9.4.3 Critical limits at critical control points

For preventive measures identified as the critical control points for the water supply system, critical limits must also be defined and validated. A critical limit is a prescribed tolerance that distinguishes acceptable performance from unacceptable performance at a critical control point in terms of hazard removal or attenuation. Critical limits often incorporate both a numerical limit and a delay period, to ensure that alarms are not activated inappropriately (Mosse 2009). The delay period should be appropriate to the system: if the delay period is too short, repeated false alarms could lead to alarms being ignored or inactivated; if it is too long, a significant quantity of potentially unsafe water could be supplied to consumers before an alarm is activated.

Critical limits and the delay period are system-specific and determined through balancing the capability of the instrumentation and equipment used in the alarm system with the risk characteristics of the water supply system (e.g. whether treated water is supplied direct to consumers, or a storage exists before consumers receive the water).

Breaching a critical limit represents loss of control of the process and the existence of a health risk, either directly through the supply of unsafe water, or indirectly, where multiple critical control points exist, by exceeding the capacity of subsequent processes. Such events should result in immediate corrective actions to re-establish operations within specification, and notification of the relevant health authority or drinking water regulator.

Setting target criteria that are more stringent than the critical limits at critical control points will enable corrective actions to be instituted before an unacceptable health risk occurs. Exceeding a target criterion at a critical control point would generally not require that the relevant health authority or drinking water regulator be notified, providing corrective action successfully prevents deviation from a critical limit.

Box 9.5 provides an example of setting operational requirements for filtration as a critical control point.

Box 9.5

Target criteria and critical limits for filtration

Where drinking water is sourced from multi-use surface water with risk of contamination by Cryptosporidium, filtration is often the primary barrier to these chlorine-resistant protozoan pathogens. It is critical that filter performance be optimised and continuously monitored to ensure that the required pathogen removal is achieved and safe drinking water provided at all times.

Whilst not a perfect measure of performance, continuous monitoring of filtered water turbidity is currently the best practical surrogate for assessing filter performance. It is strongly recommended that continuous on-line turbidity meters be installed on the outlet of each individual filter, as monitoring only at the combined filter outlet may fail to detect poor performance of an individual filter (Mosse & Murray 2008).

With filtration defined as a critical control point, a critical limit is set to define unacceptable performance contributing to a significant health risk (e.g. 0.5 NTU). Measured turbidities above this limit indicate loss of control of the process and compromised pathogen removal. To ensure critical limits are not breached, target criteria should also be established. A target criterion for filtration may be to achieve <0.2 NTU. To avoid unnecessary alarms, target criteria should incorporate a delay period in which the criteria are continuously breached before an alarm is generated (Mosse 2009).