Environmental Engineering Reference
In-Depth Information
forward on several fronts to create opportunities
for a fresh look at the strategies and tools that they
(and other agencies) use, both in the context of
real-time emergency response to floods and in
flood risk mapping conducted by planners and
engineers.
transferred and incorporated into both planning
and warning procedures. Previously, the incorpo-
ration of models into flood mapping concentrated
on fluvial (upstream-downstream watershed-
based) models and to some extent coastal floods.
Now, partly in response to the Pitt Review
(Pitt 2008), scientific modellers are extending
their activity to include surface water runoff
and urban drainage issues, such as contaminated
sewer surcharges. Professional agencies are con-
tinuously and reflexively incorporating the
science of FRM, and then tailoring the scientific
end of flood risk models as communication
strategies and tools.
Yet given the rapid pace of these changes, the
models are sometimes described as having devel-
oped beyond the capacity of the professionals to
incorporate them, and this applies particularly to
the incorporation of uncertainty tools (Kinzig
et al. 2003). Certainly, there is currently some-
what poor connectivity between radar warnings,
even top-of-the-range radar-based prediction mod-
els, and their ability to improve communications
to the public during real events. Real-time radar
predictions of flood onset are also poorly interfaced
with the models that predict inundation patterns,
depths and velocities, although as this topic tes-
tifies, these models do separately exist. Enhanced
co-operative working between the Met Office and
the EAwill improve communications (Purdey and
Davies 2008). Faulkner et al. (2007) argued that
translating complex and uncertain science is a
considerable
Science and risk communication tools
and methods
Some of the tools available at the science-
professional interface are listed in the first two
columns of Table 19.1. Of these, floodplain maps
are the most obviously useful risk communication
tool that flood risk managers can use for forward
planning. Their development is underpinned by a
widely applied classical scientific formulation of
risk. This formulation calculates risk at any time,
t, as the product of the probabilityof anevent (based
upon the statistical properties of the long-term
event series), and the consequences of that event
(Kron 2002). This is most often expressed in the
form:
R
t
ΒΌ
H
t
V
t
where R
t
is the risk estimate at time t, H
t
is hazard
or probability, and V
t
is vulnerability or conse-
quence (which, for floods, is often expressed in
economic terms as a cost of damage). The generic
term 'H
t
' is taken here to embrace distributed
physical models of rainfall events, or storm surges,
or floodplain inundations for event at time t. The
scientists included in the papers presented in this
topic are primarily concerned with modelling a
range of hydrological and/or meteorological pro-
cesses through time t (
S
additional
challenge
for
FRM
professionals.
In Figure 19.2, the flow of scientific informa-
tion, in the form of meteorological, flood and
inundation models couched largely in mathema-
ticized language, is portrayed as suffering what
Mansilla et al. (2006) would term a 'serious trans-
lational gradient' by the time this information
interfaces with the public. Since scientists, profes-
sionals and the public use the terms 'risk' and
'uncertainty' and 'probability' in entirely differing
ways (Holmes 2004), the problem is partly semi-
otic (Morgan et al. 2002; Holmes 2004; Leiss 2004;
Pappenburger et al. 2005; Wilby et al. 2008). The
EA (Environment Agency 2004b) is now moving
H
t
) at various recurrence
probabilities. Models can simulate a single runoff
or rainfall event andmake real-time predictions. In
the domain of flood risk management, however,
most physical scientists were taught about flood
prediction and management as being underpinned
by probabilistic extreme event theories. In these,
the hazardH
t
is thendefined in terms of a discharge
frequency curve (Pappenburger et al. 2005) that can
be derived from existing long-term records (at least
where such records exist).
