Growing populations and climate change place increased pressures on our water supplies. Stormwater harvesting and reuse offers a potential option to augment traditional water resources. Before it can be utilised within a given context, however, its safety must be established. This project developed a Bayesian Belief Network (BBN) model representing pathogen sources and treatment barriers within a proposed stormwater harvesting scheme. The BBN can utilise a range of data sources and be constantly updated to assist managers to engage effectively with stakeholders and identify the most appropriate combination of risk management strategies available to them.

Honours Thesis completed by Dean Albert Mensinga in October 2019.

Lake hydrodynamic models are used by water utilities to provide an estimation of the conditions within a water storage. Increasingly, water utilities are employing a pathogen module to simulate the advection and dispersion of pathogens in source water. Lake model outputs provide forecasted pathogen concentrations across the water column of lakes and reservoirs, including at the offtakes of water treatment plants. This pathogen information, in combination with any monitored pathogen concentrations, provides important inputs to water quality experts to conduct quantitative microbial risk assessment.

The quality and outcomes of a lake modelling process depend on several critical steps, some of which are implicitly used by modellers, but which are often not adopted routinely or communicated widely as part of a quality assurance process. There is an emerging need in the drinking water supply industry for information and guidance on the appropriate use of lake models to support QMRAs. Ideally, the information is provided through a set of best practice guidelines that demonstrate a series of quality assurance principles and actions to ensure that model development, implementation and application represent best practice and are commensurate with the intended purpose. This project will identify, review, and communicate guidance on specific modelling steps, such as project administration, conceptual modelling framework, model evaluation metrics, and uncertainty analysis practice, and prepare consultative draft of the detailed best practice guidelines.

Lakes and reservoirs are essential for water supply for humans and agriculture, and have an important role in flow regulation, biodiversity, and streamflow below dams. Australia has been subject in recent decades to severe drought which has heightened the importance of reservoirs for human populations and highlighted the need for careful management of water levels to maintain continuity of supply. Climate change is likely to exacerbate water shortages, with extended periods of drought, interspersed with more discrete and intense rainfall, leading to challenges for storing water in reservoirs and potentially affecting the quality of water.

In this project, Griffith University researchers examined the water quality risks from low and variable water levels in dams and reservoirs in Eastern Australia.

Predicting the effects of climate change is a complicated business. A climate change model might relate temperatures to air flow and patterns of rainfall in a defined geographical area, important information for planning the construction of future dams and water storage reservoirs. The problem is that it is difficult to extract and transfer knowledge from the climate model into the planning and design process; there is a gap between climate science and those who need to apply the knowledge. This project bridges that gap through the establishment of a suite of strategies to further improve the ability of decision-makers to understand and integrate climate change into their ‘business-as-usual’ water planning and management practices.

The development of a toolkit of climate change adaptation supporting resources, called ‘Resiliwiki’, provides Water Research Australia members a valuable resource, allowing them to rapidly identify appropriate decision evaluation methods for a given decision. Although aimed at the water industry, this decision framework for transferring climate change science into adaptive practice is also valuable for government organisations and environmental, health and economic planners.

Wastewater (WW) contains harmful chemicals, including pesticides, that can disrupt normal gene function or hormone activity. The cost of measuring each separate contaminant at the frequency needed to demonstrate the safety of recycled WW is prohibitive. This research reviewed the risk assessment and regulation of chemicals in Australian water, with a focus on ‘thresholds of toxicological concern’. Laboratory techniques were developed to extract and concentrate WW contaminants into solutions suitable for analysis using both new in vitro cell culture assays and analysis in expensive, established chemical tests. WW and treated samples were collected from nine Australian water reclamation plants. The total effect of each sample (which contained a mixture of contaminants) on cell death, gene integrity and aspects of liver, hormone, nerve and immune system activity, was determined using in vitro cell culture bioassays, and compared with the classical chemical measurement of each separate contaminant. The cheaper cell-culture tests correlated well to the levels of groups of chemicals and could be used to find thresholds of toxicological concern. Both testing regimens also demonstrated that reverse osmosis is a highly effective method that removes harmful chemicals to levels much lower than those designated safe by regulatory authorities.

E. coli bacteria naturally populate the gastrointestinal tract of humans and animals; they are usually harmless and are commonly excreted. Faeces can also contain harmful microscopic pathogens, and this has led to the assumption that if harmless E. coli are found in water, that the drinking water has been contaminated with faeces that might also have contained pathogens that pose a risk to public health. Using E. coli as an indicator of faecal contamination was recently challenged by the finding that some E. coli strains live, grow and bloom in the environment, and their presence in water might not mean that the water has been contaminated with harmful pathogens. This research examined the environmental conditions associated with E. coli bloom formation in the context of climate-change adaptation and developed multiplex PCR tests which allow the identification of environmental and faecal E. coli. This information was added to a Utility Response Protocol.

A survey of water utilities identified the top five challenges faced in daily operations, and technical, economic and literature reviews identified remote sensing strategies and technologies to address these five operational issues. The need for, and benefits of real-time monitoring which facilitates cost-effective and efficient responses to rapidly changing conditions, were common to all five challenges, which were: monitoring cyanobacteria and their metabolites, the effects of contamination and extreme climate-change driven events on water quality, the variation in climatological data over relatively small distances (need to increase focus, precision and produce ‘finer’ datasets for reservoir management), asset inspection and management and monitoring catchments for a variety of factors. This research evaluated and explained the solutions, strengths, weaknesses, and costs of products best suited for addressing each challenge.

The ADWG has methods for predicting risks to water quality, but these were not developed for managing extreme climate-change driven weather events such as bushfires or floods. This research developed a risk assessment tool for managing water-related health risks associated with extreme weather events. Real-world datasets and experience of water cloudiness (turbidity), colour and blue-green algae were used to create and validate environmental models which were developed further by applying Baysian network and System Dynamics concepts. This iteration of the model was not constrained by, and did not reflect existing risk profiles, but was judged to be flexible enough to provide a realistic representation of future hazards arising from extreme weather events.