Models in Policy Formulation and Assessment: The WadBOS Decision-Support System - Environmental Modelling: Finding Simplicity in Complexity - page 348

increasing role and the revolution in hardware and
software technologies has equipped them with very pow-
erful multimedia calculators.
The technology presented in this paper is not an alter-
native to calibration and good modelling practices. On
the contrary, as long as a model is not calibrated, it
should stay in the hands of the model developer and
should be used for policy-making purposes under very
strict conditions. Calibration remains a very difficult and
time-consuming task. It is more so for a complex inte-
grated model than for a simple and small one. However,
linking individual models into an integrated model is not
necessarily increasing the level of uncertainty and error
in the final product. Rather, the strong linkages and loops
between the individual models and the fact that the val-
ues of variables are passed from submodel to submodel at
every simulation time step, bring to the surface mistakes
in the model formulations and calculations much more
easily than when the same models would be applied in
isolation (see, for example, de Kok et al ., 2001). Nev-
ertheless, it should be emphasized that even extensively
calibratedWadBOS-like models will only generate poten-
tial developments rather than predictions of future states
or changes. This limitation is not only due to the problem
of calibration, but because of the inherent uncertainty in
the processes represented too. Consequently, the model
and the encompassing DSS should be used for explorative
rather than predictive purposes at all times.
From the WadBOS project it can be concluded that
a policy-support tool can be developed within very rea-
sonable constraints relative to budget, human resources
and development time. This development is much eas-
ier when good base material and expertise is available
and when a stimulating collaboration between visionary
end users and competent DSS developers is propelling the
development. However, the development phase described
is only the first one in the life of a decision support sys-
tem. It needs to be followed by an in-depth evaluation
of the technical contents of the system: its constituent
models, its coverage of the decision domain(s), its way
of dealing with the spatial and temporal dynamics of the
processes and the policies. An evaluation on behalf of
the end users is also pertinent. How do they believe their
decision domain has been represented? Does the system
speak their language? Does it work in the way that they
find useful and pleasant? With the answers to all these
questions a more or less major redesign of the DSS might
be possible or necessary.
Introducing and institutionalizing the DSS in the end-
user organization is the ultimate factor determining its

success (see, for example, Alter, 1980; Klein and Methie,
1995). Even the best DSS will go unnoticed or will fail if
the wrong implementation and introduction strategy is
chosen. Generally speaking, the development and accep-
tance of the DSS will be much easier if the development
is initiated by its end users right from the beginning. The
more they feel the need for a change and the more they
are involved in defining the precise role of the instrument
in the organization, the more likely it is that the DSS will
be accepted and used (see for example, Marakas, 1998).
The more the product itself fulfils tasks that are perceived
as real, in ways that are transparent to the end-user, and
in ways that solve the problem in an obviously better
way than before its introduction, the more the prod-
uct will make a chance to survive. It is therefore very
crucial to determine the right moment in the develop-
ment phase of the DSS to have it change state from a
prototype into an operational version. At this stage, a
gradual introduction, a good technical documentation,
hands-on training, and prolonged technical assistance
during the usage, become of paramount importance. A
hasty, unprepared introduction should be avoided under
all circumstances.

21.10 Acknowledgments

WadBOS was developed by a consortium of Dutch R&D
institutes consisting of Research Institute for Knowledge
Systems bv (Maastricht), Infram bv (Zeewolde), Delft
Hydraulics (Delft), Institute for Environmental Stud-
ies (Amsterdam), Resource Analysis (Delft), and DHV
(Amersfoort). The work was supported by the Land
Water Information Technology Programme of the Dutch
Government as well as the Ministry of Transport, Public
Works and Water Management. A more elaborate and
earlier version of this text was produced for the National
Institute for Coastal andMarine Management/RIKZ, The
Hague, The Netherlands as part of Contract: 42002555
(see Engelen, 2000).

References

Alcamo, J. (1994) IMAGE2.0: IntegratedModeling of Global Climate
Change , Kluwer, Dordrecht.
Alter, S.L. (1980) Decision Support Systems: Current Practices and
Continuing Challenges , Addison-Wesley, Reading MA.
Brinkman, A. (1993) Biological processes in the EcoWasp ecosystem
model , IBN-DLO report 93/6, Texel.