Sensitivity features of aquifer parameters and their implications on parameter estimation

Jiao, Jiujiu (1993). Sensitivity features of aquifer parameters and their implications on parameter estimation. University of Birmingham. Ph.D.

Full text not available from this repository.

Abstract

The sensitivity features of aquifer parameters and their influence on parameter estimation are investigated by coupling flow models with optimization methods in terms of pumping-test analysis in various aquifer configurations. Pumping tests are the most popular methods for estimating aquifer parameters. However, many features relating to aquifers and pumping wells are difficult to include in analytical solutions. The influence of these features - such as boundary effects, aquifer heterogeneities, additional recharge and well storage - on parameter estimation is examined.

The significance and usefulness of traditional pumping tests in representing the flow properties of non-uniform aquifers are challenged. These pumping tests create a very strong convergent flow towards the central pumping well, and information about the whole aquifer is so highly averaged that the information of a local zone can be easily obliterated. The parameters estimated from a pumping test may thus not reflect the influence of a local zone. Therefore, satisfactory fitting of drawdown by a model with uniform parameters does not prove that the aquifer is homogeneous.

The different hydraulic features of transmissivity and storativity are explored. Storativity has the features of two-way coordinates, but transmissivity has the features of one-way coordinates - its information can be mainly transferred from upstream to downstream, not the other way round. In terms of parameter estimation, an upstream well can produce information on storativity of both upstream and downstream, but it can produce little information about the transmissivity downstream.

The implications of sensitivity features in the design of pumping tests and the identification of different parameter zones for regional numerical modelling are discussed. A good pumping-test design should be such that parameters are most sensitive and least correlated during the pumping period; in a regional model, if the sensitivity of head at locations of interest to the parameters of a zone is very small, the existence of this zone is unrealistic and should be lumped into a neighbouring zone.

The pumping tests in a Nottinghamshire aquifer are used as the main case study. Both analytical and numerical models are used to estimated parameters, with emphasis on understanding the sensitivity features of parameters. A new equation is developed to estimate additional recharge during the pumping period.

Type of Work:

Thesis (Doctorates > Ph.D.)

Award Type:

Doctorates > Ph.D.

Supervisor(s):