Data Quality for Water Management

Purpose

To highlight the continuing need to ensure that the integrity of data quality obtained for the purpose of establishing and publishing meaningful information for investment purposes, performance assessment and beneficial changes to the environment and public health is fit for purpose.

Context

When reporting upon the quality and sufficiency of potable water supplies, associated discharges to the aquatic environment, the disposal of wastes arising from the production of drinking water, the collection and treatment of domestic sewage (including trade waste discharges), water used for a variety of industrial and food product purposes, data must be of the requisite standard to satisfy:-

• corporate and executive management of associated industries, 
• water industry, food and environmental regulators
• customers, and
• relevant stakeholders, particularly those responsible for public health and future investment planning.

The correct use of such data, taking full account of the specific purpose for which the data was originally obtained is of critical importance.  Although most comments derive from laboratory testing and water quality, they are equally applicable to measurements of flow rates and quantities.  Establishing data quality is also important when, for example, assessing general performance levels; Health and Safety records and complaints.

CIWEM's Position on Data Quality

1. A significant proportion of the data assembled by water utilities relate to water quality. Carefully designed monitoring programmes should be maintained to establish water quality at appropriate critical and sequential points in:-  
   1. the abstraction, production and distribution stages of potable water supplies;
   2. the collection and treatment of domestic sewage, industrial and commercial trade wastes, and their subsequent discharge to receiving waters.
   3. environmental waters that may be used for leisure, recreational, irrigation, industrial and transportation purposes.
2. Such monitoring programmes should have three fundamental but interlinked objectives:- 
   1. Regulatory - programmes to demonstrate compliance with the conditions of potable water regulations, product quality standards, process permits, aqua-environment standards and statutory instruments.
   2. Operational or process control - wide-ranging variable programmes aimed at observing and as appropriate, facilitating effective control of the various stages of treatment processes that influence potable water, sewage and trade effluents product and by-product quality.
   3. Investigative - specifically designed initiatives and projects to identify causes of ineffective or spasmodically problematic performance levels of treatment processes, the variability of product quality, influent and effluent quality, adverse trends in water resources, environmental damage, customer complaints and "incidents" concerning public health.
3. The objective of the monitoring programme must influence the selection of both the frequency and location of sample collection including the secure and recorded delivery of samples to receiving laboratories. This is also of critical importance in choosing either manual or automatic sampling methods - particularly with regard to the latter when due consideration must be given to utilising time; volume; or flow -proportional collection methodology.  Details of sample collection must be included in the purpose codes that identify each sample with the result(s) obtained from its subsequent examination or analysis.
4. The variable(s) of interest or concern must be clearly defined, to subsequently establish the analytical methodology that will yield meaningful data. This may need full discussion between those responsible for implementing the monitoring programme and the provision of analytical services - particularly when the latter are procured externally.  Likewise, distinctions between dissolved and "particulate" forms of relevant determinands must be delineated where appropriate.
5. The choice of analytical methodology must reflect the anticipated ranges of concentrations of determinands that are likely to be encountered.
6. The analytical methodology used by laboratories to determine product, effluent and general water quality, must be fully described in the Quality Control Manual that is specific to individual laboratories. Details should be provided of:
   • the principles and performance characteristics of each analytical method used to determine chemical, microbiological, biological, microscopic and radiological content;
   • the analytical procedures;
   • the reagents, apparatus and equipment;
   • the collection, preservation and storage of samples;
   • explanation of how the analytical result is expressed;
   • An assessment of the variability of the results obtained and published by reviewing the sources and types of errors that may influence the said results.
     
     The Quality Control (QC) procedures applicable to specific batches of samples and determinants must be published in this Manual together with any relevant published references to the protocol of individual determinations.
7. Laboratories must have in place an Analytical Quality Control system that is capable of determining the quality of its published water quality and associated data.  This is demonstrated by continually monitoring the reliability of the results obtained having regard to the use of reliable, effective equipment, high-quality reagents and, properly trained competent analysts.  All such laboratories must have precise knowledge of the accuracy and precision of their analytical data.
8. 'Responsible' analysts must manage all fields of enumeration that lead to the publication of water quality data.   Being appropriately qualified and able to demonstrate competencies in the water examination techniques used, especially method development and the ability to effectively investigate errors is crucial to the production of reliable data.  Furthermore, laboratory personnel supervised by the 'Responsible' analyst must be trained and experienced in the correct use of relevant laboratory equipment, instrumentation, the application of appropriate laboratory techniques, with proficiency in data handling.
9. Representation of accuracy, precision, repeatability and reproducibility must remain self-evident. A well-defined Analytical Quality Control (AQC) regime incorporating the use of suitable control charts is essential in obtaining reliable water quality data. The responsibility for managing the AQC programme should ideally rest with a competent scientist who is wholly independent of the management structure of the laboratory.
10. Random audits are essential to confirm that all laboratory generated data are recorded, with errors correctly identified (when they occur), acknowledged and signed.
11. A quality control, maintenance and calibration programme must be implemented for process instrumentation yielding "on-line" measurements for process control and observation purposes. Scale length, drift and concentration spans are important aspects of the calibration and operational control of such instrumentation to enable the ensuing data to be meaningful.  It is important to understand the manufacturers' specifications for such instrumentation and to fully acknowledge operating protocol for the validity of the instrumentation to be verified, or justified, as appropriate.
12. It is essential that manufacturers and suppliers of analytical and process (on-line) instrumentation be alerted to any anomalies that may be experienced particularly if deviations from specified operating characteristics are observed.
13. The comparison of data sets derived from different analytical techniques requires care and attention.  It should not be assumed that new analytical methodology will display lower or less errors than well used and practised "traditional" methods.  The newer methods may be preferred on the basis of ease of automation, speed, cost and possibly, lower limits of detection.
14. Data assembled for purposes other than water quality must also be subject to critical scrutiny.  Typical examples where data are recorded for specific reasons are:
    • Performance indicator systems
    • Complaints
    • Health and Safety trends and records.
      
      All data should be confidence graded.  Both the reliability of the source and the accuracy of the data should be assessed and specified.  Whereas the accuracy accounts for uncertainty in the acquisition or measurement of input data, the reliability will take account of the extent to which the data sources yield consistent, stable and uniform results, from repeated observations or measurements made under the same conditions each time.  Document BS ISO 24510:2007 gives examples of bands of both the accuracy and reliability of data, highlighting the ever present need to assess data source reliability and data accuracy for each input variable.
15. The forgoing paragraphs emphasise the ongoing need to use data with care.  Data produced for different purposes must not be used indiscriminately.  Purpose codes that accompany data acquisition must be respected, otherwise misleading or erroneous information will ensue.

The Chartered Institution of Water and Environmental Management (CIWEM) is the leading professional body for the people who plan, protect and care for the environment and its resources, providing educational opportunities, independent information to the public and advice to government. Members in 97 countries include scientists, engineers, ecologists and students.

May 2009

Note: CIWEM Policy Position Statements (PPS) represent the Institution's views on issues at a particular point in time. It is accepted that situations change as research provides new evidence. It should be understood, therefore, that CIWEM PPS's are under constant review, and that previously-held views may alter and lead to revised PPS's.