Bayesian networks in planning a large aquifer in Eastern Mancha, Spain
Introduction
Water demand is increasing throughout the world, while at the same time there is a general decrease in overall water quality. Faced with this situation, the European Union regards it as a top priority: efficiency in the management and use of water resources should be improved. This objective is included into the current EU Water Framework Directive 2000/60/EC.
The use of groundwater to supply large surfaces of irrigated land has been the key to agricultural development in a large number of countries for the past few years. In arid and semiarid zones, irrigation using groundwater has transformed good quality land with low productivity (caused by drought) into areas of high productivity. Consequently, the income level for the farmers has increased and the rural population base has been maintained (Bernabéu and Serna, 2002).
This exploitation of groundwater is maintained at a sustainable level balancing aquifer abstraction and recharge. Overexploitation of the system has often been a risk in the past, leading to serious environmental damage and contributing to the desertification process. Moreover, the exploitation of this land will eventually prove to be uneconomic in the future (Martín de Santa Olalla and De Juan, 2001). It is therefore imperative that strategies should be adopted to regulate the actual volume of water abstraction in order to adequately sustain the level of groundwater.
At least two conditions are necessary for this purpose (Directive 2000/60/EC):
- –
The participation of stakeholders in the management process; they need to accept that water is a limited resource. A permanent effort needs to be carried out to foster a greater degree of solidarity and co-operation among present and future users.
- –
The use of suitable tools to ensure that the decisions made on the use of water resources will be the most adequate in each individual case. For this purpose, the use of irrigation techniques, particularly those based on models representing the existing relationship among the different variables that affect the hydrological unit may well prove to be of considerable assistance.
This case study was developed within Project MERIT (Management of the Environment and Resources using Integrated Techniques, Ref. EVK1-CT-2000-00085) funded by the European Union from June 2001 to June 2004. The aim of MERIT was to construct Bayesian networks (Bns) for four study areas that have different problems in terms of water use. These networks, regardless of the way in which they may be used in the management of each hydrographic catchment, are intended to act as a pilot experience for the construction of any Bn geared towards water resource management within the framework of the European Union or even elsewhere.
As part of the MERIT Project, the University of Castilla–La Mancha has built a Bayesian network (Bn) applied to water resource management in the Hydrogeological Unit Eastern Mancha (HUEM), which includes within its limits the aquifer known as “Eastern Mancha”.
The aim of this paper is to demonstrate the flexibility of Bn to represent a complex and large aquifer, and analyse its power in helping water resources management. The network had been employed to perform a “what if” analysis of different alternative decisions (see Castelletti and Soncini-Sessa, this issue) and therefore assess their effects in a sustainability perspective, thus to support decision-making. The focus of the paper will be on the use of Bn and the interpretation of the results it provided and not on the network construction process. The paper also shows how to involve stakeholders in building and validating the network for this case study. The process followed to build the Bn has been the focus of another paper (Martín de Santa Olalla et al., in press), while the basics of Bns are provided in the paper by Castelletti and Socnini-Sessa (2006).
The paper shows the processes followed in order to build up the tool: getting the modelling system information, selection of variables, Bayesian networks operation and the stakeholders' participation; as well as the model output results, so as to achieve the sustainable management of water resources in the HUEM.
Section snippets
Description of the Unit and its geopolitical environment
The Hydrogeological Unit “Eastern Mancha” is one of the hydrogeological units of the Jucar System. The main tributary is the Cabriel River that flows directly into the Jucar, and a number of minor rivers which do not flow directly into the Jucar, but through infiltration into the aquifer “Eastern Mancha” and through the latter into the river. The whole Jucar System is regulated by three major reservoirs: Alarcón, Contreras and Tous, all of them located outside of the Hydrogeological Unit
Decision making
The HUEM supplies water for this Unit and for the units downstream. For this reason, the sustainability of the Unit means that, on the one hand, there is enough water for those units and on the other hand, the volume extracted from the aquifer is lower than the replenishment. Actually, achievement of sustainability in the HUEM is necessary to achieve the second condition. The following actions have been considered:
- 1.
Reduction of water volumes for irrigation through exploitation plans (SPEI)
- 2.
Conclusions
The prospects of modelling the integrated management of hydro-geological resources using Bayesian networks are promising. As the processes involved require the use of a large number of variables, which are in turn marked by a high degree of uncertainty, we believe the use of this technique is justified.
This paper provides a practical demonstration of how Bns may support resource managers of the HUEM in the decision-making process. The advantage of this tool is that it can predict the behaviour
References (29)
- et al.
The use of Hugin® to develop Bayesian networks as an aid to integrated water resource planning
Environmental Modelling and Software
(2005) - et al.
Monitoring irrigation water use by combining Irrigation Advisory Service, and remotely sensed data with a geographic information system
Agricultural Water Management
(2003) - et al.
Analysis of uniformity of sprinkle irrigation in a semi-arid area
Agricultural Water Management
(1999) - Aquagest, 2003. Consumo de la ciudad de Albacete y su área de influencia. Internal report. No published. Servicio...
La estrategia de política monetaria del BCE y el marco operativo del Eurosistema
- Bernabéu, R., Serna, J., 2002. Ponencia “Los grandes cultivos herbáceos de regadío en Castilla-La Mancha”. Libro Blanco...
Planning Improvements in Natural Resources Management. Guidelines for using Bayesian Networks to Support the Planning and Management of Development Programmes in the Water Sector and Beyond
(2001)- et al.
GIS tools applied to the sustainable management of water resources. Application to the aquifer system 08-29
Agricultural Water Management
(1999) - et al.
Bayesian networks and participatory modelling
Environmental Modelling and Software
(2006) - CHJ (Confederación Hidrográfica del Júcar), 2004. http://www.chj.es. Accessed February 2,...
Community Involvement in Planning and Development Processes
Introducción al Razonamiento Aproximado. Departamento de Inteligencia Artificial
Cited by (69)
Nitrate-dependent Uranium mobilisation in groundwater
2019, Science of the Total EnvironmentCitation Excerpt :Further details on BNs modelling are provided in the Supporting information. The application of BNs in understanding different environmental systems can be found in literature (Bonotto et al., 2018; Borsuk et al., 2004; Liu et al., 2018; Maeda et al., 2017; Martín de Santa Olalla et al., 2007; Rigosi et al., 2015; Ticehurst et al., 2007; Wijesiri et al., 2018). Accordingly, U mobility in groundwater was visualised as a collection of multiple processes/systems (e.g. redox systems and bicarbonate complexation), rather than considering how each variable influences U concentrations independently from other variables.
Fees and governance: Towards sustainability in water resources management at schools in post-apartheid South Africa
2019, Sustainable Cities and SocietyCitation Excerpt :The main imperative in selecting the appropriate analytical technique was the integration of both, quantitative and qualitative variables (see Table 1 above), which were identified to influence schools’ water usage. In this context, Bayesian Networks (BNs) modelling has proven to be effective in relation to a range of environmental systems/processes modelling (Bonotto et al., 2018; Borsuk et al., 2004; Liu et al., 2018; Maeda et al., 2017; Martín de Santa Olalla et al., 2007; Rigosi et al., 2015; Ticehurst et al., 2007; Wijesiri et al., 2018). In fact, Bayesian statistical methods have gained relatively little attention, although they have been used for scenario-based water demand modelling.
Factors influencing volatile hydrocarbon pollution in urban areas
2019, Emerging ContaminantsHydrology-oriented forest management trade-offs. A modeling framework coupling field data, simulation results and Bayesian Networks
2018, Science of the Total EnvironmentCitation Excerpt :This tool allows for investigating the impacts of management options through analysis of scenarios, assessing the consequences of management hypothesis or potential actions (Pérez-Miñana, 2016). Many recent efforts can be found in the literature on solving complex problems in the field of ecohydrology (Woznicki et al., 2015), water resources (Xue et al., 2017), groundwater management (Martín de Santa Olalla et al., 2007), forestry (Nyberg et al., 2006), irrigation and farming (Wang et al., 2009), climate change (Molina et al., 2013), biosecurity (Lohr et al., 2017), public health (Beaudequin et al., 2016) -and many others-, using BNs. Of special interest some systematic reviews: Aguilera et al. (2011) classified 128 case studies by fields.