Spatiotemporal modelling of groundwater contaminants

Molinari, Daniel Alberto (2014) Spatiotemporal modelling of groundwater contaminants. PhD thesis, University of Glasgow.

Abstract

Spatiotemporal data have become very common, particularly through environmental
settings where a spatial array of sampling sites generates data over time. This
thesis deals with a specific spatio-temporal setting of groundwater contamination
and aims to construct suitable statistical models. One of the motivating features
of the application is that the model has to be implemented in an unsupervised
manner and there is a high premium on the results being available very quickly,
with a response time of a few seconds only.
Many routes to spatiotemporal models are possible, but in order to achieve the
aims outlined above we have proposed a model based on P-splines. A Bayesian
approach to fitting is used to provide the stability required in an unsupervised
setting. The speed requirement makes computationally intensive methods such
as MCMC unsuitable for the determination of the optimal penalisation parameter
and so conjugate priors and highly efficient methods of linear algebra have been
brought to bear.
Use of the model identified a problematic issue due to the irregular spatio-temporal
design of some data sets, giving rise to cases of \ballooning", where unexpectedly
high predictions, not supported by the observations, can appear. This matter
was also tackled within the Bayesian framework mentioned above. The proposed
procedures were assessed both by means of a simulation study and on real data.
Finally, as an extension of the proposed methodology, we address the issue of
non-detects, namely observations which are known only to lie below some limit of
detection. The task is accomplished using a Laplace-type approximation to the
posterior distribution of the parameters and the suitability of this approximation
is analysed through examples.
The problems addressed in the thesis are motivated by the need to ensure environmental
quality in and around installations operated by the multinational company
Shell. The assistance of Shell in advising on the context of the issues, and in providing
data sets for case studies, is much appreciated.

• Item Type: Thesis (PhD)
• Qualification Level: Doctoral
• Keywords: spatiotemporal, groundwater, contaminants
• Subjects: Q Science > QA Mathematics
• Colleges/Schools: College of Science and Engineering > School of Mathematics and Statistics > Statistics
• Supervisor's Name: Bowman, Professor Adrian and Evers, Doctor Ludger
• Date of Award: 2014
• Depositing User: Mr Daniel Alberto Molinari
• Unique ID: glathesis:2014-5873
• Copyright: Copyright of this thesis is held by the author.
• Date Deposited: 20 Jan 2015 15:06
• Last Modified: 21 Jan 2015 09:34
• URI: https://theses.gla.ac.uk/id/eprint/5873