5.3 are still necessary. It's
tempting to use the package as a prop, carrying out the analyses that it suggests
without first thinking through your own hypotheses. The package is a short-cut for
the actual model-building component of the analysis, but you still need a concep-
tual model first, data, and an understanding of the underlying processes and the
uncertainty involved. You then need to carry out sensitivity analyses and model
validation. You also need to read and fully understand the users' manual so that you
can enter data appropriately and choose the right model structure.
In this section, we will go through the stages of model-building, using two simple
contrasting models as examples. One is an age-structured model of red deer (
Cervus
elaphus
) population dynamics, based on the model presented in Milner-Gulland
et al
. (2004), the other is a very simple bio-economic model of village hunting.
Even if you go no further, it is worth building a conceptual model of your system
as a way of expressing what you know about it and clarifying your logic in thinking
about potential interventions. One common form of conceptual model is a flow
diagram (Figures 5.1 and 5.2). A few things to bear in mind with these are:
●
What do the boxes and arrows actually represent? It is important to separate
processes
and
quantities
, and often the arrows represent processes and the
boxes quantities. Do the boxes and arrows represent the same thing in all parts
