Defining projects and scenarios for integrated
assessment modelling using ontology
Janssen, S.J.C.1,2, J.J.F. Wien3, Hongtao Li4, I. N. Athanasiadis4, F. Ewert1, M.J.R. Knapen 3, D.
Huber5 , O. Thérond6, A.E. Rizzoli4, H. Belhouchette7, M. Svensson8 and M.K. van Ittersum1
1 Wageningen University, Plant Production Systems Group, Wageningen
2 Wageningen University, Business Economics Group, Wageningen
3 Alterra, Environmental Sciences Group, Wageningen University and Research Centre, Wageningen
4Dalle Molle Institute for Artificial Intelligence (IDSIA), Lugano
5
AntOptima SA, Lugano
6INRA- UMR-AGIR, Toulouse
7
INRA-Agro M. UMR SYSTEM, Montpellier
8 Lund University, Centre for Sustainability Studies, Lund
Email: sander.janssen@wur.nl
Keywords: Scenarios, integrated assessment, modelling, ontology
EXTENDED ABSTRACT
Integrated Assessment Modelling provides a
systematic inter-disciplinary approach to support
coherent ex-ante decision-making by a flexible
integration of (reusable) models and datasets
across scales. Within integrated assessment
modelling a coherent and robust description of
projects and scenarios is required to facilitate data
preparation, model integration and the graphical
user-interface development.
This paper explains our experiences with a
challenging and time-consuming task, e.g.
arriving at a shared understanding on the
definition of projects, experiments and scenarios
among researchers coming from different
disciplines, who have been exposed to dissimilar
education and research experience. We
demonstrate the use of ontologies in building this
shared set of definitions and the relationship
between the ontology and the human computer
interaction through a case study. With a common
ontology that represents the joint
conceptualization of the projects, experiments and
scenarios each researcher can refer at any later
stage to the semantics of the concepts used. A
collaborative approach was used to build such a
common ontology in the SEAMLESS-Integrated
Project, funded through the EU sixth Framework
Programme, which aims at developing an
integrated modelling framework (SEAMLESS-IF)
to assess, ex-ante, agricultural and environmental
policy options, allowing cross-scale analysis of a
broad range of sustainability issues.
Through several iterations a common ontology for
projects, experiments and scenarios was built. In
our common ontology a project has one and only
one problem definition, and it can handle at least
one or more Experiments. Experiments represent
the assessment of one or a combination of policy
options in a given context and outlook on the
future. The indicator(s) should be the same between
experiments which are part of the same project,
allowing the comparison of different experiments.
Each of the concepts Policy Option, Context and
Outlook capture one part of the input parameters
required for running each of the models.
As a first validation of the project ontology, a set of
four fictitious sample projects were made. One of
these sample projects is an integrated assessment
for one region Midi-Pyrénées in the South of
France concerning the impacts of the CAP2003
reform, which is described in this paper
The common project ontology highlighted the
imprecise meaning of the word scenario and it links
projects to problems, outlooks on the future,
indicators, context of the problem, policies and
ultimately to model runs in experiments. Also, by
this common ontology the assumptions in building
the assessment are clarified, moving the focus away
from the tools to the assumptions underlying
models and scenarios. In any integrated assessment
project, it is recommended to clarify with its
participants the meaning of scenario and associated
concepts. We achieved this by the use of a common
ontology, which forces participants to be clear,
precise and coherent in their description of concepts
and relationships between concepts, while the
common ontology can be directly used for
development of databases, models and graphical
user interfaces.
2055

1. INTRODUCTION
Integrated Assessment Modelling (IAM) is more
and more frequently used methodology to assess
the impacts of policies, technologies or societal
trends on the environmental, economic and social
future sustainability (Parker et al., 2002), for
example in mitigation to climate change (Weyant
et al., 1996; Cohen, 1997) or water quality in
catchment areas (Turner et al., 2001). Integrated
assessments are defined by Rotmans and Van
Asselt (1996) as an interdisciplinary, participatory
and future-oriented process of combining,
interpreting and communicating knowledge from
diverse scientific disciplines to allow a better
understanding of complex phenomena.
Scenario analysis is identified as an important tool
in integrated assessment (Rotmans, 1998), where
scenarios are used in the interaction between
scientists and stakeholders to describe what the
future could be. Many different definitions of
scenario exist in scenario literature. For example,
Rotmans (1998) defines scenarios as ‘archetypal
descriptions of alternative images of the future,
created from mental maps or models that reflect
different perspectives on past, present and future
developments’, while Parry and Carter (1998)
define scenarios as ‘a coherent, internally
consistent and plausible description of a possible
future state of the world.’ Peterson et al. (2003)
provides a definition of scenario which is closer to
modelling, ‘as variation in the assumptions used to
create models.’
Given that a wide range of definitions is available
for scenarios, there is a risk for confusion and
misunderstanding in any integrated assessment
modelling project. For example, does scenario
refer to the outcomes of model runs? Or does it
refer to the set of input parameters to a model? Is it
only related to policies as suggested by the term
policy scenario or is it broader? This reinforces the
need for a clear set of rules and protocols for
integrated assessment, in particular with respect to
the understanding of scenarios, as concluded by
Rotmans and Van Asselt (1996), to avoid the
dangers of unclear, inconsistent, narrowly-defined
scenarios and ad-hoc setting of parameters
(Rotmans, 1998; van Asselt and Rotmans, 2002).
SEAMLESS is an integrated assessment modelling
project (Van Ittersum et al., 2007), which aims to
provide an computerized framework to assess the
sustainability of agricultural systems in the
European Union at multiple scales. As in
SEAMLESS over one hundred of scientist
participate from different disciplines and dissimilar
research background, many different views exist
on the meaning of scenario and its implications for
the computerized SEAMLESS-IF framework.
As SEAMLESS builds a computerized framework,
one and only unified view on the meaning of
scenarios and the definition of assessment projects
is required within the group of scientist to enable
data preparation, model integration and graphical
user-interface development. There is no
established procedure to develop such a unified
view on the meaning of scenarios and the
definition of assessment projects. This paper
explains our experiences with this challenging and
time-consuming task, e.g. arriving at a shared
understanding on the definition of projects and
scenarios among researchers from different
disciplines with dissimilar research experience. We
demonstrate the use of common ontologies in
building this shared conceptual model through a
case study.
In the next section, some background will be
provided on common ontologies and the process of
ontology engineering. Also, our case study set-up
will be introduced for the SEAMLESS-project. In
Section 3, the shared conceptual model on scenario
and project definition is presented, supported by a
fictitious example of the use of the common
concept in a regional integrated assessment
project, while this common concept will be
discussed in Section 4. In the final section we
highlight the most important lessons we learned in
our case study.
2. MATERIALS AND METHODS
2.1. Common Ontologies
In the context of integrated modelling, ontologies
are useful to define the shared conceptualization of
a problem, as ontologies are written in a language,
e.g. Web Ontology Language, that is
understandable by computers and as ontologies
consists of a finite list of concepts and the
relationships between these concepts (McGuinness
and van Harmelen, 2004). The term ontology
originates from philosophy, originally coined by
classical philosophers Plato and Aristotle
(Aristotle, 336-332 BC) in the study of types of
being and their relationships (metaphysics). An
ontology in computer science is considered as a
specification of a conceptualization (Gruber,
1993), where a conceptualization is ‘an abstract,
simplified view of the world “e.g. systems under
study (addition by author)” that we wish to
represent for some purpose’ (Gruber, 1993). In
integrated modelling research, scientists from
various disciplines can define a common
2056

conceptual schema that their domains share. A
common project ontology, i.e. ontology which is
shared by all domains to-be-integrated , serves as a
knowledge-level specification of the joint
conceptualization of the project and scenario
definition. Each scientist can refer to and should
adhere to the semantics of the concepts in the
common project ontology, including restrictions
on the concepts and relationships between the
concepts.
2.2. Ontology Engineering
In developing a common ontology, the scientific
challenge of adopting tight, well-reasoned and
shared conceptualizations among a group of
scientists or one individual scientist should be
overcome. The development of a common
ontology by a group of researchers is a complex,
challenging and time-consuming task (Musen,
1992; Gruber, 1993; Farquhar et al., 1995;
Holsapple and Joshi, 2002). Tools are available
that help in ontology development (Farquhar et al.,
1995) and to store the ontology once it was
developed (Knublauch, 2005). To achieve
ontological commitment, i.e. the agreement by
multiple parties to adhere to a common ontology,
when these parties do not have the same
experiences and theories (Holsapple and Joshi,
2002), a collaborative approach is suggested to be
used. A collaborative approach has the advantages
that researchers from different disciplines are
diverse in their contributions, which avoids
blindspots and which has more chances of getting
a wide acceptance (Holsapple and Joshi, 2002) and
that it can incorporate the other approaches, e.g.
synthetic approach, as required for development of
parts of the ontology.
2.3. Case Study: SEAMLESS project
The SEAMLESS integrated project (System for
Environmental and Agricultural Modelling;
Linking European Science and Society), EU sixth
Framework project, develops a computerized and
integrated framework (SEAMLESS-IF) to assess
the impacts on environmental and economic
sustainability of a wide range of policies and
technological improvements across a number
scales. This aim should be achieved by
overcoming the gap between micro-marco level
analysis, overcoming the bias in integrated
assessments towards either economic or
environmental issues, facilitating the re-use of
models and providing methods to technically link
different models together (Van Ittersum et al.,
2007).
Within SEAMLESS, both modelling and
stakeholder involvement are seen as important
elements of the assessment procedure proposed by
SEAMLESS-IF. With respect to modelling,
macro-level economic partial or general
equilibrium models are linked to micro-level farm
optimization models and field crop growth models,
while in between macro and micro-level steps of
aggregation and des-aggregation occur by other
models. A participatory approach is foreseen for
the use of SEAMLESS-IF with stakeholders at the
end of the project. The SEAMLESS-IF should be
designed to facilitate such a participatory
approach. Prime Users as the Directorates General
of the European Commission are involved in this
process through a User Forum. In the project 30
partners and more than 100 researchers participate.
Thus, the common ontology for project and
scenario definition acts on these interfaces between
modellers and other scientists and between
scientist and stakeholders after the development of
the SEAMLESS-IF (Fig. 1). The common
ontology is used to construct the database schema
to store data on projects and scenarios and Javatm-
beans (Athanasiadis et al., 2007) for development
of the Graphical User Interface (GUI) and for
structuring input for the models.
Figure 1 Role of a common project ontology in a
integrated assessment modelling project
The collaborative approach for ontology
engineering in our case study was based on
developing one shared document between a group
of seventeen researchers working in different parts
of the SEAMLESS project. Ten iterations of the
document were used and after each iteration a
small ontology constructed in Protégé OWL
(Knublauch, 2005) was synchronized with the
iteration. With each iteration, more scientist were
involved starting from four for this first iteration
up to seventeen for the tenth iteration. At the tenth
iteration both the document and the ontology were
Modeler
Scientist Stakeholder
Common project
ontology
Database
Graphical
user
interface
Model(s)discuss
discuss
discu
ss
d
ete
rm
in
e
s
determines
determines
2057

‘closed’ after the approval of the SEAMLESS
management group and a set of actions was
formulated to elaborate specific parts of the project
and scenario definition.
3. RESULTS
3.1. The project ontology
A project ontology was developed for integrated
assessments as carried out in the SEAMLESS-
projects, and here, a description in words is
provided, together with an overview in diagrams.
In SEAMLESS, we foresee in an application of the
SEAMLESS-IF an integrative modeller (scientist)
working together with a policy expert or other
stakeholders. In general, a project in SEAMLESS
refers to the assessment of the effects on
agricultural sustainability and profitability of
changes in policies. When the integrative modeller
wants to work with the SEAMLESS-IF, he/she
always has to build a project that will handle the
specific problem the integrative modeller wants to
tackle, after discussing it with a policy expert. A
project has one and only one problem definition,
and it can handle at least one or more Experiments
(Fig. 2). This implies that different perspectives on
a problem can be investigated through different
Experiments, representing the assessment of one or
a combination of policy options in a given context
and outlook on the future.

 




 



















 


 
 
Figure 2. A data model describing the project
ontology.
Thus, an experiment is ONE run of the models
within SEAMLESS-IF that assesses ONE or A
COMBINATION of policy option(s) within a
context and an outlook on the future. The
indicator(s) should therefore be the same between
experiments which are part of the same project,
allowing the comparison of different experiments.
Each quantitative indicator selected for a project
gets one value for each Experiment. Impacts are
the changes in indicator-value due to changes in
policy options, context and outlook on the future
as compared to a reference situation (Fig. 3).
Expected impacts are then changes in indicator
values defined by the Policy Expert before running
the experiment in the SEAMLESS-IF, while
calculated impacts are the changes in indicator
values as calculated by the SEAMLESS-IF.
Each of the concepts Policy Option, Context and
Outlook capture one part of the input parameters
required for running each of the models. These
parameters are unchangeable by the models,
meaning that the model run does not affect the
value of the parameter, and so these are
exogenous. Next to these exogenous parameters,
the models have endogenous parameters, which
are parameters that can be changed by other
models in the model chain. A policy option refers
to one or more policy measures as part of it and
each policy option is described by a set of
parameters that are exogenous to the models. An
example of a policy option is a reform of the
Common Agricultural Policy of the European
Union, which has as measures a decoupling of the
subsidy payments from the area of the farm to
income support and the lowering of the support
prices for products. Relevant policy parameters are
for this example the percentage of decoupling for a
region, the reference yield for a region and the cut
in premiums.
Figure 3. An ontology snapshot showing the
concepts ‘Impact’, ‘Indicator’ and ‘Experiment’
(large circles), their relationships (arrows) and
their data-properties (small circles).
Outlook on the future describes trends and trend
deviations foreseen to occur in society that might
affect the implementation of policy options within
a given context. These trends or trend deviations
are not modelled endogenously in SEAMLESS.
One reference outlook is always required that
describes the prolongation of the current situation
into the future, sometimes called business-as-usual
outlook. Outlooks are usually contrasting, for
example a positive versus a negative outlook, a
globalization versus a regionalization outlook.
Each outlook has several exogenous parameters
that capture the different trends occurring in
society. Examples of these parameters are
Indicator
Dimension
Impact
Experiment
Name
CalculatedValue
OfExp
erime
nt
HasIndicator
ExpectedValue
Description
Title
2058

atmospheric CO2-concentration, GDP-growth and
unemployment rate.
Finally, the context of a problem defines the object
of interest, which is delimited by the boundaries to
the biophysical and agro-management system.
These boundaries determine what is inside and
what is outside the system. Each experiment
within a problem will be based on one agro-
management and biophysical context that can be
different from those of other experiments.
The experiments thus define the changes or driving
forces as compared to the reference situation, by
capturing the changes in policy options, context
and outlook, either as changes in isolation (only
one policy option/outlook/context-change) or
simultaneously (more than one policy option/
outlook/ context-change). This implies that the
maximum number of experiments to be defined is
the factorial combination of policy options,
outlooks and context, but not all of these
experiments make sense.
3.2. An example project
As a first validation of the project ontology, a set
of four fictitious sample projects were made. One
of these sample projects is an integrated
assessment for one region Midi-Pyrénées in the
South of France concerning the impacts of the
CAP2003 reform as requested by Mrs X and Mr.
Y of a regional government agency. They also
want to evaluate if CAP2003 reform will favour
conservation agriculture in the Midi-Pyrénées
region and they are curious to know if a subsidy on
conservation agriculture would increase the uptake
of conservation agriculture. They discuss their
assessment with three scientists, working at a
research institute in Toulouse, Midi-Pyrénées.
The three scientists define the following two
policy options, based on their discussions with the
policy experts. The policy option CAP2003 reform
comprises a set of European Union policies,
related to income support for farmers, while the
policy option subsidies for conservation
agriculture also comprise a set of policies, but
targeted at the sustainability of agriculture in terms
of soil conservation.
After discussing with the policy experts, the
scientists think they need three outlooks. One
business-as-usual-outlook in which there are no
trend deviations in society, so the current situation
is prolonged. Next to this a globalization and
regionalization outlook are defined. In the
globalization outlook, atmospheric CO2-
concentration is expected to rise steeply by 5%,
and unemployment in Midi-Pyrénées is expected
to decrease by 3%. In the regionalization Outlook
atmospheric CO2-concentration is expected to rise
mediocre by 2%, and unemployment in Midi-
Pyrénées is expected to increase by 3%. (Table 1)
Table 1. Outlooks for the fictitious example
concerning Midi-Pyrénées
Outlook Atmospheric
CO2
concentration
Unemployment
Business-as-
usual
No change No change
Globalization +5% -3%
Regionalization +2% +3%
Two contexts should be defined. These contexts
describe the technological innovation as driving
forces. One context is the situation without the
possibility for the farmers to choose conservation
agriculture, and in the other context farmers have
the option to choose conservation agriculture on
the farm.
The scientists define four experiments, which are a
combination of one context, one outlook and one
or a combination of policy options (Table 2). The
first experiment is the business-as-usual
experiment where no CAP 2003 reform takes
place, the farmers do not have options for
conservation agriculture and the outlook for the
future is that no particular important trends occur.
This experiment acts as a reference point for
comparison of the other experiments (Table 2).
Table 2. The experiments for the fictitious sample
project in Midi-Pyrénées
Experiments Policy
option
Outlook Context
1. Business
as Usual
Only current
policies
Business
as Usual
No
conservation
agriculture
2. CAP 2003
reform
CAP 2003
Reform
Globa-
lization
No
conservation
agriculture
3. No
support
CAP 2003
Reform
Regiona-
lization
Conservation
agriculture
4.
Conservation
oriented in
regional
world
CAP 2003
Reform and
subsidies for
conservation
agriculture
Regiona-
lization
Conservation
agriculture
Finally, the scientist discuss with the policy
experts the relevant indicators, which are the
regional cropping pattern, the farmer income, the
amounts of subsidies, the % of no-plowing tillage,
2059

the area for the intercrops mustard and clover and
the level of erosion.
4. DISCUSSION
4.1. Scenario and its meaning
The fictitious example for the integrated
assessment in the Midi-Pyrénées region
demonstrated that all relevant aspects of the
integrated assessment could be captured by the
project ontology. In our project ontology as
presented in the previous Section, we have no
concept Scenario as part of it. In the iterative
process of building the common project ontology,
it was apparent that scenario had different
meanings for scientists. Some scientists thought of
scenarios as experiments, so a perspective of what
could change in terms of policies, outlooks and
context, and thereby determining the input
parameters for the models. Other scientists thought
of scenarios as a set of impacts, so values for
indicators that were the ‘end-result’ of changes in
policies, outlooks and contexts. Economic
modellers limited their definition of scenario to
policy options, while biophysical modellers were
more inclined to think of scenario as outlook. The
concept scenario was thus not included in the
project-ontology for risk of confusion and other
concepts were chosen that could be un-
ambiguously defined and agreed upon. This proves
that the project ontology is able to cover all the
different meanings scenario can have, and offers
an opportunity to comprehensively describe an
integrated assessment problem. One definition of
scenario could be centrally decided on by scientists
participating in an integrated assessment.
4.2. Project ontology and models
Models are not explicitly mentioned as a separate
concept in the project ontology and the fictitious
sample projects as presented in Section 3, although
a link exists between the properties of the context,
outlook and policy option and input parameters for
the models. The required models to analyse an
assessment problem can be deduced from the
selections by the integrative modeller with respect
to the properties of context, policy option and
outlook. This allows to focus on the assumptions
made while defining values for the different model
input parameters and defining the experiments
instead of focusing solely on the technical
capabilities of the models (Rotmans, 1998;
Greeuw et al., 2000). Many different types of
models could be linked to the project ontology, for
example optimization models and deterministic
simulation models.
4.3. Use of ontologies and ontology
engineering
By using ontology engineering as our
methodology, scientists participating in this
collaborative process had to be precise in their
meaning of concepts they proposed for the
common ontology, and they had to ensure
consistency and coherence between the concept
they proposed and the other concepts in the project
ontology. With ten iterations and seventeen
participating scientist, the collaborative approach
required a clear objective and set of actions for
each iteration, which lead it to be a time-
consuming task. The collaborative approach was
an appropriate solution in our case, as knowledge
from scientist from different domains could be
disclosed and as the project ontology has become a
common reference point for scientists in the
project, reflecting the shared understanding.
4.4. Future developments
More evaluation is required by peer review and
stakeholders outside the science community of the
common project ontology. Also, the concept of
‘scale’ needs to be included, which is recognised
as an important concept in integrated assessments
(Parker et al., 2002).
5. CONCLUSIONS AND
RECOMMENDATIONS
As concluded by Cohen (1997), ‘if stakeholders
and their knowledge are to be drawn into
integrated assessments, integrated assessments
must become less of a “black box” and more
human.’ Through the development of a common
ontology on projects and scenarios, we opened the
black box surrounding scenarios in terms of
meaning and content. Our common ontology
improves the consistency and transparency of
scenarios, as (i) a set of concepts is provided to
describe different types of model input parameters,
as (ii) the focus is more on assumptions made in
defining these input parameters instead of the tools
themselves and as (iii) experiments should be
explicitly constructed capturing the different
perspectives on the future.
In any integrated assessment project, it is
recommended to clarify with its participants the
meaning of scenario and associated concepts. We
achieved this by the use of a common ontology,
which forces participants to be clear, precise and
coherent in their description of concepts and
relationships between concepts, while the common
ontology can be directly used for development of
databases, models and graphical user interfaces.
2060

6. ACKNOWLEDGMENTS
We thank all scientists in the SEAMLESS project
who contributed to development of the common
ontology on projects and scenarios. This work has
been carried out as part of the SEAMLESS
Integrated Project, EU sixth Framework
Programme, Contract No. 010036-2.
7. REFERENCES
Aristotle (336-332 BC), Metaphysics,
Athanasiadis, I. N., F. Villa and A. E. Rizzoli
(2007), Enabling knowledge-based software
engineering through semantic-object-relational
mappings, Paper presented at 3rd International
Workshop on Semantic Web Enabled Software
Engineering, 4th European Semantic Web
Conference, 15p., Innsbruck, Austria.
Cohen, S. J. (1997), Scientist–stakeholder
collaboration in integrated assessment of climate
change: lessons from a case study of Northwest
Canada, Environmental Modelling and Assessment
2(4), 281.
Farquhar, A., R. Fikes, W. Pratt and J. Rice
(1995), Collaborative Ontology Construction for
Information Integration, Secondary Collaborative
Ontology Construction for Information Integration,
32 p. Knowledge Systems Laboratory, Department
of Computer Science, Stanford University,
Greeuw, S. C. H., M. B. A. Van Asselt, J.
Grosskurth, C. A. M. H. Storms, N. Rijkens-
Klomp, D. S. Rothman and J. Rotmans (2000),
Cloudy crystal balls- An assessment of recent
European and global scenario studies and models,
Secondary Cloudy crystal balls- An assessment of
recent European and global scenario studies and
models, 112 p. European Environment Agency,
Copenhagen
Gruber, T. R. (1993), A Translation Approach to
Portable Ontology Specifications, Knowledge
Acquisition 5, 199-220.
Holsapple, C. W. and K. D. Joshi (2002), A
collaborative approach to ontology design,
Communicationss of the ACM 45(2), 42-47.
Knublauch, H. (2005), Protege OWL, p., Stanford:
Stanford Medical Informatics
Mcguinness, D. and F. Van Harmelen (2004),
OWL Web Ontology Language Overview,
Secondary OWL Web Ontology Language
Overview, p. WWW Consortium,
Musen, M. A. (1992), Dimensions of knowledge
sharing and reuse, Computers and Biomedical
Research 25(5), 435-467.
Parker, P., R. Letcher, A. Jakeman, M. B. Beck,
G. Harris, R. M. Argent, M. Hare, C. Pahl-
Wostl, A. Voinov and M. Janssen (2002),
Progress in integrated assessment and modelling,
Environmental Modelling & Software 17(3), 209-
217.
Parry, M. and T. Carter (1998), Climate Impact
and Adaptation Assessment, Earthscan
Publications Ltd., London, UK
Peterson, G. D., G. S. Cumming and S. R.
Carpenter (2003), Scenario Planning: a Tool for
Conservation in an Uncertain World, Conservation
Biology 17(2), 358-366.
Rotmans, J. (1998), Methods for IA: The
challenges and opportunities ahead, Environmental
Modelling and Assessment 3(3), 155.
Rotmans, J. and M. Asselt (1996), Integrated
assessment: A growing child on its way to
maturity, Climatic Change 34(3), 327.
Turner, R. K., L. Ledoux and R. Cave (2001), The
Use of Scenarios in Integrated Environmental
Assessment of Coastal-Catchment Zones,
Secondary The Use of Scenarios in Integrated
Environmental Assessment of Coastal-Catchment
Zones, 25 p. School of Environmental Sciences,
University of East Anglia, Norwich
Van Asselt, M. B. A. and J. Rotmans (2002),
Uncertainty in Integrated Assessment Modelling,
Climatic Change 54(1), 75.
Van Ittersum, M. K., F. Ewert, T. Heckelei, J.
Wery, J. Alkan Olsson, E. Andersen, I.
Bezlepkina, F. Brouwer, M. Donatelli, G.
Flichman, L. Olsson, A. Rizzoli, T. Van Der
Wal, J.-E. Wien and J. Wolf (2007), Integrated
assessment of agricultural systems- a component
based framework for the European Union
(SEAMLESS), Agricultural Systems In Press.
Weyant, J., H. Davidson, H. Dowlatabadi, J.
Edmonds, M. Grubb, R. Richels, J. Rotmans, P.
Shukla, R. S. J. Tol, W. Cline and S. Fankhauser
(1996), Integrated Assessment of climate change:
An overview and comparison of approaches and
results, In Climate Change 1995 - Economic and
Social Dimensions Eds J. P. Bruce, H. Lee & E. F.
Haites), Cambridge University Press, Cambridge

2061