Ensuring the integrity of data

It is increasingly important to consider and report the quality of process control and environmental measurement data. For example Integrated Pollution Prevention and Control (IPPC) will impose significant monitoring responsibility on operators. The Department for Environment, Food and Rural Affairs’ (DEFRA) Practical Guide to IPPC para 16.2 states: “The conditions should generally require operators not just to provide basic data (for example, the actual results from the monitoring equipment), but also to demonstrate whether they are meeting the conditions of the permit. This may include showing that they are not exceeding ELVs, (emission limit value) that they are monitoring using the required techniques and that they have the necessary management systems in place.”

However environmental measurement is multi-discipline and a multi-department undertaking, often involving several layers of management. It is important, therefore, that there is a common understanding of the requirements which need to be met in order to obtain relevant, consistent and high-quality data. This is the only way to achieve mutual confidence in the data.

The fundamental problem is shown simplistically in Figure 1, and perhaps dramatised by choosing biosolids as the body of material being measured. The outfall in Figure 2 provides a dramatic picture of stratification and a lack of homogeneity. The challenge is less obvious in Figure 3 but it does not take much imagination to realise finding a single representative point for a quality measurement for one of the channels might be difficult.

It is rare for the body of material, whether aqueous, sludge or relatively solid, to be homogenous. As a result, when a sample is taken away for later analysis it might be quite tricky to select a representative few grams from a pile of bio-solids. Users of on-line and

at-line instruments also face frequent difficulties in selecting a single point which is representative of the whole stream or body of material. Even an apparently simple measurement such as the composting temperature of a bio-solids heap is not straightforward since temperature profiles will undoubtedly exist. Further, no matter whether a sample is taken to the laboratory for analysis, or whether an on-line or at-line instrument reports data to a remote display, a chain of events is involved and hence an uncertainty chain must be considered.

Although many interesting and useful standards and documents are available, there appears to be no unifying document to provide advice to a designer of

measurement installations. The Environment Agency specification E32 Automatic Wastewater Sampling Equipment dealt only with the samplers themselves and the system with which to transport the sample from the point of extraction to the storage

chamber – with only an exhortation to make sure the sample is relevant. E32 has now been subsumed into The Environment Agency’s, MCERTs specification of performance standards and conformity

testing procedures.

ISO/TC 147/WG2 N22 Water Quality – On-line Sensors/Analysing Equipment for Water – Specifications and Performance Tests is also an instrument accreditation guide rather than a user’s guide and while this 26-page document goes to considerable detail the part on page 19 dealing with field tests takes less than a quarter of a page and stresses the site-specific nature of such tests, including the sampling considerations within the on-line measuring chain. Other relevant documents include:

• Urban Waste Water Treatment Regulations. Statutory Instrument No.2841, 1994, HMSO,

• ISO 5667-1: 1980, Water Quality – Sampling. Part 1: Guidance on the Design of Sampling Programmes,

• ISO 5667-2: 1991, Water Quality – Sampling. Part 2: Guidance on Sampling Techniques,

• BS EN ISO 5667-3: 1996, Water Quality. Part 3:

  Sampling. Guidance on the Preservation and Handling of Samples,

• ISO 5667-10: 1992. Water Quality – Sampling. Part 6: Section 6.10 Guidance on Sampling of Waste Waters.

It is very unlikely a busy non specialist will trawl these documents to gain information on how to design and install equipment and extraction systems that ensure the sample measurement made is representative and the final data properly represents the process fluid or solid at the time of sampling.

Once the sample has been selected and deemed to be representative, the quality of data arising from the sample needs to be assessed. Most frequently it is assumed United Kingdom Accreditation Service (UKAS) or other accreditation of the laboratory determination, or the declared performance of the on-line instrument, provides the quality assessment of the data.

However, as Figure 1 shows, there is a chain of events from the sample to the availability of the data, and this uncertainty chain determines the quality of the data. The sample might change during the period of storage in the sampler and/or in the transport to the laboratory and the laboratory procedure may remove some parts of the sample in the preparation phase of testing. In the case of on-line measurements it is not unknown for the various signal conversions and transmission to the point of reading to modify or corrupt the data.

In 1997 Bo Neergard Jacobsen, VKI reported on an EC DGX1 contract (B4/3040/93/000924/JS/B1) that showed a need for harmonisation of report formats and comparability of data between member states. There still appears to be much to do. There is currently a great deal of discussion within the EU about the need for properly agreed and validated methodologies for the determination of some organic and most hygienic and health measurements.

In parallel with the need for widely agreed methodologies there is a need for more work on certified reference

materials (CRMs). For many of the measurements that

need to be made CRMs are either unavailable or not available in sufficient quantities to check the calibration of an

on-line instrument.

The ability to provide data with a demonstrable quality and relevance will be of increasing importance in communicating with a wide range of interested people such as managers, regulators, pressure groups and local communities. Although there is increasing pressure to consider the relevance and quality of the data (often incorrectly thought to be information) many of the interested people will be ill informed on the niceties of measurement technology and might be cynical about the providers of the data. However they will be influenced by the logical presentation of the circumstances and details of the measurement chain. British standards and International standards and reference materials

generated by an internationally quality accredited laboratory will clearly assist.

However these take several years to generate and put in place.

A great deal could be achieved in a relatively short time with widely accepted and authoritative best practice guides. For example there is a proposal from within Water-Monet (www.Water-Monet.org), an EPSRC Network run from the University of Cranfield, to generate a best practice guide for sampling for bio-solids and related water quality measurement in collection, distribution, process and waste recycling. The intention is to assemble a very broad range of expert and no-expert inputs, led by a small steering group, to produce a best practice guide that is authoritative, user friendly and widely accepted, for use by experts and non-experts alike. Such a guide would provide a basis for communicating unambiguously with those who have infrequent but important needs to understand what the data is saying