Statistical regression models and control charts for the Real Time Water Quality Network in Newfoundland

• Main Author: Harvey, Robert Richard
• Format: Thesis
• Language: English
• Published: Memorial University of Newfoundland 2009
• Subjects: Newfoundland
• Online Access: https://research.library.mun.ca/9286/; https://research.library.mun.ca/9286/1/Harvey_RobertRichard.pdf

The Newfoundland and Labrador Real Time Water Quality (RTWQ) network was established by the Water Resources Management Division (WRMD) of the Newfoundland and Labrador Department of Environment and Conservation in 2001. Digital sensors continuously record water temperature, pH, specific conductance, dissolved oxygen and stage level for the rivers and streams in the network. This technology is still a relatively new and unfamiliar approach to collecting water quality data in the province. This thesis presents the complete findings of research and development carried out to further enhance the WRMD's capability to work with the RTWQ data in new and innovative ways. -- Statistical regression models for predicting water temperature and dissolved oxygen levels are developed for RTWQ stations owned and operated by the WRMD: Humber River, Peter's River, Leary's Brook and Waterford River. A logistic S-shaped model using air temperature can accurately predict water temperature. An exponential model using water temperature can accurately predict dissolved oxygen. Investigations are also carried out into developing statistical regression models using RTWQ data as a surrogate for grab sample chemical concentrations (alkalinity, chloride, etc.). There is more potential to develop these grab sample regression models for urban rivers than there is for rural rivers. -- Investigations are made into designing statistical process control charts for monitoring the RTWQ data. Modifications to the traditional process control chart need to be made so that it can monitor the highly autocorrelated RTWQ data.