Content uploaded by Ilke Borowski - Maaser
Author content
All content in this area was uploaded by Ilke Borowski - Maaser on Aug 01, 2016
Content may be subject to copyright.
Content uploaded by Ilke Borowski - Maaser
Author content
All content in this area was uploaded by Ilke Borowski - Maaser on Aug 01, 2016
Content may be subject to copyright.
Exploring the Gap Between Water Managers
and Researchers: Difficulties of Model-Based Tools
to Support Practical Water Management
Ilke Borowski &Matt Hare
Received: 1 December 2005 / Accepted: 2 June 2006
#Springer Science + Business Media B.V. 2006
Abstract Supported by EU funds, the European research community has been putting much
effort into providing model-based tools to support water resource managers in implementing
water management as well as the implementation of the Water Framework Directive. This
paper presents the results of a two-year long elicitation phase which aimed to explain why the
use of tools in water management is not as great as the corresponding investment in applied
research in this area might suggest it should be. The paper identifies a gap between water
managers and research community that is evidence of a mutual misunderstanding of the
fundamental activities of both communities. We elaborate on these misunderstandings
between these two communities by focussing on their attitudes towards seven assumptions
that derived from an elicitation phase carried out between 2003 and 2004. These
misunderstandings appear to revolve around the issues of the role and importance of
model-based tools in water management; the transferability of models to new target sites; the
role of participatory modelling in water management; how to solve lack of confidence in
model-based tools; the development of computer user interfaces to improve tool usability;
and the nature of model integration. Based on these insights, recommendations for improving
research, development and ultimately the use of model-based tools in river basin management
processes are proposed. The recommendations include improving researchers’understanding
of water management processes and the role their tools play within such a process; identifying
for both communities the importance that such tools can play as part of social learning-
oriented management processes; improving the role of software consultancies as carriers of
research results; considering new methods of model transferability between target basins; and
expanding the structure of funding for academic research and development projects to allow
Water Resour Manage
DOI 10.1007/s11269-006-9098-z
I. Borowski (*)
Institute of Environmental Systems Research,
University of Osnabrueck, Barbarastrasse 12, 49076 Osnabrueck, Germany
e-mail: borowski@usf.uos.de
M. Hare
Seecon Deutschland GmbH, Westerbreite 7, 49084 Osnabrueck, Germany
e-mail: matt.hare@seecon.org
the greater provision of non-technical requirements such as post-development tool
maintenance and transferability, required by water managers.
Key words use of models .tools .participatory river basin management .WFD .
policy-science interface .social learning
1 Introduction
Acknowledging the complexity of integrated water resource management, the research
community has been putting much effort into providing model-based tools. Those tools aim
to increase scientists’and resource managers’understanding of the interaction between
physical, chemical, biological and socio-economic processes. This effort has been
supported within the FP5 research programmes of the European Union (key action water):
250 M euros have been invested into 180 research contracts trying to explore these
interactions –often involving the development of new model-based tools, especially in
support of participatory, integrated water management (IRBM) according to the European
Water Framework Directive (2000/60/EC) (WFD).
To bring the research results to practice, the European Commission has not only
prescribed the support of the European policies as one of the objectives in the framework
programmes. It has also established mechanisms which allow small- and medium sized
software houses and consultants to be involved. The European Commission is convinced
that “European research is delivering more than knowledge; indeed it is providing
sustainable solutions to problem owners. The ‘problem solving’approach combines
scientific expertise with industrial involvement to secure reliable and exploitable results
with highly marketable potential”(European Commission 2004).
This important role of model-based tools is confirmed by scientific literature which
claims potential advantages of using computer based models and tools to integrate
knowledge and provide support in management activities (e.g., Larsen et al. 2000; Welp
2001; Dørge and Windolf 2003; Wasson et al. 2003; Rekolainen et al. 2004). However,
there are also more critical voices. Many of the tools developed for direct (decision) support
are not suitable and have not been close enough designed along the requirements of the user
group (e.g., McIntosh et al. 2004; Robinson and Freebairn 2001; Ubbels and Verhallen
2000; Vonk et al. 2005; Greeuw et al. 2000).
The direct transfer of research results to practical management is still not considered
satisfying by the EU, e.g., the development of WISE (Water Information System for
Europe; (Quevauviller 2006) is an attempt for improvement by improvement of
communication.
In this paper we show that ‘simply’improving communication on the research outcomes
may not be enough. We present results from a two years elicitation phase to explain further
why model-based tools are little applied in comparison to the amount of research done by
the research community. We elaborate on the gap between water managers and researchers
by focussing on the misunderstandings that the work has identified between these two
communities in terms of their attitudes towards seven assumptions. These assumptions
concern the role of model-based tools in water management, how to solve the problem of
lack of water managers’confidence in models, what needs to be integrated into an inte-
grated model and the transferability of model-based tools to new target sites (see Table V).
Water Resour Manage
Before continuing with the next section it is important to point out that the results of the
elicitation phase have to be seen within the context that the participants were all self-
selecting and were drawn to the workshop since they were interested in modelling and
participation. All results of the workshop are well documented and were run through a
feedback cycle with the participants. Still, this naturally does not, however, mean that all
members of the group were in agreement. We also do not assume that individual readers
who would consider themselves either a water manager or a researcher would agree with all
of our arguments. What we would like to achieve, though –by using grounded, yet starkly
formulated arguments –is to challenge the readers to reconsider their own assumptions
about model-based tool use.
For this purpose, we structured the paper as follows: In Section 2, we outline the
methodological approach used in the activities during the elicitation phase, as well as a
description of the participants from the two communities. Section 3synthesizes the results,
presenting and explaining the difference in assumptions between water managers and
researchers. Finally, in Section 4, recommendations for improving the research,
development and use of model-based tools in river basin management processes are made.
2 Methodological Approach
The requirements elicitation phase took place between 2003 and 2004 as part of WP5 of the
Concerted Action Harmoni-CA (http://www.harmoni-ca.info). The phases’results have fed
into the design and implementation of further activities of Harmoni-CA and especially
Harmoni-CA/WP5 such as three further policy workshops, two workshop for researchers, a
second set of river basin meetings and the development of documents summarizing
available support by model-based tools for water managers. In the design of the
requirements elicitation phase it was acknowledged that there is some lack of commu-
nication between the groups of model developers and water managers; and that both groups
need to learn more of the requirements and capabilities of the others. Therefore one aim of
the elicitation phase was to learn about the details of the reasons why model-based tools
developed by research are not frequently applied in water management. We also wanted to
learn how their application in practical water management –either by using them online in
participatory processes, using their results and insights to support decision making –could
be improved. In this paper, the term model-based tool refers to computer-based
management support tools that make use of some form of model of the system to be
managed. Application of these tools refers here to both the integration of outputs from these
tools (model results, scenarios) within a decision making process as well as the integration
of the tool into a communication process to communicate the results of decision making.
For the convenience of the reader we use tool, model or model-based tool synonymously.
Another aim of the elicitation phase was to provide the floor where exchange between
both groups could take place and potential for cooperation could be explored. We initially
targeted both groups separately: inviting the water managers and the researchers to different
workshops. After analysing the main issues in their commonalities and differences, we then
set up a series of river basin meetings with the water managers. Here we aimed to increase
our knowledge on the issues from the water managers’perspective. We then invited both
groups to a third workshop, the Synthesis Workshop, to discuss them. During the three
workshops and the five river basin meetings that were organised, 23 water managers and 29
researchers participated. The results of this two-year long phase include 128 requirements
for models, 28 research directions (issues or fields of research) and hours of intensive
Water Resour Manage
discussions on how, when and why model-based tools may or may not be appropriate to
support the implementation of the European WFD.
The results of all workshops have been recorded in official project reports (Hare 2004a,b,
2005). These reports form a large proportion of the material used in developing the
arguments in this paper. It is therefore important to know that the reports were written as
faithfully as possible to reflect the attitudes of the participants within the workshops. To
reduce mistakes to a minimum and to make sure that the contents of the reports could be
seen as valid representations of the attitudes of the participants, as part of a feedback cycle,
the participants were sent draft versions of the reports and asked to send in their corrections
which were added to the final version of the documents.
In the following section, we first describe the water managers in their context. We then
show how the group of researchers was composed. The third part will describe the different
events and their results, of which an analysis is provided in Section 3.
2.1 1st Target Group: Water Manager
In the context of this paper, we use the term water manager to refer to participants in our
workshops and other meetings who are working at public authorities which are involved in
the implementation of the WFD. In the context of Harmoni-CA water managers are often
differentiated from water policy makers. Water managers work on an operational level
directly implementing the WFD including the preparation of plans, programs or public
participation activities. Water policy makers are, in general, placed at national or European
level; their task refers to coordinating national, European or inter-regional implementation
strategies like the work in the EU Common Implementation Strategy working groups. Of
our participants 16 were mainly working at regional level. Six persons were working at
national level: three of them had rather specific tasks linked to modelling or public
participation. The other three participants were working at national level who also had
coordinating tasks qualifying them as water policy makers. Following the definition above,
four water policy makers were involved in the elicitation phase. Nevertheless, in the
following we refer to all of them as water managers.
To seek participants, we announced the workshop through our contacts and also publicly
through water management newsletters. During the elicitation phase, a total of 23 water
managers participated from 11 different countries (Denmark, Belgium, Germany, Estonia,
Finland, France, Hungary, Portugal, the Netherlands, Ireland, and Poland). To characterize
the water managers we asked all participants for a short description of themselves including
their area of responsibility before the 1st Policy Workshop or during the river basin
meetings. The water managers had different expertise regarding modelling and public
Number of water
managers....
...with main expertise/ interest in modelling 12
...with main expertise/ interest in public
participation
10
...working in leading position 6
...with all characteristics 2
Total number of participating water
managers
23
Table I Water managers partici-
pating in the elicitation phase
Water Resour Manage
participation. Some of them were working in a leading position, being responsible for the
tools applied and for the conclusions drawn from them including the recommendations
being then sent to water policy makers. Table Igives an overview on the participants.
2.2 2nd Target Group: Researchers
When we use the term researcher, in this context , we refer to scientists involved in
research and development of model-based tools. We use the terms model developer,
researcher, and modeller synonymously. For our 1st Method Workshop, we invited senior
researchers (Post-docs or Professors) who are involved in European research on river basin
management. Of the 20 participants, 15 had between 5–25 years of experience with
developing models. Nine participants had experience with stakeholder integration in water
management. Five participants had expertise in both. Researchers from the European
projects SMURF, Mulino, Advisor, WADI, MERIT, HarmoniQUA, BMW, HarmonIT,
TransCat, Harmoni-CA, and HarmoniCOP were present. At the Synthesis Workshop, we
invited researchers to present relevant tools to the water managers. This way, five
researchers from the 1st Method Workshop and 10 additional researchers took part.
2.3 1st Harmoni-CA/WP5 Policy Workshop
The principal topic of the 1st Harmoni-CA/WP5 Policy Workshop (16/17th October
2003) was the identification of the requirements of 15 water managers for model-based
tools to support the participatory elements of the WFD. The basic methodological approach
for design of the workshop was taken from the field of user-centred requirements
engineering (McGraw and Harbison 1997). In requirements engineering one attempts to
identify the functional and non-functional requirements for the design of a new piece of
software by eliciting information from potential users of the software and/or from experts in
the area of expertise for which one is trying to provide software support. The principle
assumptions behind such user-centred approaches (McGraw and Harbison 1997) is that by
so doing (a) the software engineer elicits valid requirements from such people and (b) that
these requirements will help produce a tool that will be both functional and used.
There are many different methods of doing requirements engineering. Maiden and
Hare (1998), for example, interviewed individual experts (using card sorting techniques), to
elicit their categorical perspectives on particular business processes (e.g., assigning
passengers to airline seats). The results of the interviews were collated, analysed and then
categorised, so that requirements for re-usable software components could be identified for
a tool to support the development of software for automating these business processes. One
does not have to always interview experts or users, however. McGraw and Harbison (1997)
describe how experts can alternatively be observed carrying out their work tasks, which can
then be decomposed and analysed for requirements in a process known as task analysis.
Rather than ending in a simple categorisation of the requirements, a scenario matrix can be
produced, describing the activities to be automated and the way they may be implemented.
Through use of a storyboarding approach (visually linking the scenarios in a coherent and
logical set of activities), these matrices can be presented to the users or experts to check that
the tool does what is required, before it is built. More requirements can then also be elicited.
Whether or not from a conscious effort to adopt the principles of user-centred
requirements engineering, the involvement of users in the development of water
management tools is being increasingly adopted (e.g., Letcher et al. 2004). Such principles
Water Resour Manage
were also underlying work by Swinford et al. (2004) when they opted to test and evaluate
Information and Communication Technology tools for use in participatory water
management not only on the tools’original target audiences (schoolchildren), but also on
future, potential users (postgraduates), citing the following reason:
Both groups ... took part in the evaluation because designing future tools ... would
require that individuals of all ages and ability would need to be able to use [them]. (p4)
The 1st Policy Workshop was in effect a user-centred, group requirements engineering
process, for the design of model-based tools to support the implementation of the WFD,
which drew on elements of the approaches described above. The water managers acted as
both potential users of the model-based tools to be designed and as experts in the WFD
management areas for which support is needed. Instead of dealing directly with individual
experts, or carrying out task analysis, we posed interview questions to moderated groups of
experts within the workshop. The use of group moderation for eliciting meaningful
information from a group is well established, not only evidenced by the appearance of
‘pocket guides’to moderation, such as Edmüller and Wilhelm (2002), but also its
application in serious scientific endeavours, particularly in operations research (e.g., Vennix
1996; Delbecq et al. 1975), knowledge elicitation (e.g., Grabowski et al. 1992) and more
recently in participatory water management (e.g., Mostert 2003; Hare et al. 2006) and
participatory modelling (Lynam et al. 2002; Hare et al. 2003). Such group work for
requirements engineering, allows not only the experts to be the provider of requirements,
but also to allow them to do the work of the analyst in terms of verifying and selecting the
most important requirements to be put forward by the group, since each individual’s
requirements will go through a peer-review process within the group, forcing it to be
justified for inclusion in the group’s end results. Obviously working in groups can have its
negative consequences as in terms of group censorship, or group-think, but good
moderation of the group, among other things, can help mitigate such factors (see Vennix
1999), as long as the groups are small (8–12 persons maximum).
Before the workshop the following interview questions were posed to the participants in
order to prime them for the work to be done:
(1) In what ways is guidance for participation in river basin planning under the WFD
different in comparison to current practices?
(2) How are models currently used in river basin planning and to what level of success?
(3) How are models currently used to support participation in current planning practices?
To what level of success have they been used?
Once at the workshop the experts were split into small groups, each one of which to
answer one of the questions above. The external moderators of each group took the group
members first through a brainstorming exercise in which answers to the question from each
individual were written onto meta-plan cards. Each individual had three to four cards
allowing one answer per card. Only answers written on such cards were then presented and
considered by the group for inclusion in the group’s‘answer’to be presented to the other
groups. Implausible or unjustified answers as perceived by the group, were excluded. Thus
a peer-reviewed group answer was created and documented.
The next step lies at the heart of the Brücke Method (Hare 2003) subsequently employed
in the workshop. Small group work in workshops is very useful to make sure that most
participants can have the opportunity to contribute, however, often the price for this is that
participants only get a very fragmented view of the work that has gone on in the workshop
Water Resour Manage
and thus much information generated is lost to the general participant. To make sure that the
results from small group work were not lost and that each participant in the workshop had a
meaningful opportunity to read, critique and influence answers from other groups, the
documented answers were then exchanged between groups. The groups then met again in
the next workshop session to comment on the documents. Again, brainstorming and
moderated discussion led to the group adding comments to the documents before returning
to the original groups who wrote them, so that they could read and then make final changes
to them before they were presented in plenary and entered into the formal results protocol
of the workshop. Hence, not only did this allow each participant to know what was going
on in the other groups in some detail, the final output of each group was also peer-reviewed
by all the participants.
At the end of the day, based on the answers to the three questions, each participant was
asked to select the type of tool they felt was most important to be designed to support the
WFD work they did. According to the answers, the new small groups were formed of like-
minded participants, who were then given the job, inspired by the storyboarding approach,
to draw a design of their most important tool and its functionality. Four tools were
designed: an information provision tool for both public and the stakeholders; a consultation
tool for stakeholders; a tool for education/awareness raising for public and stakeholders;
and an active involvement tool for stakeholders. The designs were then made available to
the other groups who could discuss and reflect on them and suggest what important
requirements could be implied from them. The project team collected these further
requirements and added them to the workshop results protocol.
2.3.1 Analysis of Workshop Results
Each day’s activities had the common goal of eliciting water managers’requirements and
based on the results of them, a set of 119 requirements for tools and models to be used to
aid the participatory element of the WFD were elicited from the various stages of the
workshop. After the workshop, these requirements were analysed, categorised and grouped
by the project team, according to 13 categories (see Table II). The mode number of
requirements per category was 10. The participants were then asked to prioritise the
requirements in each category. This was done so by asking for them to rank the top six most
important requirements per category. Nine of the 15 participants provided us with a
complete list of ranking of all requirements. The Condorcet rule (see accuratedemocracy.
com/elect.htm) was then used to calculate an aggregate ranking (a preference list) for the
requirements in each category, i.e., a ranking of the requirements that could be said to
reflect the ranking of the participants as a whole group. The Condorcet rule was used since
it claims to avoid the problem of vote splintering caused when there are a large number of
options to rank (amongst which, as in our case, there are many similar ones) and thus an
extreme option (i.e., the one with less in common with other options) ends up winning.
Table II shows an overview with the top two priority requirements per category common
to all participants. The complete list of requirements and discussion results can be found in
the workshop’s report (Hare 2004b).
2.4 1st Harmoni-CA/WP5 Method Workshop
In the 1st Harmoni-CA/WP5 Method Workshop (16/17th February 2004), the principal
topic for the 20 senior researchers taking part was to consider future research directions for
Water Resour Manage
Table II Overview of the top two priority requirements per category by water managers at the 1st Policy
Workshop
Category Top two priorities requirements
1. Tools –models, to be used by participants,
often need to be packaged into useable tools.
In what sort of tools could models be packaged
under in order to support particular purposes?
Information provision tool for public
and stakeholders
Management tool for competent authorities
2. Goals –what purposes should models
and tools serve?
Promote common understanding
between stakeholders
Act as supporting material for talking
with the public and stakeholders not as
the only source of information
3. Constraints –under what constraints
should models carry out these purposes?
For stakeholders as users
Models should be based on reliable scientific data
Models should always inform about solutions
as well as problems
For the public as users
Models should be based on reliable scientific data
Models should have contents that are explainable
and understandable
For managers as users
Models should be based on reliable scientific data
Models should allow the user to understand
the limitations of the model
4. Trust –for models to be used in participation
they must be trusted. How can trust be instilled
in models?
By being based on a solid scientific basis
By providing sensitivity and statistical analyses
5. Integration –participation in river basin scale
management means that models need to integrate
a large amount of different knowledge and
support different roles. What should be integrated?
Ecology, economics, social, physical, chemical and
biological impacts on water quantity, quality
Many tools together to support active involvement
in management by stakeholders
6. Results presentation –for participation,
presentation of model results needs to be
well done. How?
There should be on-screen cross referencing of
information on the screens i.e., it must be clear, the
links between action (input) and effect (output)
The presentation of indicators that mean something
to the stakeholders and public
7. Communication –without good communication
of models to the participants, participation may
fail. How can good communication be achieved?
By communicating its limitations
By including scale of impacts
8. Usability –in participation, models need to be used
easily and effectively by a wide variety of people.
How can high levels of usability be achieved?
By making them simple to use
By including a GUI to design scenarios
9. Data requirements –participation in river-basin
scale management has particular demands on data.
What demands?
More data exchange processes
More plan exchange processes
10. Maintaining involvement –participation needs
to be maintained over long periods of time or else
models need to take into account changes in their
users. How?
Models should be interactive
Models should be open to all participants
11–13. Models –participation in river basin management requires
a range of models to support the entire planning process. Which ones?
Water Resour Manage
the development of model-based tools to aid water managers involved in participatory
RBM. This workshop used an almost identical group requirements engineering approach as
to that of the previous workshop. On the first day, the participants were divided into small
groups to answer the following questions:
(1) What implications does the participatory element of the WFD have on the role of
models for use in WFD? At what stage of the management process should particular
models be used?
(2) What is the state of the art in integrating human dimensions into models and how far
does this meet the requirements of participatory river basin management under the
WFD?
(3) What is the state of the art of river basin models in terms of purpose, what they
integrate, and how they are applied in management? How far does this meet the
requirements of participatory river basin management under the WFD?
Based on the results of the first day, on the second day the model developers were ask to
visually describe (‘storyboard’) how a future model-based tool could support a particular
stage of the management process for the implementation of the WFD. The tool should
integrate participatory modelling and management linked to the WFD. In four different
groups the modellers designed a model to support the setting of environmental targets; two
designs for models for the development of a river basin management plan; and a flexible
and modular process for participatorily eliciting proposals and conditions for models to be
used to develop river basin management plans.
2.4.1 Analysis of Workshop Results
By the end of the workshop, 28 research directions were identified and recorded in the results
protocol of the workshop. The participants were then asked during the final session to mark
three key research directions that they thought were important. The research direction with
the highest number of votes was “defining linkages between socio-economics and
hydrology”. Table III shows the 10 research directions which received two or more votes
Table II (continued)
Category Top two priorities requirements
11. Meta-level requirements What-if, what then scenario models
Just one good model per problem made up from the
best parts of all the alternatives
12. Domain-level requirements Cost-effectiveness of measures models
Models that take a list of problems and output a list
of measures to deal with them
13. Instance-level requirements River quality models including physical biological
and chemical impact assessment for different
activities
Agriculture and land use
Water Resour Manage
including the categorisation by the project team. The full list of research categories and
results of group discussions can be found in the workshop report in (Hare 2004a).
2.5 Analysis of the Results of the First Two Workshops
After the both workshop, the results were analysed to identify what were the main issues
discussed by each group. Similarities and differences were identified in the issues and the
way they were discussed? Table IV shows the results of the analysis.
2.6 River Basin Meetings
Based on the results of the analysis, we set up five river basin meetings (between January
and September 2004). We intended to learn more about the issues which had arisen from
the previous two workshops. The river basin meetings took place with selected water
managers in Denmark, Germany, France, Belgium and Hungary. All of the meetings were
organized with authorities that are responsible for the implementation of the WFD also at
operational level. Except for the interview in Germany, all river basin meetings took place
with water managers who had participated in the 1st Harmoni-CA/WP5 Policy Workshop.
This ensured openness to the issue of model usage as such.
Methodologically speaking, an approach, in which individual water managers were
interviewed using a semi-structured interview, was adopted. The questionnaire used
included questions related to the issues identified in the previous workshops about the water
managers’role, the role of models in their tasks, the importance of different sources of
information in their decision making; the role of uncertainty in their decision making; the
role of public participation for their work and the contents of any messages they would like
to communicate to the modelling community. The results of the interviews were written
down in a discussion protocol. Since the aim was to find more detailed information to
support or refute the emerging findings of the first workshop, these results were
qualitatively analysed by coding the protocol (Glaser and Strauss 1998) according to a
pre-coding defined by the issues in Table IV. The main results of this series of meetings will
be described in Section 3below. For more details, see Hare (2005).
Table III Overview of the 10 research directions with two or more votes. Categories can be translated as
(see also in text) HD=Human Dimension; PS = Participation Support; MD =Model design
Category Research direction
HD Defining linkages between socio-economics and hydrology
HD Integrate human dimensions (e.g., perceptions and values; human activities as driving forces for
environmental change; decision making processes; human activities affected by environmental.
change)
MD Guide/expert system for allowing managers to tailor a participatory multi-tool approach
MD Field experience on ‘cyclic modelling’
MD Best practice/cases/narratives for participatory modelling in scenarios/DPSIR etc.
PS Matrix for efficiency and ongoing evaluation of participatory process
PS Cost effectiveness of public participation/pre-assessment –case studies based
PS Pitfalls of selecting representatives from different groups –how, what it means
PS Role playing as an analytical tool for stakeholder identification
PS How to use workshops for participatory river basin modelling?
Water Resour Manage
2.7 The Synthesis Workshop
The Synthesis Workshop (4/6th October 2004) invited both water managers and model
developers together to consider the results of the 1st Harmoni-CA/WP5 Policy and Method
Workshops and the river basin meetings. The goal of the workshop was to encourage the
communities to share experiences through discussion and presentation of tools in
development. Also, it was an attempt to form intercommunity working groups which
would work on developing position papers to explain and suggest means of tackling the
four thematic issues identified in Table IV. Of the 22 participants 12 persons were
researchers and 10 policy makers.
In this workshop, the group requirements engineering approach of the first two
workshops was not replicated. The approach adopted included a mixture of presentations,
demonstrations and small group work. First, the water managers were invited to present
their major challenges linked to the WFD implementation. This included their current use
and expected support of models for their work. Afterwards, the model-developers gave short
Table IV Overview of issues from 1st Method and 1st Policy Workshop (adapted from Hare 2005). Italics
show areas of full agreement between the two communities
Area Researchers Water Managers
Issue 0:
Role of models
Models are central to management Models should be seen as one
source of information for
managers (of many)
Issue 1:
Participatory
modelling and
management
Participatory management can
happen through participatory modelling
What is participatory modelling?
Include stakeholders/public during
model development, use and/or
access to results
Deeper involvement of public
in decision making only after
awareness raising and education
Issue 2:
Overcoming a lack
of confidence
Make models more transparent/
understandable
Make models more transparent/
understandable
Improving the measurement
and communication of uncertainty/
limitations of models
Improving the confidence level/
reliability of basis data and
results of model
Issue 3:
Interfaces
Better user interfaces are needed Better user interfaces are needed
... for better access to and
communication of results
... for better access to and
communication of results
... for public/stakeholder joint-use
and development of models
... mainly for public/stakeholder
education
Issue 4:
Integration of
models
Integration of socio-economic
models are needed
Integration of socio-economic
models are needed
Integration of human dimensions
i.e., human impacts on
environment; environmental
impacts on humans; human
behaviour & decision making
Integration of ... river and lake
quality models; land use and
agricultural models; models to
support whole WFD process
e.g., gaps analysis; cost-efficiency etc.
Integration within one composite
model or through the use of many
separate models providing alternative
results?
One model, one answer, please?
Water Resour Manage
introductions in terms of how their approach or model supported the implementation of the
WFD. All participants also presented their major issues of interest for the workshop. The
project team finally presented a summary of the previous two workshops and river basin
meetings and introduced the major issues developed over this preceding period of time.
The participants were then divided into four groups to consider the following issues:
(1) interaction of participatory modelling and management; (2) improving confidence in
models by improved uncertainty analysis; (3) interfaces to support communication with
stakeholder; and (4) integration of sociological, ecological and economic models into
management processes. As with previous workshops, the groups were moderated by
external moderators using brainstorming and meta-plan cards techniques. Results peer-
reviewed by each group were then recorded in the results protocol of the workshop.
To increase the ‘intercommunity’exchange and to help the moderated group discussions
focus on issues raising from concrete tools, time for modellers’demonstrations of model-
based tools that were of relevance to the issues was allowed. This allowed all managers to
get more information on relevant tools for use in the break-out discussions. It also allowed
the modellers to get focussed criticism on their tools.
At the end of the workshop, each group presented their findings on the issues to the
workshop plenary. General findings were discussed and recorded in the results protocol of
the workshop.
After the workshop, the break-out groups were invited to develop their discussion results
into a scientific publication. Two of these papers are published in this special issue (see
paper from Horlitz 2006 and paper from Brugnach et al. 2006) Full workshop results are
available in Hare (2005).
3 Analysis of Results of the Requirements Elicitation Phase
From an analysis of the results from the elicitation phase, it was possible to identify seven
main assumptions belonging to the researchers that can be used to illustrate the gaps
between the two communities. As we intend to stir thoughts among our readers, we have
phrased these assumptions as starkly as possible, formed from some of the most potentially
Table V Linkages between workshop issues and the assumptions
Issues from first three workshops Assumptions used in this paper (researchers’perspective)
Issue 0: Role of models ASSN 1: Models can play a central role in water management.
ASSN 2: If you develop a model for a case study this tool can then be
transferred to the water managers’river basin target site.
Issue 1: Participatory modelling
and management
ASSN 3: If you get the user involved in designing the model within a
participatory modelling exercise, then the user will use the model.
Issue 2: Overcoming a lack of
confidence
ASSN 4: If lack of confidence in models is to be overcome, then
investment has to be made in carrying out and communicating
uncertainty analysis.
ASSN 5. Predictive models cannot be built; therefore decision makers
will accept scenario-based tools offering a range of possible outcomes.
Issue 3: Interfaces ASSN 6: Work on improving the interface and the usability will increase.
Issue 4: Integration of models ASSN 7: Large integrated models are needed, providing decision makers
with many insights into different sectors and scales.
Water Resour Manage
controversial yet persistent positions of the researchers. Understanding these positions as
well as the arguments for and against them, we assume, is of importance if both the
research and water management communities are to improve the application of model-
based tools in water management.
To elaborate on the findings of the elicitation phase, we also refer to scientific publications
and EU project documents. Project documents included both the publicly accessible Web site
(see last section for Web sites of all projects mentioned in this article) as well as official
project reports. These documents were considered more easily accessible (e.g., to water
managers) than articles published in scientific journals which is why we took them as
seriously as peer reviewed publications.
Table Vshows the assumptions and how they are linked to the main issues of the first
three workshops.
In the following section, the assumptions will be elaborated on, comparing the
perspectives from the water managers and the researchers. The comparative analysis will
lead to recommendations in the final section as to how the gap between water managers and
researchers might be bridged.
ASSN 1: Models can play a central role in water management.
Researchers at the 1st Method workshop perceived that model-based tools have a major
role in water managers’decision making. There were a number of different roles considered:
tools for information provision; scenario simulation platforms for testing measures and to be a
mediator for stakeholder participation and discussion; support creating and inventory of river
basins as required by WFD; work as educational tools; helping communicate unpleasant
outcomes; communicate policy uncertainty; or even to break down power structures.
Researchers see a stronger need for more participation of users, especially water managers in
model design. This leads in turn to a bigger future role for models. Researchers expected also
that multiple models of the same management system will become the norm to contribute to
the diversity of tasks and aspects in IRBM.
As shown in the introduction, this important role of model-based tools is confirmed by other
scientific literature but there has been also already extensive literature showing that many of the
tools developed for direct (decision) support are not suitable and have not been close enough
designed along the requirements of the user group (e.g., McIntosh et al. 2004; Robinson and
Freebairn 2001; Ubbels and Verhallen 2000;Vonketal.2005; Greeuw et al. 2000).
This confirms the results from the elicitation phase that models were not as widespread in
day-to-day decision making as was initially expected. Water managers at the 1st Policy
Workshop stated that so far models are mainly used for management/ operational issues,
testing of measures and their outcomes, mainly physically based models for water allocation,
water quality. If they are applied, they have been designed to answer specific questions.
During our river basin meetings, we asked each water manager to rank the importance of
different sources of knowledge in the decision making process. Information from models
and the public consistently ranked low in importance (as low as the use of information from
the public), below the use of their own experience and the expertise of in-house and
external experts. This is a very important issue for modellers to take into account when
trying to find users for their models. However, note that the results are obtained from
discussions and interviews with water managers. An analysis made by the project
HarmoniQuA showed a higher confidence in models but the respondents were mostly
external experts/ consultants, supporting water management (Old et al. 2005). This might
show an indirect potential for models to be of support through being used by external
consultants supporting water managers, rather than the water managers directly.
Water Resour Manage
Being only one of many sources of information, and due to the fragmented nature of
water management authority duties, model-based tools are expected to fill specific gaps.
The models applied by water managers in the river basin meetings tended to be simple, data
rich (having a solid data basis), spatial, hydrological models used for specific tasks. These
tasks included water quality assessment, precipitation simulation, risk assessment, flood
management and water abstraction assessment. Among the top-two priorities of the water
managers (see Table II) were the need for information provision tools for public and
stakeholders as well as a management tool for competent authorities for implementing
specific tasks of the WFD. It was also expected that models could act as supporting material
for decisions, rather than providing the decision.
Actual models that were used by the water managers at the river basin meetings included
MIKE-11, MODFLOW, IHACRES, SIMCAT, PEGASSE and MODCOU as well as self-
constructed empirical and scale models.
When asked about barriers to model use or to the reuse of existing models, the water
managers mentioned the high costs of re-applying a model in their management area. The
lack of validation of the desired model in their management area was also a major issue.
Here, the lack of available data for their management area and lack of tool maintenance
provided by the researchers played an important role. This also appeared in the top two
requirements of the water managers (see Table II). Researchers in the 1st Method Workshop
linked these constraints to the need for improved documentation of tools. Other research
confirms these points: More then the half of professionals claimed in a recent analysis,
carried out by HarmoniQuA (Old and Packman 2005), that their models were re-used with
new data, but that the documentation of models was often in a rather poor state, making it
difficult to actually re-apply them. Consequently HarmoniQuA has developed a computer
based modelling support tool to provide a user-friendly guidance and quality assurance
framework that will contribute toward enhancing the credibility of catchments and river
basin modelling (Scholten et al. in press).
The barriers to use go far deeper, however, than simply providing better documentation
and guidance. It became apparent during the river basin meetings that models produced by
scientists in academia faced competition from those developed in-house and most crucially,
from those developed and sold by consulting software companies. The advantages for in-
house models were that their development could be controlled and focussed on the target
area. They would also be well understood and documented by the organisation. The
principal disadvantage was that they would consume large amounts of resources during
their development. The academic model therefore has got the advantage that the resources
consumption for their model development is very low and the fact that open, non-
proprietary software is used in their development. Unfortunately, as it was pointed out, the
academic model is severely disadvantaged in comparison to software consultancies in three
ways: (1) that the researchers offer lower levels of control to the water manager during tool
development; (2) they often provide less quality documentation and (3) they can provide, in
comparison, negligible levels of maintenance support afterwards. A number of cupboards in
the organisations of the water managers visited had therefore turned into graveyards for
unused, un-maintained academic models.
Another constraint identified by water managers in the river basin meetings was the lack
of communication between water managers and researchers. Both groups are not talking to
each other enough, especially linked to the first definition of questions to be approached by
a model. Neither are they sharing the same language for expressing results or the require-
ments of models.
Water Resour Manage
As for the role of models, it looks like as if the problem as to how to increase the use of
model-based tools in water management may have reached an impasse. Water managers,
due to lack of resources and time have a need for better model-based tools as a means to
support increasingly complex decision making but that same lack of time and resources
prevents them from being able to invest time and resources into the models, if the
researchers cannot invest the time in providing the type of non-technical requirements (such
as maintenance, model re-calibration; documentation, confidence, usability) they require.
Many of these issues will be revisited in consideration of the next assumptions.
ASSN 2: If you develop a model-based tool for a case study, this tool can then be
transferred to the water managers’river basin target site.
Case studies are often carried out in research projects to extract lessons learnt and any
models produced are meant to be transferable for other cases. Among researchers at the 1st
Method Workshop there was scepticism whether existing models (particularly ones which
had been participatorily developed) could be reused to support participatory decision
making outside the area that they are actually developed for. At the Synthesis Workshop we
observed often that modellers would, being asked if their tool might be applicable in a
different basin, answer, “in principal yes”. It was suggested by the researchers that the water
managers can still take the generic concepts of the tool with them and use it in building and
setting up their own tool specific to their own area. This agrees with the researcher’s
attitude at the 1st Method Workshop that they thought that we had got the models we need
for IRBM though there would be still a need for a better resolution of space and time within
models. The difficulties of technical transferability of the model were acknowledged not
only by the researchers. Also, water managers state that the transfer and re-application of
model-based tools fails often due to the fact that water managers have limited time and
money to invest in what may sometimes be an enormous task to collect the data and
parameterise it for their own river basin. There is still a lack of ideas/ methods that allow
water managers to use model-based tools designed for one location analogously with
respect to their own location.
Also other research showed that modellers state often that one of the major difficulties during
the development phase is to get access to and to harmonize the needed data (Old and Packman
2005). Thus, re-application of model-based tools would most certainly face the same
difficulties if not worse as for different basins there exist most times different data formats.
ASSN 3: If you get the user involved in designing the model within a participatory
modelling exercise, then the user will use the model.
Researchers suggested that getting water managers actively involved in the design,
development and testing of models would increase the chances that the model would be
used by those co-developers. This was seen as an approach to improve the understanding
and the perception of the model results. User education and better interfaces were required.
Many of the researchers at the 1st Method Workshop saw modelling as integral part of the
management process. They were interested to better interweave the different activities in
the modelling process and water management. The idea is that when the model comes to
the water managers as completed, perfected science, the water manager has no way of
assessing its validity and needs therefore stringent reassurances from the developers, which
will, on the evidence above, not be available. When the model comes to the water manager
through the familiarity of having helped to develop it, the model will be treated in the same
way in which all other familiar information sources with whom they have worked for a long
Water Resour Manage
time (e.g., colleagues, consultants), would be treated, i.e., with a known personal and
intuitive understanding of the validity of the information source.
The activity of getting users deeply involved in modelling is known as participatory
modelling. Participatory modelling has developed out of a synthesis of the group model
building (see Vennix 1996) applied in operations research and the participatory
management concepts developed for natural resource management (see (Mostert 2003)).
Its use and claims for its benefits are widespread in the research literature (Pahl-Wostl 2002;
Lynam et al. 2002;D’Aquino et al. 2003; Hare et al. 2003; Letcher et al. 2004).
So far, many of the European projects try to improve the relevance of their work by getting
stakeholders to give advice during the project. For example, Mulino, Transcat, HarmoniQuA
and HarmoniCOP have followed this approach quite closely: they involved water managers
as stakeholders during the development of their products and respectively as external
advisors. Also, the approach for participatory scenario building has been used (see ASSN 5).
In general, however, water managers showed little enthusiasm to spend much time on
their own education about the correct use of models. They claimed already that their limited
resources don’t really allow a lot of additional activities, like participating in a modelling
process. The water manager has little extra time to find out about different models and
especially little time to use more than one model. If the manager has a modelling team in-
house then this becomes more of a possibility.
ASS 4: If lack of confidence in models is to be overcome, then investment has to be made
in carrying out and communicating uncertainty analysis.
Water managers see a good potential to use models but they displayed an obvious lack of
confidence in them –even in some of their own in-house models. This obviously is strongly
linked to the liability and responsibility the water managers have for their decisions.
Researchers acknowledge that users may lack confidence in the models they develop.
One problem researchers see is the lack of data for validating complex management system
models apparently required by the WFD. In the 1st Method Workshop, it was contended
that this lack of confidence can be overcome by investing resources into better forms of
model uncertainty analysis and its communication. This assumption is supported by
Refsgaard et al. (2005) who claim that water managers will only be able to act adequately
under different situations of uncertainty “when they know how uncertain the available
information is and when they know how to incorporate this in their decision making”.In
Refsgaard et al.’s guidance for professionals involved in modelling water resources
management, methodologies and tools for uncertainty assessment are described which may
be used in the decision making process, allowing decision makers to take sensible decisions
on how to act under a situation of uncertainty.
Water managers agree with the researchers that a trustable tool avoided misleading the
users by clarification of limitations and uncertainty of the model. However, the attitude of
water managers towards uncertainty in models is more complex.
To understand this attitude their context of decision making has to be understood. When
communicating decision making information to policy makers, water managers tend to
simplify their message. There is no need from water policy makers to be told about all the
uncertainties in the decision. With the public, uncertainty is not tended to be communicated.
Therefore there is little need for communication of sophisticated uncertainty analysis in the
results. However, the water manager has to make a sound decision about what the best
solution/ recommendation is. Therefore, s/he needs an option to assess the risk of making
the wrong decision. The most frequent requirement linked to confidence in a model
Water Resour Manage
mentioned from the water managers was a solid scientific basis of the model which could
be proved by providing statistical analysis. An important requirement from the water
managers’workshop was that the models should be based on very accurate data sources
and that the results of models should have accuracy levels within ±10%. In terms of the
water managers’own use of uncertainty analysis, the assessment of the risk of making the
wrong decision can only be done when inaccuracies in data used in their models are
recorded. At the river basin meetings it appeared that their data was not recorded with
uncertainty levels. If the water managers do not formally take data uncertainty into account
in their decision making, in this way, then it seems optimistic to assume that they will be
looking for sophisticated uncertainty analyses in their modelling.
Therefore, it can be seen that, for water managers, a lack of uncertainty analysis is not
the reason for the lack of confidence in the models; it is the lack of certainty contained in
the model itself. What the water managers seem to want is that the uncertainty in models is
significantly reduced, not simply better represented. Obviously they do not wish to apply
these standards to the recording of their own data, but it is their own data and they can
assess the quality of it themselves. When receiving data and models from a second party,
they do not have that implicit confidence –they need extra evidence for its quality.
One could argue that the expectations about model accuracy on the part of the water
managers are unrealistically high. However, model-based tools are not singled out for
special treatment in this regard. This need for accuracy is also demanded from any other
second party sources of information. External experts, for example, from consultancies are
also required either to provide accurate advice or to have very good liability insurance
should their advice turn out to be flawed.
Beyond that, water managers identified several dimensions of confidence in a model: in
the mathematics of the model; in the reliability of the model in terms of predicting
processes in nature; use & familiarity; simplicity and complexity. It includes convincing
people of the merits of a model and the trust that comes with experience of a model and
with modelling. Water managers ask for transparent models with their limitations clear and
communicated. However, water managers also acknowledged at the 1st Policy Workshop
that it will be nearly impossible to make a model totally transparent. Water managers
suggested that stakeholder’s trust and competence in a model can be generated by getting
support and involvement by a particular key stakeholder. Naturally this only works if
somebody in the stakeholder community has (a) confidence in the model and (b) if s/he is
well trusted by that community.
Finally, generally, the researchers from the 1st Method Workshop also found it difficult
to deal with water managers who don’t see the technical limits of applying their model-
based tools in decision making. A lack of understanding in the impact of and water
managers’use of uncertainty in their decisions was found on the researcher’s side. Probably
some training of researchers is required in this area, if they are to sensibly deal with water
managers’requests for “one answer, one model please”.
ASSN 5: Predictive models cannot be built; therefore decision makers will accept
scenario-based tools offering a range of possible outcomes.
In the face of increasing management complexity and subsequent uncertainty inherent in
integrated model building, the researchers, in the 1st Methods Workshop, felt that models
would have to change from being considered as predictive machines and into being tools
for awareness raising and exploration of management issues. Generally this view has been
mirrored in the literature. In certain fields, in which modelling techniques are applied,
Water Resour Manage
doubts are being registered as to the possibility of being able to predict anything with
models in complex systems and in turn they are lowering their own expectations of models.
Prediction is often no longer a requirement. In ecology, for example, Judson (1994) has
called the search for predictive general ecological models as a quixotic quest for an illusory
‘holy grail’, and has suggested the attempt be abandoned. Gilbert, a leading proponent of
the use of agent-based simulation in social sciences, declares (Gilbert 1998) that the
practice of using models for prediction is a bad use of simulation.
Alternative requirements for the purpose of models are therefore being explicitly
developed in various fields. Models have begun increasingly to be seen as instruments for
aiding the understanding of reality, not for replicating it. For example, in the field of
Artificial Intelligence modelling, Winograd (1990) has called for model-based tools to
“support the discourse that creates reality, not the reality itself”. Almost identically, but in
the field of systems research, Checkland (1995) proposes the Soft Systems Methodology
that rules out the modelling of reality in favour of “models relevant to the debate about the
reality”which can aid the “process of enquiry”about the world. Oreskes et al. (1994)
representing the earth sciences, redefine the purpose of models as “representations, useful
for guiding further study”. In the social sciences one should look for such models to
provide “understanding”,“insights”and “new choices”(Gilbert 1998). Due to the inability
to validate a predictive model in these systems (i.e., to use classical validation techniques),
the requirement of Artificial Intelligence models applied in complex systems is therefore
moving away from the predictive oracle or accurate classifier of old, towards becoming
ideas generators or “epistemological devices”(Checkland 1995). Checkland goes on to
write that “All this means is that model validation is not now an issue of great moment”.
Among the research directions from the 1st Method Workshop (see Table III) were three
that linked to scenario research as an alternative to prediction. This is also reflected in EU
research on water management tools, in which modellers offer alternative scenarios
covering different settings. These settings are describing what could be, rather than what is
or will be (e.g., Tisza River Project, TEmpQSim, TransCat, and CLIME).
However, the difference in opinion between the two groups seems not to be that models
are perceived as prediction by water managers as the latter stated at the Synthesis Workshop
that “prediction is not reality”. In any case, they had strong interest in what if–what then
scenario models (see first requirement from category 11, Table II). The problem seems to
come from the fact that, whilst they were aware of the unrealistic ambition behind the
following requirement, water managers stated that they were strongly interested in “just one
good model per problem made up from the best parts of all the alternatives”(see Table V)
which should provide unambiguous results. The problem therefore seemed not to be to
differentiate between prediction and scenarios but that water managers are requiring a level
of model validity and of a lack of ambiguity in results that modellers are increasingly
disinclined to provide.
Even if modellers create scenario-based tools the water managers still demand that the
scenarios are based on good, valid, data-rich models. Supporting this view, a study from the
European Environment Agency, states that “in order to be useful for decision-making
purposes, scenario studies should provide recommendations for decision-making ... [whose]
Usability criteria ... [include] whether the recommendations are concrete, challenging,
realistic and practical ”(Greeuw et al. 2000; bold added for emphasis by authors).
This means that doing validation (something that less than 50% of professional
modellers do according to Old and Packman (2005)) cannot be avoided simply by moving
into providing scenarios rather than predictive results. Validation still must be done or else
Water Resour Manage
another conceptual solution is required. Simply deferring classical validation (Doran 1996)
won’t help as neither will throwing it out completely (Checkland 1995). Also, calling on
water managers to see their tools as providing insights, does not solve the problem as to
how to convince the water managers that the models behind the scenarios provided are
valid and ‘good enough’to consider as serious possible futures. Having concrete reasons
for trusting the scenarios is even more important when it comes to the power of scenarios to
affect public perception of a problem. The over-emotive use of un-validated scenarios in the
public arena can lead to understandable water manager concern.
Similar as in ASSN 3, one possible solution to the lack of confidence and trust in models
is to get the water managers involved in the development of scenarios. Participatory
scenario development including water managers as well as other stakeholders may
contribute to an enhanced understanding of interactions in a basin. In addition, the
scenarios would probably be more appropriate for water managers’requirements. This
would be because the water managers would not only make sure that the scenarios are
relevant for the measures they need to assess, but would also make sure they understand the
limits and validity of them. This would create an important change in the conceptualisation
of the validity of the model on the part of the water manager. However, a major barrier to
this is linked to the issue of time and resources’constraints mentioned under ASS3,
meaning that even this effort would not come for free to the water managers.
ASSN 6: Work on improving the interface and the usability will increase.
The researchers at the 1st Method Workshop saw the need for model-based tool
improvement relying on visualizing better the relevant contents of the model to the different
stakeholders. They also considered the interface as of major importance in making the
model accessible and transparent to the user. Often, however, experts will be necessary to
explain the model. In the Synthesis Workshop such experts were called model moderators.
The water managers also showed strong interest in easy understandable, graphical
interfaces. Those interfaces should include on-screen cross referencing between cause and
effect in modelled processes. Also, interactive devices allowing all participants to access the
models and to design scenarios were asked for. These features were especially of interest
where models should be used in participatory processes. Here, also indicators should be
chosen which mean something to stakeholders and public.
However, water managers stressed also that they had neither the time to train the public
to use such tools nor to train themselves to use complicated systems. Models should be easy
to use. The researchers commented this as not being not easy to achieve when modelling
complex integrated river basin systems. Improved interfaces, with sophisticated help
facilities could support learning to use complex models easier.
Additionally, for the use of models in their internal water management processes, the
water managers pointed out that a large obstacle to greater access is not an improved
interface per se, rather access to the hardware and software. Most authorities don’t have
enough money to procure and provide computers to allow wider access to the models that
run on them. Also there are issues of copyright associated with propriety software that
limits the use and adaptation of software to fit the information provision needs of cash-
strapped authorities. However, non-commercial software will most probably not solve the
problem of lack of maintenance of a model.
ASSN 7: Large integrated models are needed to provide decision makers with insights
into different sectors and scales.
Water Resour Manage
Confronted with the integrated river basin management approach of the WFD, water
managers are being encouraged to consider the integration of many things, such as ecology,
economics, social, physical, chemical, and biological impacts on water quantity, quality.
Since Article 4 of the WFD asks for the development of plans for the achievement at the
river basin scale of good ecological status for all water bodies based on ecological/
biological, chemical and morphological indicators it means that, as well as having to tackle
ground and surface water issues, land use changes also need to be considered within
borders defined on hydrological rather than administrative criteria. Traditional governmen-
tal water management is organised differently. Often, water quality, water quantity,
groundwater or surface water, diffuse or point pollution, shipping, flood protection,
environmental economics, land use planning and hydro-power is distributed in separate
departments, even in different–sometimes competing–ministries but in almost every case
with different persons working on it. Water managers are used to be experts in quite specific
aspects, able to respond to the specific needs of the different public users and interests.
However, with the WFD the focus shifts from focusing on specific details to integration of
different sectoral aspects. The water managers’interest for integrated models can therefore
be seen as a response to the challenges of integration in water management itself.
Water managers are also expected to facilitate river basin planning and the
communication of complex problems and their solutions in the different river basins. This
requirement is also expressed in some European or national guidance documents (e.g., EU
2002,2003;LAWA2003).
Researchers at the 1st Method Workshop approached integration in a comprehensive
way, befitting the comprehensive nature of integration required by the WFD. The
integration is grandly and broadly across sectors and –where possible–scales in the form
of a large Decision Support Systems (DSS). Some went further, proposing in addition, the
integration of behavioural models of actors in the river basin using, for example, agent-
based modelling (see Bousquet and Le Page 2004 or Hare and Deadman 2004 for a
review), as a way of capturing the necessary socio-psychological elements that need to be
considered in testing policy options. This led to tools that are data-hungry and thus very
difficult to transfer and parameterise to the manager’s target site, let alone to assess the
validity of the model (see also ASSN 4).
On the other hand, water manager’s requirements were less comprehensive in some
ways, still focussing on specific needs. Beyond showing the physical, ecological and
chemical interactions in river basins, water managers required models to be of support for
the assessment of different measures, in terms of cost efficiency as well as ecological
impact (both aspects were top-two requirements at the 1st Policy Workshop (see Table II)).
Unfortunately, model-based tools supporting the assessment of the introduction of
economic measures (taxes, charges...), legal and administrational instruments, or the
introduction of best practice guidance are not yet readily available from the research
community. This need was evidenced by the limited amount of economic modelling
represented in the model-based tools discussed and brought to the model makers and
Synthesis Workshops (Hare 2004a), as well as the second policy workshop (Borowski and
Heeb 2005).
The different perspectives of water managers and researchers can be also understood in
the way how requirements for integration of model-based tools are expressed. Gilbert et al.
(1998), for example, describe how policy makers specify their needs always in terms of policy
fields in which economics (e.g., cost effectiveness) of policy measures play an important role.
That is, economics is always part of their reckoning whether they are considering water policy or
Water Resour Manage
agriculture. Researchers, on the other hand, focus their interests in terms of sectors, considering
economics as separate from ecology, agriculture, society and hydrology.
One agreement on integration, however, was found on the need for the integration of
economic aspects into tools.
4 Conclusions & Recommendations: How Can the Role of Tools/Models
be Improved in River Basin Management?
Summarizing the preceding sections, the differences between the perspectives from water
managers and model-based tool developers on the reasons why the application of model
based tools in water management is relatively low, seem to be rather small at first. Both
communities name similar issues –integration, interfaces, confidence. To explore why the
research funding provided by the European Commission has not had the expected impact,
we have studied the misunderstanding between researchers and water managers.
Water managers and researchers are both experts in their own fields but they are driven
by their own goals and methods of working. Researchers do not consider themselves as
producing the ‘directly applicable’products (with all the implications this has on providing
maintenance, reliability, topic focus, etc.), that the water managers expect they should be
getting. More specifically, the misunderstanding is a result of structural differences in
research and water management that ultimately lead to differences in attitudes to what we
have identified as seven major assumptions.
Water managers involved in our elicitation phase can be described as to have conflicting
interests to balance. They have to manage a resource properly and in accordance with legal
constraints and opportunities. They have to justify decisions to superiors and the public
(and perhaps to a court of law). Water managers need also to get used to integrated river
basin management. If they need to learn more about the interactions of physical, ecological
or any other factors, let alone the economic and social impacts, they need to learn about it
now i.e., as soon as possible. Water managers have a time horizon influenced by the
deadline of legal frameworks, annual reports and sometimes public elections (1–15 years).
Water managers, generally, have limited time to invest in models: either on being ‘guided’
on their use; being involved in their development; or reapplying tools developed in their
own basin.
The researchers involved in our elicitation phase showed strong interest that their work
is of support to practical water management. However, it became also clear that the
researcher’s first interest is to do research: to explore deeper into their field of interest such
as climate change, groundwater degradation, ecological modelling, and in a lot of cases
they are eager to provide their insights to the policy community, to spread new perspectives
and to influence society. Researchers need to produce scientific results on the focus of their
work which might include lower scientific quality in areas which are not directly linked to
the project’s objectives. Often, they have to legitimate their efforts and products to their
funding bodies. Researchers’time relation is linked to funding cycles and project duration
(three to five years).
These structural differences lead to sometimes very different answers towards the
question how role model-based tools can support water management. First of all, while
researchers saw them central to water management (ASSN 1), water managers consider
model-based tools as only one source of information. Both saw potential to improve this
role. However, researchers favoured better documentation, participatory modelling,
Water Resour Manage
improved interfaces and extended uncertainty analysis as trust building approaches (ASSN
3, 4, 6). Water managers were very reluctant to spend more time or money on model-based
tools. They in contrast stressed their need for a wider range of non-technical requirements
(including maintenance contracts, access to cheap software and hardware, control over
development). Water managers stressed also their preference for a solid scientific database
of models instead of for better documentation of the unsatisfying state of the database. This
became also clear in the role water managers saw for scenario tools. Those tools were
considered an important approach to assess different options but not as an alternative to use
models for predicting (because they wouldn’t use model-based tools for that). In total this
showed a mutual misunderstanding about the resources (time and funding) available to each
community to develop and use model-based tools that should be applicable. Also the
transferability of model-based tools from case studies to other basins (ASSN 2) was
assessed differently.
Finally, the need for integration (ASSN 7) was assessed differently from the different
groups. While researchers often voted of the full spectrum of sectoral integration to reflect
the complex interactions in river basins, water managers’requirements were in general
closely related to needs such as the integration of cost effectiveness of measures into river
basin planning.
This means that simply improving communication between the two will not solve the
problem alone. If left ignored and unresolved, these structural and attitudinal differences
threaten to limit the present and future application of model-based tools. The important
question is which community has the most need to adapt itself to the other? Unfortunately,
for the researchers, the use of model-based tools is not the job of water managers, they have
many important decision making activities to carry out, with or without such tools.
However, there might be the chance that model-based tools can improve and facilitate
participatory, integrated river basin management. Thus, we conclude with the following
recommendations.
4.1 Recommendations
4.1.1 Improving Understanding of the Management Process
The research community needs to better understand the decision making process that they
are trying to support as well as the constraints and requirements of the water managers. To
this end, one recommendation is that a future EU funding area should be for tool designers
to be funded to learn more about management processes, e.g., to shadow water managers in
their daily work. Another option might be for specific projects to adapt their timeline (and
duration) with the managers’and to develop a joint process for participatory modelling and
managing which implies that the modelling process is designed as part of management.
Thought needs to be given to how to compensate the managers and their organisations for
their time taking part. Also, activities on other continents like the CREM (The Council for
Regulatory Environmental Modeling; see http://cfpub.epa.gov/crem/) can be taken as
example for how to bring the models closer to management.
4.1.2 Model-Based Tools as Part of a Social Learning Management Process
The results have also shown that integration may be one of the most important and
challenging features in both modelling and managing. Both communities are still used to
Water Resour Manage
fragmented thinking. Integration asks in general for a generalization as the amount of details
can not be processed or understood. The hope for a technical solution like a detailed database
linking well validated, model-based tools will have to be kept on for some time. The
dependency of the data formats on the collecting authorities and the resulting diversity of data
is only one factor here. So far, model-based tools have to be considered as only one method
among many to support the river basin management. Their role might be much bigger if they
facilitate mediation between different experts and interest groups (and support this way social
learning) to shift the focus from the individual interests to integrated thinking. Therefore
approaches to promote social learning as featured in HarmoniCOP (2005) and in Pahl-Wostl
and Hare (2004) should get a more prominent place. Social learning encompasses a structured
approach facilitating participatory decision making by improving mutual learning of all
involved actors. It acknowledges that different interests may be linked with different
knowledge and the time which experts need to integrate their knowledge.
4.1.3 Involving Software Consultants as Carrier of Research Results to Water Management
The attempt to bring together the research community and water managers should be
continued. However, it must change its emphasis. The integration of small and medium-
sized enterprises in research projects in European research programmes need to focus on
their potential as permanent interface between researchers and water managers. This means
that already from the very beginning the tool should be developed in close link to the water
management questions. This link could early be established and maintained by firms with a
commercial interest of success. In addition, only software development consultancies are
going to be able to provide the quality of tool combined with the non-technical
requirements such as maintenance contracts, reliability, documentation, training, product
liability etc. These services have to be paid. If available they will allow a water manager to
take the risk of using the tool to support his or her legal responsibilities. Universities are not
set up for this type of service and nor do they expect that they have to provide it. This does
not mean that Universities should stop developing model-based tools.
4.1.4 Transferability Through Analogy: Supporting Cognitive
Instead of Technical Transferability
The problem of transferability can also be tackled. For technical transferability, it is too
optimistic to expect that sufficient money and time will be provided to parameterise and test
the model in the new target basins. If current model-based tools are ever to be used in time
for the WFD, rather than try to achieve technical transferability, it could be better to
encourage a more ‘cognitive transferability’. That is the water managers will have to learn
how to transfer the findings of current models, applied to analogous river basins, to their
own basins, rather than transferring the models themselves. In turn, modellers will have to
learn how to support this activity.
4.1.5 Funding for Providing Non-Technical Requirements
Modellers have to consider both the non-technical as well as the technical requirements of
their tools, especially their maintenance. This means that there needs to be funding not only
for tool development but also for maintenance of that tool for current and future users.
Water Resour Manage
4.1.6 Keep Listening
Finally, the following exchange between two participants of the Synthesis Workshop is a
salutary lesson to all of us who are trying to support water managers in their work: keep
actively listening to your users!
Water
manager
I have a computer tool used in my administration which I want and need to use for my
management work but I cannot understand how to use it. I need help. Can such tools be
better designed to help me use them?
Modeller Look, that is not a problem –you’ll find that anyone who is motivated enough will be able to
use it.
5 Internet Sites of Projects Referrer to:
Advisor (Integrated Evaluation for Sustainable River Basin Governance) http://ecoman.
dcea.fct.unl.pt/projects/advisor/
BMW (Benchmark Models for the Water Framework Directive): http://www.ymparisto.
fi/default.asp?node=11687&lan=en
Clime (Climate and Lake Impacts in Europe): http://clime.tkk.fi/
EuroHarp (Towards Harmonised Procedures for Quantification of Catchment Scale
Nutrient Losses from European Catchments) http://www.euroharp.org
Harmoni-CA (Harmonised modelling tools for integrated basin management): http://
www.harmoni-ca.info
HARMONICOP (Harmonising Collaborative Planning) http://www.harmonicop.info
HarmoniQuA (Harmonising Quality Assurance in model based catchment and river
basin management): http://www.harmoniqua.org/
HarmoniRiB (Harmonised techniques and representative river basin data for assessment
and use of uncertainty information in integrated water management): http://www.
harmonirib.com/
HarmonIT (IT Frameworks): http://www.harmonit.org/
MERIT MULINO (MULti-sectoral, INtegrated and Operational decision support system
for sustainable use of water resources at the catchment scale) http://www.feem.it/web/loc/
mulino/
SMURF(Sustainable Management of Urban Rivers and Floodplains) http://www.smurf-
project.info/
TempQSim (Evaluation and improvement of water quality models for application to
temporary waters in southern European catchments): http://www.tempqsim.net/
Tizsa River (Real-life scale integrated catchment models for supporting water- and
environmental management decisions): http://www.tiszariver.com/
TransCat (Integrated water management of transboundary catchments): http://www.
transcat-project.net/
W.A.D.I. (Sustainability of European Irrigated Agriculture under the WFD and Agenda
2000) http://www.uco.es/investiga/grupos/wadi/
Acknowledgments We would like to thank all participants in our workshop and river basin meetings for
their support in our work. We also thank the European Commission for the financial support of Harmoni-CA
(Contract no [EVK1-CT1-2002-00192]). The views expressed herein are the authors’own and do not
necessarily reflect those of the European Commission. Also the European Commission is not responsible for
any use of data and information appearing herein.
Water Resour Manage
References
Borowski I, Heeb J (2005) Evaluation of tools/models for integrated & participatory river basin management
focusing on the interaction with agricultural management. Report of 2nd Policy Workshop of Harmoni-
CA/WP5. Harmoni-CA report HCA-WP5-2005-05/Final Version, Osnabrueck, University of Osnab-
rueck. http://www.harmoni-ca.info
Bousquet F, Le Page C (2004) Multi-agent simulations and ecosystem management: a review. Ecol Model
176(3–4):313–332
Brugnach M, Tagg A, Keil F, de Lange W (2006) Uncertainty matters. Computer models at the science–
policy interface. Journal Water Resources Management Special Issue: Participatory Methods for Water
Resources Management.
Checkland P (1995) Model validation in soft systems practice. Syst Res 12(1):47–54
D’Aquino P, Le Page C, Bousquet F, Bah A (2003) Using self-designed role-playing games and a multi-
agent system to empower a local decision-making process for land use management: the selfCormas
experiment in senegal. JASSS 6(3). http//jasss.soc.surrey.ac.uk/6/3/5.html
Delbecq AL, Ven AHVd, Gustafson D (1975) Group techniques for program planning: a guide to nominal
group and Delphi processes. Scott, Foresman & Co., Brighton
Doran J (1996) From computer simulation to artificial societies. Trans Soc Comput Simul 14:69–77
Dørge J, Windolf J (2003) Implementation of the Water Framework Directive –can we use models as a tool
in integrated river basin management? International Journal of River Basin Management 2(1):165–171
Edmüller A, Wilhelm T (2002) Moderation. Haufe, Freiburg
EU (2002) Guidance on public participation in relation to the Water Framework Directive. Active
involvement, consultation, and public access to information. Prepared in the framework of the Common
Implementation Strategy of the European Commission and the EU Member States. http://forum.europa.
eu.int/public/irc/env/wfd/library
EU (2003) Guidance on the planning process. Best practices in river basin planning. Prepared in the
framework of the Common Implementation Strategy of the European Commission and the EU Member
States. http://forum.europa.eu.int/public/irc/env/wfd/library
European Commission (2004) Water cycle and soil-related aspects. Water technologies: results and
opportunities. Research Directorate-General. Global Change & Ecosystems, Brussels. http://europa.eu.int
Gilbert N (1998) The simulation of social processes. Edited transcript of talk given at SMAGET Conference,
6–8th October 1998, Clermont-Ferrand, France
Gilbert A, Goosen H, Tol RSJ (1998) Water and environment. Scoping study for the European forum for
integrated environmental assessment. Institute for Environmental Studies, Amsterdam
Glaser BG, Strauss AL (1998) Grounded theory: Strategien qualitativer Forschung. Verlag Hans Huber
Grabowski M, Massey AP, Wallace WA (1992) Focus groups as a group knowledge acquisition technique.
Knowl Acquis 4:407–425
Greeuw SCH, Asselt MBAv, et al (2000). Cloudy cristal balls –an assessment of recent European and global
scenario studies and models (experts’corner report prospects and scenarios no 4). Environmental Issues
Series. E. E. Agency, Copenhagen, Denmark
Hare M (2003) Plan for integration of the First Policy and Method Worskhops for Workpackage 5 of
Harmoni-CA. Harmoni-CA Document HCA-WP5-2003-Re01. University of Osnabrueck, Osnabrueck
Hare M (2004a) Modeller’s recommended research directions for models to support the participatory
elements of the Water Framework Directive. Report on the 1st Modellers’& Scientists’Workshop of the
Harmoni-CA WP5 Project, Harmoni-CA report HCA-WP5-2004-Re02/Final Version. University of
Osnabrueck, Osnabrueck (available at http://www.harmoni-ca.info)
Hare M (2004b) Policy makers’requirements for models to support the participatory elements of the WFD.
Report on the 1st Policy Workshop of Harmoni-CA/WP5. Harmoni-CA report HCA-WP5-2004-01/Final
Version. University of Osnabrueck, Osnabrueck (available at http://www.harmoni-ca.info )
Hare M (2005) The use of models to support the participatory elements of the EU Water Framework Directive:
creating a dialogue between policy makers and model makers. Harmoni-CA report HCA-WP5-2004-04/
Final Version. University of Osnabrueck, Osnabrueck (available at http://www.harmoni-ca.info)
Hare MP, Deadman P (2004) Further towards a taxonomy of agent-based simulation models in
environmental management. Math Comput Simul 64:25–40
Hare MP, Letcher RA, Jakeman AJ (2003) Participatory modelling in natural resource management: a
comparison of four case studies. Integrated Assessment 2(4):62–72
Hare MP, Barreteau O, Beck MB, Cogan V, Mostert E, Letcher R, Ridder D, Tabara JD (2006) Methods for
stakeholder participation in water management. In: Giupponi C, Jakeman AJ, Karssenberg D, Hare MP
(eds) Sustainable management of water resources: an integrated approach. Edward Elgar, Chichester
Water Resour Manage
HarmoniCOP (2005) Learning together to manage together –improving participation in water management.
Handbook of the HarmoniCOP project. http://www.harmonicop.info
Horlitz T (2006) The role of model interfaces for participation in water management. Journal Water
Resources Management Special Issue: Participatory Methods for Water Resources Management
Judson OP (1994) The rise of the individual-based model in ecology. TREE 9(1):9–14
LAWA (2003) Arbeitshilfe zur Umsetzung der EG-Wasserrahmenrichtlinie. Teil 3 & 4. Bearbeitungsstand
30. 04. 2003. Länderarbeitsgemeinschaft Wasser. http://wasserblick.net/
Larsen H, Mark O et al (2000) The application of models in integrated river basin management. Internal
Publication des Asian Institute of Technology and DHI Water & Environment. DHI.
Letcher RA, Jakeman AJ, Croke BFW (2004) Model development for integrated assessment of water
allocation options. Water Resour Res 40:W05502
Lynam T, Bousquet F, Page CL, d’Aquino P, Barreteau O, Chinembin F, Mombeshora B (2002) Adapting
science to adaptive managers: spidergrams, belief models and multi-agent systems modeling.
Conservation Ecology 5:[online] URL: http:www.consecol.org/vol5/iss2/art24
Maiden NAM, Hare MP (1998) Problem domain categories in requirements engineering. Int J Human-
Comput Stud 49:281–304
McGraw KL, Harbison K (1997) User-centred requirements –the scenario-based engineering process.
Lawrence Erlbaum, Mahwah, NJ
McIntosh BS, Seaton RAF, Jeffrey P (2004) Tools to think with? Towards understanding the use and impact
of model-based support tools. Environ Model Softw
Mostert E (2003) The challenge of public participation. Water Policy 5:179–197
Old G, Packman J (eds) (2005) Position paper on professional response. HarmoniQuA report. HarmoniQuA,
Wageningen
Old G, Robinson M, Packman J (2005) Position paper on stakeholder response. WP4-4 Harmonising Quality
Assurance in model based catchment and river basin management. HarmoniQuA, Wageningen
Oreskes N, Shrader-Frechette K, Belitz K (1994) Verification, validation and confirmation of numerical
models in the earth sciences. Science 263:641–646
Pahl-Wostl C (2002) Participative and stakeholder-based policy design and modeling processes. Integrated
Assessment 3(1):3–14
Pahl-Wostl C, Hare MP (2004) Processes of social learning in integrated resources management. J
Community Appl Soc Psychol 14:193–206
Quevauviller P (2006) Introduction to WISE–Water Information System for Europe. Presentation at the 3rd
Harmoni-CA Forum & Conference, 5–7th April 2006. http://www.harmoni-ca.info
Refsgaard JC, van der Sluijs JP, Højberg AL, Vanrolleghem P (2005) Harmoni-CA guidance uncertainty
analysis. Guidance no 1. Harmoni-CA, Brussels
Rekolainen S, Kämäri J, Hiltunen M (2004) A conceptual framework for identifying the need and role of
models in the implementation of the Water Framework Directive. International Journal of River Basin
Management 1(4):1–6
Robinson JB, Freebairn DM (2001) The decision-maker has an important but often neglected, role in model
use. MODSIM 2001 4:1595–1600
Scholten H, Kassahun A, Refsgaard JC, Kargas T, Gavardinas C, Beulens AJM (in press) A methodology to
support multidisciplinary model-based water management. Environ Model and Softw
Swinford A, McIntosh B, Jeffrey P (2004) Supporting the strategic objectives of participative water resources
management: an evaluation of the performance of four ICT tools. In: Pahl-Wostl C, Schmidt S, Rizzoli
AE, Jakeman AJ (eds) Complexity and integrated resources management. Transactions of the 2nd
Biennial Meeting of the International Environmental Modelling and Software Society. iEMSs: Manno,
Switzerland. ISBN 88-900787-1-5
Ubbels A, Verhallen AJM (2000) Suitability of decision support tools for collaborative planning processes in
water resources management. RIZA rapport 99.067. RIZA, Wageningen
Vennix JAM (1996) Group model building. Wiley, Chichester
Vennix JAM (1999) Group model-building: tackling messy problems. System Dynamics Review 15(4):379–401
Vonk G, Geertman S, Schot P (2005) Bottlenecks blocking widespread usage of planning support systems.
Environ Plann A 17:909–924
Wasson J-G, Tusseau-Vuillemin MH, Andreassian V, Perrin C, Faure JB, Barreteau O, Bousquet M, Chastan B
(2003) What kind of water models are needed for the implementation of the European water framework
directive? Examples from France. International Journal of River Basin Management 2(1):125–135
Welp M (2001) The use of decision support tools in participatory river basin management. Phys Chem Earth
26(7–8):535–539
Winograd T (1990) Thinking machines: can there be? Are we? In: Partridge D, Wilks Y (eds) The
foundations of artificial intelligence: a sourcebook. Cambridge, Cambridge
Water Resour Manage
