History

Bayesian networks in water resource modelling and management

Description
Bayesian networks in water resource modelling and management
Categories
Published
of 2
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
Share
Transcript
  Editorial Bayesian networks in water resource modelling and management The srcin of this Special Issue lies in a research project of the European Union: the project FP5-MERIT (Management of the Environment and Resources using Integrated Techniques).Its target was to enquire if Bayesian Networks (Bns), a model-ling approach mainly adopted for diagnosis in the fields of medicine and system maintenance, may be useful in the fieldof environmental participatory and integrated planning andmanagement.In the search for tools that could help in coupling participa-tion to integration, the Bns seemed, in fact, to be quite appeal-ing. Indeed, they are graphical models that exploit the dualitybetween an interaction graph and a probability model. Thegraphical structure provides a visual representation of thecause e effect relationships in the problem at hand, thus facili-tating participation, while thepotential ofthe probabilitymodelsupports integration. Moreover, Bns can also be interpreted asa decision-support tool, when decisions and objectives (utili-ties) are explicitly considered among their variables. To inves-tigate the Bns’ potential, four case studies in the field of waterresources were designed and carried out in four different re-gions of Europe which presented different planning and man-agement tasks. The results of three of them are presented inthree papers that constitute the central part of this issue. Theissue is completed by three other contributions, which, respec-tively,providethe theoretical background and two further inter-esting examples of the application of Bns in the water resourcefield.As already mentioned, Bns can be used as models as wellas decision-support tools. However, it is extremely useful toseparate these two functions, enquiring about their reciprocalrole and nature. This is argued in the first paper (Castellettiand Soncini-Sessa, 2007a), which imbeds Bns in the widerspace of making decisions in a participatory and integratedcontext, and compares them with different types of models.The paper also provides a short introduction to the notion of Bns that is assumed to be known in the other papers.The subsequent three papers concern the MERIT casestudies. Martin de Santa Olalla et al. (2007) consider the planning of sustainable withdrawal from a large aquifer inEastern Mancha, Spain; the planning actions they deal withare the reduction of irrigation withdrawal, the improvementof the efficiency of the irrigation system and the increasein water price. Also Hans Jørgen Henriksen et al. (2007)address the planning of an aquifer, in Denmark. However,this time the focus is on groundwater quality and on the ef-fects that voluntary farming contracts may have on it. Fi-nally, Andrea Castelletti and Rodolfo Soncini-Sessa (2007b)describe planning issues regarding a complex water reservoirnetwork in a river basin in Central Italy. The actions hereconsidered are the modification of the irrigated crop surface,the introduction of financial incentives to promote the use of more water-saving irrigation techniques, the definition of minimum inflows on some river stretches and the revisionof the management policies of the reservoirs.Two contributions close the issue. In the first, Jenifer Tice-hurst et al. (2007) show the use of Bns in assessing the sustain-ability of coastal lake-catchment systems in New South Wales,Australia. The actions considered are different farm manage-ment practices and new urban developments. In the secondpaper, Carmel Pollino et al. (2007) use Bns to further develop parameterisation methods, using an ecological risk manage-ment case study to assess the impacts of changed environmen-tal conditions on native fish communities in another Australiancatchment (in Victoria). The actions here are all those factorsthat are presumed to have an impact on fish abundance anddiversity, such as flow alteration, introduction of non-nativespecies, etc.The differences in the contributions of this issue are notonly due to the characteristics of the water system and theproblems posed, but also to the procedures adopted for the net-work population (calibration), the type of stakeholder partici-pation, the way the Bns are used to solve the problem and theextent of the system to which they are applied.  Bn population . In the Spanish study, the Bn is populated bymixing statistical manipulations of historical data and ex-pert judgements; in the Danish study, by interviews withthe farmers (qualitative data) and expert judgements (quan-titative data); in the Italian study only through interviews;while in the New South Wales case, it is populated on thebasis of manipulation of published data, expert opinionand model simulation. Network parameterisation is a centraltopic in the Victoria case study, where knowledge elicita-tion from domain experts is combined with machine learn-ing techniques. Here, the results of a sensitivity analysis onthe Bn’s parameters are also presented. Environmental Modelling & Software 22 (2007) 1073 e 1074www.elsevier.com/locate/envsoft1364-8152/$ - see front matter    2006 Elsevier Ltd. All rights reserved.doi:10.1016/j.envsoft.2006.06.001  Stakeholder participation . In the first two cases the stake-holders participation is informative and consultative, whilein the third stakeholders are actively assigned a co-designrole in the model construction process. In the New SouthWales case study, close participation with catchment man-agers is a feature of the process of Bn development. Finally,in the Victoria case study only domain experts are consultedto obtain the information required to populate the network.  Bn’s use . In the first three cases and in the New SouthWales case study the problem is formulated as a feasibilityproblem, i.e. as the search for an alternative that satisfiesa given set of constraints, and the solution is determinedthrough iterative ‘‘what if’’ questions, i.e. a trial-and-errorprocedure. Conceptually analogous is the approach adoptedin the Victoria case study, where the Bn is used to predictfish abundance and diversity in response to the human ac-tivities on the basis of the existing or predicted environmen-tal scenarios. Instead, in the Italian case study, the problemis formulated within a wider decision making framework and the alternatives to be evaluated are designed by solvingan optimal control problem. System described  . In all cases except for the Italian one Bnsare used to represent the whole water system: in the Italianone the network is used to model only one component (theirrigation district), while the others are described with othertypes of models and the model of the whole system is there-fore obtained by integrating different types of models.We conclude by thanking the authors for their efforts in re-writing and extending the srcinal version of their papers, andfor taking pains to achieve clarity for a wide readership. Wehope that you will find the papers stimulating and interesting. References Castelletti, A., Soncini-Sessa, R., 2007a. Bayesian networks and participatorymodelling in water resource management. Environmental Modelling andSoftware 22 (8), 1075 e 1088.Castelletti, A, Soncini-Sessa, R., 2007b. Coupling real time control and socio-economic issues in participatory river basin planning. EnvironmentalModelling and Software 22 (8), 1114 e 1128.Henriksen, H.J., Rasmussen, P., Brandt, G., von Bulow, D., Jensen, F., 2007.Public participation modelling using Bayesian networks in managementof groundwater contamination. Environmental Modelling and Software22 (8), 1101 e 1113.Martin de Santa Olalla, F., Dominguez, A., Ortega, F., Artigao, A., Fabeiro, C.,2007. Bayesian networks in planning a large aquifer in Eastern Mancha,Spain. Environmental Modelling and Software 22 (8), 1089 e 1100.Pollino, C.A., Woodberry, O., Nicholson, A., Korb, K., Hart, B.T., 2007.Parameterisation and evaluation of a Bayesian network for use in an eco-logical risk assessment. Environmental Modelling and Software 22 (8),1140 e 1152.Ticehurst, J.L., Newham, L.T.H., Rissik, D., Letcher, R.A., Jakeman, A.J.,2007. A Bayesian network approach for assessing the sustainability of coastal lakes in New South Wales, Australia. Environmental Modellingand Software 22 (8), 1129 e 1139. Andrea Castelletti*Rodolfo Soncini-Sessa  Dipartimento di Elettronica e Informazione Politecnico di Milano, Milano, Italy *Corresponding author. Tel.:  þ 39 0223999632;fax:  þ 39 0223993412.  E-mail address:  castelle@elet.polimi.it (A. Castelletti)26 May 2006Available online 20 July 2006 1074  Editorial / Environmental Modelling & Software 22 (2007) 1073 e 1074

Joana

May 14, 2018

Nuno rufino

May 14, 2018
Search
Similar documents
View more...
Tags
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks
SAVE OUR EARTH

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!

x