Geoscience Reference
In-Depth Information
Traditionally, of course, engineers and others have had agreed-upon heuristics for dealing with uncer-
tainties (factors of safety, freeboard, etc.) that err in the direction of more robust design. While involving
significant subjectivity in the choice of appropriate values, that has not stopped them from being incor-
porated into standards and codes of practice (and, without doubt, preventing many engineering failures).
The types of guideline for good practice argued for here represents a formal extension of this approach.
Does it matter to robustness in decision making that the underlying model structure, or the assumptions
about the relevant sources of uncertainty might be quite wrong? This would suggest that, for whatever
reason, we have not (yet) been able to detect a Type I error in choosing a model representation. So
we would not therefore have a good reason to know that the model is wrong - until some information
came along to question that conclusion (although in some cases, such as climate change projections, we
are already very well aware of deficiencies in reproducing past observations). This might involve the
collection of more observations that reveal the deficiencies of the model; it might be that an evaluation
of the predictions of potential future outcomes does not seem to produce sensible results; it could be that
specific experiments are carried out with a view to testing a model as hypothesis about how a particular
part of the system functions. In either case, a continuing review of the heuristic assumptions on which the
analysis is based will be justified as part of an adaptive management strategy. If neither case is evident,
then we have no evidence to question the assumptions.
12.6 Models of Everywhere and Stakeholder Involvement
Once models of everywhere are available however it will very soon become apparent that modelling is
a learning process about places and all their idiosyncracies (Beven, 2007). This is because there will be
a continuing local feedback process about how well a model is performing, particularly if visualisations
of model predictions are made available to stakeholders who have local knowledge. The stakeholders
might be local authority or agency employees, they might be farmers, they might be environmentalists.
They might be interested in flood levels, nutrient levels, or patterns of erosion and sediment deposition.
They will all be interested in checking whether what is being predicted for their place is consistent with
their local knowledge and experience. If it is not, they will say so - requiring some action by those
who set up and run the model (as in the example of the Danish National Water Resources Model noted
in Section 12.5).
An interesting strategy within such a framework is then to involve local stakeholders in setting up the
model. The type of decision framework embodied in the guidelines for good practice is a way of providing
a structure for thinking about local issues relevant to the problem under study. A practical example of
this approach has been demonstrated in a project involving a variety of local stakeholders contributing
to improved flood risk management for the town of Pickering in Yorkshire, UK (Lane
et al.
, 2011). The
stakeholders were directly involved, through the local Ryedale Flood Research Group, in formulating a
model of the runoff generation, flood storage areas and mitigation strategies in the catchment. The result
was a learning process for both the scientists and the stakeholders involved (see the quotation from Stuart
Lane at the head of this chapter). A similar approach has been taken in addressing water quality problems
in Sweden (see Olsson and Berg, 2005; Olsson and Andersson, 2007).
Some people have gone further and suggested that many environmental problems are so complex,
multi-faceted and uncertain that decisions should only made by such a process of discussion amongst
the relevant stakeholders. This view was stimulated by the topic by Funtowicz and Ravetz (1990) which
introduced the concept of “post-normal science”, although the idea that some problems might be too
difficult to address using normal scientific methods goes back much further (e.g. Weinberg's (1972)
concept of “trans-scientific problems”). More recent expressions of similar concepts include the “tangled
thickets” of Wimsatt (2007). Examples of the use of such an approach in water management are provided,
for example, by Pahl-Wostl (2002, 2007). It is clear that the question of how far model predictions can
