Author: Ian McDonald
McDonald, Ian, 2020 GIS-enabled water quality risk assessment in a drinking water catchment, Flinders University, College of Science and Engineering
Terms of Use: This electronic version is (or will be) made publicly available by Flinders University in accordance with its open access policy for student theses. Copyright in this thesis remains with the author. You may use this material for uses permitted under the Copyright Act 1968. If you are the owner of any included third party copyright material and/or you believe that any material has been made available without permission of the copyright owner please contact copyright@flinders.edu.au with the details.
South Australian Water (SAW) provided a well-established environment and workflow to model the movement of Cryptosporidium from agricultural sources in the Mount Lofty Ranges (MLR) to primary water storages. This environment included a methodology for maintaining project data in Excel spreadsheets. The modelling system was ArcMap and its Raster Calculator was used to implement the model. SAW users expressed frustration with the Raster Calculator approach and its lack of record-keeping facilities.
This project sought to address these, and other, issues through the development of a decision support infrastructure. Superficially it comprises a simple interface for data maintenance and reporting. Underlying this is a suite of support functions that enable model definitions in a spreadsheet and facilitate the record-keeping necessary for long-term sensitivity studies. The development environment for the project was Python and the script was written in such a way that it could readily accommodate different pollutants and new model definitions.
The new model added to the capabilities of the previous one by:
• accommodating the absence of pathogens
• focusing on just that part of the catchment, i.e., a watershed, that supplied water to the monitoring stations
• introducing a flow parameter that was dependent on the watershed definitions
• providing for interactions between variables in addition to simply adding their effects.
The project has developed the position that flowing water is required to deliver pathogens and nutrients to the SAW monitoring stations. As defined in previous SAW work, the main drivers for flowing water were slope and rainfall.
Previously the risk evaluated at a cell was determined solely by the values of the contributing variables at that cell. Now, a new spatial characteristic, flow, has been introduced. Its value at a cell is influenced by what is occurring upstream of that cell. This variable has been employed in two ways, either to add to the risk like other terms, or to interact with the source and transport terms to influence their contributions.
The model has been modified considerably during this project. The generation of negative risk values has been an ongoing concern. This has been addressed by reducing the impact of the sources of negative risk – buffers and fencing. Also, the growth of the land use file from 10,000 rows to over 44,000 rows changed the disposition of risk significantly. In the light of changes to the model, and weight and risk estimates, it is recommended that the system now go back to SAW for a comprehensive review by the catchment management scientists.
Keywords: water, pollution, risk, model, Python, GIS
Subject: Environmental Science thesis
Thesis type: Masters
Completed: 2020
School: College of Science and Engineering
Supervisor: Professor Howard Fallowfield