Safeswim uses water quality models to predict when the concentration of faecal indicator bacteria (FIB) in the water is above predefined thresholds that pose an elevated risk to health from swimming.

The models use the results of water quality sampling and their relationships with environmental factors to predict water quality in real time. The models consider rainfall intensity, duration, and location, as well as tide, wind speed and direction, and sunlight, and are calibrated to each individual beach.

Three different types of models are used – criteria models, 'black box' models, and 'white box' models – which differ in sophistication, functionality and information needs:

• Criteria models produce a unitless water quality index (WQI), which is used to generate alerts when over a certain threshold, based on rainfall triggers, previous FIB results, and time of day. Criteria models are the least sophisticated of the three model types and, while easy to develop, provide only simple information.
• ‘Black box’ models are statistical models that predict FIB concentration based on empirical relationships between FIB and other variables such as rainfall and tides. The black box models are developed using software published by the US Environmental Protection Agency (EPA).
• ‘White box’ models estimate FIB concentration by simulating the generation, movement, dispersion, and die-off of FIB in the environment. White Box models are the most sophisticated, and hence demanding to develop, and provide more dynamic information that has greater utility (e.g., to test infrastructure improvement options).

The model type used at any location is designed to utilise the information available for that location to ensure the best possible health risk advice is provided to beach users.

Auckland and Northland Regional Councils have invested in targeted sampling to support the continual development and improvement of predictive water quality models used in Safeswim. Targeted sampling has allowed a gradual transition from criteria models to more sophisticated black and white box models over time as it more accurately represents the real water quality at beaches (compared with weekly monitoring).

Find out more:

• How contaminants move in the ocean: Principal Coastal Scientist Ben Tuckey from DHI uses two animations to show how environmental conditions affect the movement of contaminant plumes in the ocean, demonstrating why a beach might be a ‘red alert’ or ‘green alert’ under different circumstances.
  
  
• How the Safeswim models work: Principal Coastal Scientist Ben Tuckey from DHI explains how the Safeswim models work, what the Safeswim models allow, how they are built and improved with sampling data, and how we know their predictions are accurate.