Models in Policy Formulation and Assessment: The WadBOS Decision-Support System - Environmental Modelling: Finding Simplicity in Complexity - page 347

limitations on Infrastructure; Taxes & Subsidies ;and
Quota on fishing and mining. The costs and benefits
of the policies for the Government can be assessed. In
an iterative process, s/he can tune his policy measures
in an attempt to approach the desired state with a
minimum amount of effort and costs.
4. The Scenarios View (not represented) shows the parts
of the model that are most subject to external influ-
ences. These are elements not under the control of
the policy-maker. In order to test the robustness of
the chosen policy measures he can impose effects
on the system that in the real world are beyond his
control. The user can enter hypotheses for the follow-
ing external influences: Atmospheric and Climatologic
variability; Economic growth and decline in each eco-
nomic sector; development of the level of Prosperity in
the Netherlands; and exchanges of water and dissolved
matter between the Wadden Sea, the North Sea and
Lake IJssel.

integration available from the scientific literature. Thus,
the development of integrated models seems more an art
than a science at this moment. It is a deep scientific prob-
lem but also a pragmatic multi-criteria multi-objective
problem as it requires dealing with end-use aspects - what
is appropriate to be integrated in view of the intended
use, scientific aspects - what can and cannot be integrated
on scientific grounds, and technical aspects - how the
integrated model will be assembled and run.
TheWadBOS example has shown that integratedmod-
els used to support the policy-making process come with
a set of requirements of their own, distinguishing them
rather clearly from research models (see also Mulligan,
1998). Policy-makers are most served by models in which
the time horizon, the spatial and the temporal resolution
are policy-problem oriented and not so much process ori-
ented as in research models. They need adequate rather
than accurate representations of the processes modelled
and sketchy but integral rather than in depth and secto-
rial models. While research models are as complicated as
necessary and scientifically innovative , the policy-maker
is better served with an instrument that is as simple as
possible and scientifically proven . A model, too, which
produces usable results as fast as possible , but certainly
within 10 minutes. If these differences are ignored in the
integration process, the result will often be a large-scale
sluggish model not tailored to the needs and expectations
of its end user. Clearly, a fast, interactive model will do
much better for policy exploration.
A model will only serve the policy end-user if it is
presented in a format that enables it to be worked
with. In decision support systems, models are supple-
mented with sets of tools to structure and carry out
the analysis in a manner that makes intuitive sense to
the policymaker. One of the crucial elements in the
DSSisthe user interface . Attaining the level of user-
friendliness and flexibility that policymakers seem to
desire remains a very big challenge for computer sci-
entists, model developers and domain specialists alike.
Indeed, it is very difficult to package in a single appli-
cation a system with the level of interactivity, flexibility
and the fast response times wanted, the user-friendliness,
simplicity and the transparency desired and the level
of accuracy and certainty expected. A lot more inno-
vative research, design and implementation work will
need to be carried out to get to this point, if ever
we will. The demand for the kind of instruments is
certainly real. Policy questions have reached a level of
complexity that can no longer be dealt with by politi-
cians alone. High-level technicians are playing an ever

21.9 Discussion and conclusions

When the intention was originally formulated to develop
asystem' to gather, order and link the knowledge available
about the Wadden Sea to facilitate the policymaking pro-
cess, ' it was clear that WadBOS was primarily (1) a tool
for analysis and evaluation of autonomous dynamics and
the effect of policy interventions thereon. Other views
coexisted relative to its function: (2) a storage tank of
data, information, and knowledge, (3) a tool for commu-
nication (of results), or (4) a tool for the management of
information and knowledge about the Wadden system.
Practice so far has shown that the role of WadBOS as an
instrument for communication has been at the least as
important. The fact that this instrument integrates in a
formal manner the many processes that make and change
the Wadden Sea gives direction to a lot of the discussions.
It does not necessarily make stakeholders and policy-
makers agree on issues more easily, but it helps in clarify-
ingwhat it is that they do or donot agree about. The capac-
ity to visualize conflicting views and reduce these conflicts
in an interactive session is of paramount importance.
A core element in WadBOS is the integrated model of
the Wadden Sea. The use and need of such instruments is
strongly advocated in new disciplines such as integrated
assessment (see, for example, Gough et al ., 1998). Despite
the fact that the term 'integrated model' is used all over
to mean a wide variety of things, there are very few oper-
ational definitions and recipes or procedures for model