Stochastic Analysis of Well Head Protection and Risk Assessment (W-SAHaRA)

a project sponsored by the EU (5^th framework)

CONTRACT N°: EVK1-CT-1999-000041 W-SAHaRA

Problem to be solved
Major problems that European drinking water companies and Environmental Agencies are facing are (a) the development of tools to reliably predict the pattern of the groundwater flow due to drinking well fields operating in heterogeneous formations and (b) the definition of a reliable strategy for the quantification of the risk associated to such predictions.  Regulators now understand that once a portion of an aquifer has been severely contaminated its strategic importance is compromised and protection of such resources by the reliable prediction of flow field around pumping stations has a strategic impact at a European level.  Traditional deterministic models inherently offer predictions of undetermined quality.  The W-SAHaRA Consortium is motivated by the need to recognise the importance of spatial heterogeneities and related uncertainty, and incorporate these elements into a comprehensive action aimed to the development of general and robust criteria for an efficient and cost-effective planning and management of groundwater drinking well fields.  Many of the techniques being developed within the scope of this project are (in principle) amenable to application to a wide range of problems involving the impact of groundwater pollutants on the environment.

Scientific objectives and approach
Objectives include: (a) development of a methodology for the quantification of the concept of vulnerability of groundwater drinking wells, in a probabilistic framework; (b) application of this strategy to a specific situation and a selected site and simulation of the impact of the obtained solution in decision-making policies; (c) definition of the philosophy of risk assessment for well fields in a stochastic context and (d) production of guidelines for drinking water companies and Environmental Agencies on how to reduce uncertainty on heads/fluxes prediction by geological field investigations and monitoring of groundwater heads/concentrations.  We will attack the problem on different fronts: (a) theoretical/conceptual; (b) laboratory and field scale; (c) deterministic and stochastic model development of synthetic and real-world cases.  Numerical Monte Carlo (MC) techniques will be employed, addressing issues such as (a) development of efficient algorithms for MC simulations of catchments in 3D using alternative methods of particle tracking and Kolmogorov backward equation; (b) clarification of conditioning formalisms, through analysis of the importance of conditioning data in order to reduce the uncertainty of well catchments; (c) adoption and development of inverse methods that can use a variety of data types to decrease uncertainty about aquifer properties and wellhead regions of influence.  Algorithms for optimum unbiased prediction (together with corresponding prediction errors) of heads and fluxes around wells will be developed according to the novel nonlocal formalism of groundwater flow moment equations.  Laboratory and field data will be collected and analysed.  A traditional deterministic model will be applied to a selected field situation and results will be evaluated in light of probabilistic concepts.  Our new numerical schemes and methodologies will be demonstrated in selected practical situations.

Expected impact 
Expected results from theoretical/conceptual work and laboratory/field applied experiments will (a) increase the ability of predicting groundwater heads / fluxes in exploited aquifers, together with related uncertainty bounds, allowing to quantify the degree of environmental risk connected to a specific pumping / injection situation; (b) allow a rational planning of new avenues for exploitations of the groundwater resource, with respect to the safe definition of new sites for drilling and exploitation of drinking wells for urban use, and inclusion / elimination in a urban-development planning action of a series of restrictions / bans for land use; (c) define the settings of requirements for strategic groundwater monitoring in Aquifer Protection Areas: this will be achieved by a dynamic planning of defense of existing well fields; and (d) set new standards for field/laboratory experiments in randomly heterogeneous formations.

Duration 
36 months (starting April 2000)

Participating organizations 
(Project Co-ordinator: prof. A. Guadagnini - Politecnico di Milano)