Trends of organic carbon in Scottish rivers and lochs

Reid, Stephen James (2011) Trends of organic carbon in Scottish rivers and lochs. MSc(R) thesis, University of Glasgow.

Abstract

In Scotland, the Scottish Environment Protection Agency (SEPA) is the regulatory agency responsible for monitoring water quality and reporting back to the Scottish and UK governments and the European community. In order to improve water quality in surface waters such as rivers and lochs, the European Parliament has established directives over the past twenty years outlining targets for nutrient levels and water quality status. Moxley (2010), states that the concentration of organic carbon in many Scottish rivers, has approximately doubled over the last twenty years, with soils being the most likely source. According to Moxley (2010), the rate of total organic carbon (TOC) increase, averaged across all sites with increasing concentrations, was 0.12 milligrams per litre per year (mg/l/y). This is an increase in TOC concentration of nearly 2.5 mg/l over a twenty year period. Consequently, the behaviour of organic carbon in Scottish rivers and lochs has become of interest and is the focus of analysis within this thesis.

Chapter 1 introduces organic carbon, providing an insight into observed trends in the United Kingdom, but also, other parts of the world. Furthermore, Chapter 1 discusses environmental and physical factors which are thought to be associated with changing levels in organic carbon. Moreover, Chapter 1 provides a description of the data and sampling techniques which have been used. The exploratory analysis in Chapter 2 reveals that the log TOC levels in rivers and lochs have been increasing up until the early 2000’s, and that the log TOC follows a seasonal pattern. Furthermore, the exploratory analysis reveals the high level of association between total organic carbon and dissolved organic carbon. The exploratory analysis also highlighted issues with the covariates; therefore Chapter 2 explores suitable methods of dealing with values at the limit of detection, as well as appropriately imputing missing values.

Chapter 3 explores log TOC at a selection of river and loch sites, and the relationship between log TOC and covariates at each site in detail. In addition, Chapter 3 explores the use of different regression techniques (e.g. linear and additive modelling) in order to appropriately capture the behaviour of log TOC at each site. Chapter 4 progresses from investigating and modelling individual sites, to exploring sites which are connected in some manner. Chapter 4 considers the behaviour of log TOC in sites which are part of the River Dee network, where the distance between each site and how the river flows between each of the sites had to be taken into consideration. Chapter 4 investigates the behaviour of log TOC across the river network over time and space visually; but, also explores appropriate modelling techniques which suitably capture the behaviour of log TOC over time and space, taking into consideration suitable covariates to plausibly explain the observed trends.

Chapter 5 addresses the main theme of the thesis: coherency is defined and explored there. A literature review was conducted to consider possible methods of measuring coherency. The seasonal Mann-Kendall test was found to be an appropriate method of measuring the heterogeneity of a group of sites; and dynamic factor analysis was found to be an effective technique of identifying common trends in a group of sites; hence, these methods were applied in Chapter 5 to measure the level of coherency between sites in the River Dee network, but also, sites located in the same Scottish region. Progressing from the analysis carried out in Chapter 5, Chapter 6 aims to appropriately model the levels of log TOC in Scottish regions, taking into account time and space, but also, possible covariates thought to be driving such trends. Finally, Chapter 7 provides a summary of the findings, and discusses limitations of the study and possible areas of future research.

• Item Type: Thesis (MSc(R))
• Qualification Level: Masters
• Keywords: Total organic carbon, coherency, trends, common signals, generalized additive models, euclidean and river distance, Mann-Kendall test, dynamic factor analysis
• Subjects: H Social Sciences > HA Statistics
• Colleges/Schools: College of Science and Engineering > School of Mathematics and Statistics > Statistics
• Supervisor's Name: Scott, Prof. E. Marian
• Date of Award: 2011
• Depositing User: Mr Stephen James Reid
• Unique ID: glathesis:2011-3379
• Copyright: Copyright of this thesis is held by the author.
• Date Deposited: 17 May 2012
• Last Modified: 10 Dec 2012 14:06
• URI: https://theses.gla.ac.uk/id/eprint/3379