Water treatment plant operators remove cyanobacteria and the toxins they produce from source waters but calculating the amount of treatment needed for effective removal is difficult, particularly in bloom conditions when cyanobacterial cell numbers and toxins change quickly. Current cell counting and toxin measurement can take hours or days to complete, and the results are not available quickly enough to help treatment plant operators respond to changing conditions. There is a need for a real-time method that gives instant results. This research examined the utility of fluorometers; probes that emit light that is ‘reflected’ back at different wavelengths by living cells and other matter in the water and is detected by the fluorometer. It was found that when only one species of cyanobacteria was present, there was a good correlation between the fluorescent signal and cell number, particularly when source waters were clear and not cloudy. Cell numbers did not relate well to levels of toxins or taste and odour compounds. When fluorometers were installed in 13 water treatment plants the correlation between cyanobacteria cell numbers and fluorometer signals was validated, and this led to the conclusion that fluorometers can give early warning of blue-green algae blooms.

‘Microbial source tracking’ (MST) is a technique that aims to identify the animal that excreted faeces and polluted water. There are a number of ways to do this, but the problem is that no one method accurately identifies the origins of faecal pollution in environmental water samples. This research found that faeces could be stored in a freezer or a laboratory -80°C cold-store for up to a month without changing the relative numbers of the different types of bacteria in the samples of faeces. Up to seven faeces samples from different animals were mixed together and examined using 17 techniques to identify the original animals. Three of the most accurate and reliable methods used mitochondrial DNA, the analysis of a bacterial enzyme sequence (beta-glucuronidase), and specific DNA sequences form bacteria known to come from humans, horses and cows. These three types of tests were selected for inclusion in a ‘Toolbox’ from which a combination of methods will allow accurate and reliable management of faecal contaminants in source waters.

This research has provided the most comprehensive account of the geographical distribution of blue-green algae (cyanobacteria), and the toxins they produce, in Australia. The blue-green algae cells were collected and stored. This collection now forms a valuable national asset which is particularly valuable for managing the complex array of factors that affect the accurate assessment of risk posed by any one algal bloom. Not all cyanobacteria produce toxins, and the identification of species and the presence of toxin is an important step in the decision-making process necessary to produce high quality, safe water. This research led to some notable conclusions; one being that a traditional method that uses cell-shape to identify algal species is unreliable, and also that the number of cyanobacterial cells does not necessarily correlate to the amount of toxin in source waters. Five laboratory tests were reviewed and it was found that tests for cylindrospermopsin were reliable, but tests for microcystin and saxitoxins differed as to the amount they measured, although they reliably identified the presence or absence of toxin. Problem cyanobacteria species are ubiquitous in Australia and if climatic events create favourable conditions, blooms can occur in unexpected locations.

It is prohibitively expensive and time-consuming to monitor drinking water by individually quantifying every possible polluting contaminant. Instead, living cells can be cultured in samples of the test water and if one or more toxic contaminants are present the cells die. Less toxic, more subtle effects can also be measured, for example, a toxicity test that uses cells collected (decades ago) from one monkey kidney, quantifies cell death but also measures the amount of protein made by each live monkey kidney cell. The problem with this, and other toxicity tests, is that chlorine disinfectants and harmless low levels of other substances, such as aluminium or copper, which occur naturally in water, can sometimes have inhibitory effects on the ‘bare’ cells that are often more sensitive when cultured inside laboratory culture vessels than when they were in a normal situation within a body. This research identified commonly used toxicity tests that are not affected by disinfectants or naturally occurring harmless substances, and also worked out some solutions that quench disinfectants and allow cost-effective and useful cell-based toxicity tests to be used to monitor the safety and quality of drinking water.

This paper discusses various water quality risk management techniques and proposes a step-by-step catchment risk assessment methodology that is compatible with the Australian Drinking Water Guidelines.