Skip to main content

The Treatment of Uncertainty in Groundwater Flow and Solute Transport Modelling

Objective

Analysis and understanding of the groundwater flow in the neighbourhood of a radioactive waste repository play important roles in a performance assessment. Such analysis rely on numerical modelling in order to study the flow and transport over the very long times that must be considered. Two vital issues which must then be considered are the way in which the available data are used in constructing the mathematical model of the site and the uncertainty that is implied in the results of the model by uncertainties in the model parameters and in the model itself. The two tasks in this project address these issues.
Analysis and understanding of the groundwater flow in the neighbourhood of a radioactive waste repository play an important role in a performance assessment. Such analyses rely on numerical modelling in order to study the flow and transport over the very long times that must be considered. 2 vital issues which must then be considered are the way in which the available data is used in constructing the mathematical model of the site and the uncertainty that is implied in the results of the model by uncertainties in the model parameters and in the model itself. The 2 tasks in this project address these issues. The first task is concerned with the investigation of novel approaches to the construction of mathematical models of a site. The second task with the investigation of methods for the estimation of uncertainty in groundwater flow and transport calculations.

An initial review of methods which can be used to represent a site in a numerical model was carried out. Indicator methods appear to be the most promising. Indicator methods have been used for several years in mining studies and are finding increasing application in the water resources and petroleum industries.
The Gorleben site, which consists of a sedimentary sequence overlying a salt dome, was identified as suitable for use in this project, in part because the data set contains a considerable amount of lithological information.
Initial test calculations were then carried out to investigate the usefulness of the proposed method. A simple code has been written to transform the petrographic data for the site into a form suitable for use in a variogram estimation code.
Work programme:

The first task is concerned with the investigation of novel approaches to the construction of mathematical models of a site. The second task is concerned with the investigation of methods for the estimation of uncertainty in groundwater flow and transport calculations.

Task 1 Site Characterisation
1.1 Site Models
Mathematical models that can be used to represent a site will be reviewed. A computer package based on a selected method will then be written.
1.2 Inverse Problem Techniques
Methods for the inverse problem will be investigated and a selected approach will be applied in conjunction with the site models developed in Task 1.1.
1.3 Effective Properties
Methods for relating measured data to model parameters will be investigated using numerical calculations, based on the site models developed in Task 1.1.

Task 2 Treatment of Uncertainty
2.1 Methods for the Treatment of Uncertainty
Methods for the estimation of uncertainty in the results of groundwater flow and transport models will be investigated.
2.2 Parameter Sensitivity
Methods for the estimation of sensitivity coefficients will be investigated. The coefficients will be used to study the topic of data worth.
2.3 Model Uncertainty
The study of this topic will build on experience of "what-if" studies and will pay particular attention to long-timescale changes in hydrogeology.

Funding Scheme

CSC - Cost-sharing contracts

Coordinator

United Kingdom Atomic Energy Authority (UKAEA)
Address
Harwell Laboratory
OX11 0RA Didcot
United Kingdom