Guidance

Monitoring discharges to water: guidance on selecting a monitoring approach

Formerly part of M18, technical guidance for industrial plant operators (and their contractors) who monitor effluent discharges to water and sewer.

• From: Environment Agency
• Published: 11 June 2020
• Last updated: 18 May 2021 — See all updates

Applies to England

This guidance covers:

• the management and quality assurance of monitoring activities
• the role of the Monitoring Certification Scheme (MCERTS)
• collecting samples and sampling
• laboratory analysis – selecting methods and quality control
• continuous and portable monitors and laboratory equipment – calibration and quality control
• what to look for when auditing (operator internal audits and Environment Agency audits)

Operator self-monitoring

The Environment Agency is responsible for making sure that operator self-monitoring (OSM) is done correctly and to an acceptable standard. For this reason, we encourage operators to use quality management systems.

There are tools to support OSM which give operators and the Environment Agency confidence in the accuracy and reliability of data. These are:

• technical guidance notes – specify the standards and procedures operators should use
• the MCERTS scheme – provides independent quality assurance of monitoring services and equipment and supports EU Directive requirements
• the Operator Monitoring Assessment (OMA) scheme – assesses the management and performance of an operator’s monitoring arrangements to identify any improvements needed

MCERTS and discharges to water

MCERTS covers the:

• product certification of instruments
• competency certification of personnel
• accreditation of laboratories and on-site inspection

This is done according to European and international standards.

MCERTS relevant to monitoring discharges to water

Get more information about MCERTS for monitoring discharges to water, including copies of the performance standards, guidance, and lists of certified instruments.

Performance standard for organisations undertaking sampling and chemical testing of water

This performance standard is for sites that have to send test results to the Environment Agency because they are regulated under either:

• OSM
• the Urban Waste Water Treatment Regulations (SI 94/2841)
• the Industrial Emissions Directive (IED)

The scheme sets out what you must do if you carry out the sampling and chemical testing of:

• untreated sewage
• treated sewage effluents
• trade effluents

Laboratories should be accredited to EN ISO/IEC 17025 for the MCERTS performance standard.

Performance standards and test procedures for continuous water monitoring equipment

This standard is in 3 parts.

Part 1: Performance standards and test procedures for automatic water sampling equipment

Part 2: Performance standards and test procedures for online monitors – covers the monitoring of a number of determinands

Part 3: Performance standards and test procedures for water flow meters

See minimum requirements for the self-monitoring of effluent flow and flow measurement for more details of the overall flow monitoring scheme.

Performance standards and test procedures for portable water monitoring equipment

Performance standards and test procedures for portable water monitoring equipment covers the monitoring of a number of determinands.

Minimum requirements for the self-monitoring of effluent flow

Minimum requirements for the self-monitoring of effluent flow sets out the minimum standards we require from operators that carry out self-monitoring of effluent flow. It also establishes a competency standard for independent MCERTS inspectors who will inspect the operators’ effluent flow monitoring arrangements.

Approaches to monitoring discharges

Monitoring can be classified into the following 2 types:

1. Periodic monitoring

For discharges to water, periodic monitoring is usually removing a discrete sample from the effluent flow and sending it to a laboratory for analysis.

Samples can be single spot samples or composite samples collected over a period of time, for example over 24 hours.

For a number of determinands, you can take portable instrumentation to the discharge site, for example, for pH and dissolved oxygen measurements.

2. Continuous water monitors (CWMs)

These provide continuous automatic measurements, with few if any gaps in the data produced.

The measurement may be taken in the effluent flow, or from a sample withdrawn from the effluent flow to a permanently sited instrument.

The IED specifies continuous monitoring for certain determinands. CWMs are often used to trigger alarms when permit limits are approached so effluent can be diverted automatically to storage before the receiving water becomes polluted.

Defining the substance to be measured

Before carrying out any monitoring it is essential you clearly define both the:

• determinand to be measured in the discharge
• accuracy of result required – see specification of analytical method performance in the MCERTS standard

This will allow you to select the most appropriate analytical system with the performance characteristics needed to meet regulatory requirements. If you use a third party laboratory you should fully discuss the analytical requirements before monitoring starts.

For laboratories accredited to ISO 17025 or MCERTS this will be part of the contract review procedure.

Substances can exist in various forms. A particular analytical method may not respond equally to all forms. You should carefully state the exact form to be determined. Here are some examples.

Dissolved, total or particulate

For metals and nutrients, the dissolved portion of a sample is defined as that which will pass through a 0.45µm membrane filter.

Normally this filtration will take place immediately after sampling. You should fully document the procedure you have used. You can use disposable single use filters.

Total mercury

The method you use should be able to determine organo-mercury compounds as well as inorganic mercury. You may need to demonstrate the effectiveness of sample digestion procedures to break down some compounds.

Phosphorus

Phosphorus can exist in various forms in the water environment, including:

• orthophosphates
• condensed phosphates
• organophosphates

Recognised methods of analysis use an acid medium. This can partially hydrolyse some of the condensed phosphates and break down any labile organic phosphorus compounds present.

It is therefore not possible to specify exactly the form of phosphorus being measured, so we use the term reactive phosphorus. The fractions normally quoted are:

• dissolved reactive phosphorus (sample filtered through 0.45µm membrane)
• total reactive phosphorus (unfiltered sample)
• total phosphorus (unfiltered sample pre-digested)

Phenols

You can identify and determine individual phenols chromatographically. But to comply with regulations you may need to use the colorimetric phenol index method to measure the total phenol content of a test sample.

Many common substituted phenolic compounds (for example, cresols) will be detected by this method, but it is not equally sensitive to all. The system is calibrated using phenol itself, but substituted phenols will be determined as phenol, regardless of their relative response to the method. The phenol index only includes those phenolic compounds it can determine under the specified conditions.

Groups or classes of determinand

Some determinands are grouped into classes such as polycyclic aromatic hydrocarbons and polychlorinated biphenyls. As they consist of numerous compounds, you must specify which individual species you monitor. You should consider factors such as analytical response and harmfulness. Regulation may prescribe which compounds you need to monitor.

Non-specific (empirical) methods

You should carefully define and apply these methods, as the methods themselves define the determinand. An example is biochemical oxygen demand (BOD) where you must define both:

• the length of the test (5 days is normal)
• whether or not allylthiourea (ATU) is to be added to suppress nitrification

Electrical conductivity is another example. You should either make the determination at a specified temperature, or you must temperature compensate the measuring electrode to a standard value, usually 25°C.

Specifying analytical method performance

Analytical results are estimates of the true value or concentration. To make sure that results are fit for their intended purpose, we have set performance targets for analytical accuracy.

For regulatory monitoring the results must show (within acceptable limits of uncertainty) that an operator is meeting the conditions of the permit.

These performance targets are set in terms of both systematic errors (bias, trueness) and random errors (precision). You can find an up to date list in the latest version of the MCERTS standard.

Initial estimates of precision and bias are calculated during method validation studies, see laboratory analysis. To obtain MCERTS accreditation a laboratory must both:

• show evidence of being able to achieve the performance for each determinand during method validation
• maintain the performance during routine operation

The United Kingdom Accreditation Service (UKAS) will check for continued compliance with MCERTS requirements during annual surveillance audits. In-house unaccredited laboratories should also meet these targets – we will assess compliance during OMA audits.

Some trade effluents and discharges to sewer may be more difficult to analyse. This is due to the nature of the matrix, for example very high organic content, or high solids. It may not be possible to attain the targets set in the MCERTS standard. If this situation arises, you should tell us the actual analytical performance that you can obtain. We will then be able to decide if the reported result will allow us to assess permit compliance, or whether you need to develop your analytical method further or change it.

Quality assurance

For quality assurance we would expect you to have the following in place.

Management systems

The operator responsible for self-monitoring must ensure their management system covers all aspects of self-monitoring, including:

• the management of self-monitoring
• the sampling programme design
• sampling procedures – see sampling
• analysis and reporting procedures – see laboratory analysis
• staff training
• the process for auditing and reviewing sampling and analysis operations
• addressing non-conformities

You should fully document the system used, for example in a quality assurance manual. You can use a contractor for some or all aspects of sampling and analysis.

The contractor should be accredited to ISO 17025 for the MCERTS standard, where necessary. You will not need a quality system for those contracted parts but you will remain responsible for the overall quality of the results.

ISO 17025 is an international standard that specifies the general requirements laboratories need to meet to demonstrate their technical competence.

UKAS accredits laboratories to ISO 17025 for specified tests. This provides independent recognition of a laboratory’s competence to perform certain tests or calibrations.

ISO 17025 is also the standard used to accredit sampling activities. This standard is not restricted to laboratories, so organisations that specialise in sampling may also gain accreditation.

ISO 17025 covers management and technical requirements. The management requirements are written to make sure laboratory management systems comply with ISO 9001.

As ISO 17025 only specifies general requirements – further explanations may be needed about the general criteria. MCERTS provides such an application for the sampling and analysis of effluents.

The management of self-monitoring

The quality assurance manual must include a clear quality policy statement, endorsed by a senior executive. This is to demonstrate the operator’s commitment to quality. This policy must encompass all monitoring activities.

You must identify a person with overall responsibility for the self-monitoring quality policy (often called the quality manager). You must also identify the person (or persons) responsible for controlling and implementing the self-monitoring process (technical management). Organisational charts must be available that include defined lines of responsibility.

You should have policies and procedures in place to make sure that you maintain the independence and integrity of your sampling and monitoring and protect them from operational and commercial influences. You can present evidence of this in your organisation’s quality manual.

You must have a clear document control system in place to make sure that you only use the latest versions of documents and procedures. All documents and amendments to documents must be authorised.

The quality manual must contain a procedure to investigate complaints and anomalies regarding the self-monitoring process.

You must keep records that provide an audit trail from defining the sampling programme through to reporting the results.

Sampling programme design

The operator should set out a sampling schedule in advance of an assessment period, at frequencies stated in their permit. The schedule is usually annual, on a calendar year basis, but this may not be suitable for batch processes or plants that operate seasonally.

You should include contingency arrangements for sampling event failures, for example malfunctioning auto samplers.

You may need to agree the sampling schedule with the Environment Agency in advance. If you do not comply with the agreed sampling schedule we may take enforcement action. However, in certain circumstances we will agree to a sampling event being rescheduled. For example, because of adverse weather conditions or plant operational problems.

For some regulatory regimes the operator will need to report any missed samples within 24 hours of the scheduled event.

Sampling frequency will vary depending on the reason for sampling and the nature of the process being monitored.

Staff training

You must show that all staff involved in the self-monitoring process are competent in carrying out their duties. Staff must be trained in the necessary skills and you must keep a detailed training record for each individual as part of the quality system. An appropriate trainer should sign off each documented procedure used by the staff. You should detail a programme of training updates in the management system. Training records can include details of:

• formal qualifications
• internal courses
• instrument manufacturer training
• in-house training

Internal audit and review

The operator must carry out audits of their management system. The audits should cover all aspects of the monitoring process and verify that documented procedures are being followed. The audit programme should be on a yearly cycle. Audits should be carried out by trained staff who are independent of the procedure being audited.

You must record audit findings and any corrective actions that are needed. You must carry out follow up audits to make sure that any corrective actions are effective.

You should carry out a management review of your management system at least once a year. This should:

• review the results of internal audits and assessments by external bodies
• review actions taken to correct non-conformances
• consider feedback or complaints from the Environment Agency and others

You should record details of these management reviews along with any corrective actions taken.

You should assess laboratory performance through inter-laboratory proficiency tests and internal analytical quality control (AQC). You do not need to do this if you use a third party accredited laboratory. However, we may ask for information about a laboratory’s performance in analytical quality control and its participation in proficiency testing schemes.

OMA

OMA is the audit of operators by the Environment Agency. We carry out a general assessment of an operator’s monitoring performance using the OMA scheme. This may include vertical audits of specific samples.

We developed this auditing scheme to assess an operator’s self-monitoring arrangements and identify if any improvements are needed. OMA covers:

• the management, training and competence of personnel
• periodic and laboratory monitoring
• continuous monitoring
• quality assurance of monitoring

See guidance on what to expect in an audit and how to carry one out.

Sampling

This section provides guidance on the sampling requirements we would expect an operator to meet.

Choosing your sampling point

You should agree the sampling point location with the Environment Agency. Document it and clearly and permanently label it. The following guidance applies to both manual and automatic sampling techniques.

You must locate discharge sample points at places where you can take a sample that is truly representative of the discharge. A sampling position in a pipe or channel must be far enough downstream of the last inflow so that mixing of the two streams is complete.

You should take samples at an outfall from regions of high turbulence and good mixing, usually at the centre of the discharge. Solid materials will have little chance to settle out here.

When collecting samples in channels you should collect the sample away from the sides and bottom of the channel to avoid contaminating the sample with sediment and biological growths.

When sampling from chambers (for example manholes) avoid contaminating the sample with deposits from the:

• cover – these could be disturbed when the cover is lifted
• chamber walls
• bottom of the chamber

You can also draw off samples from effluent streams at a tap. You should make sure you flush out any dead space with effluent before you collect the sample.

Sampling staff should be aware that manholes and similar confined spaces are dangerous. They must not enter them unless they follow a safe system of work and only after they have received appropriate training.

You must make sure that the sample probe remains in the effluent flow for the entire time each sample aliquot is being taken. Variations in effluent flow should not result in the sample probe being left dangling in the air or in contact with the bottom of the channel. Some operators may discharge effluent in batches from hold-up tanks, for example, if they are only allowed to discharge at high tides. If this applies you should mix the tanks wherever possible, and in some cases, take samples from recirculation lines. If good mixing is not possible, you may need to increase the frequency of sampling during discharge.

Composite sampling

Two types of composite sample are commonly used:

• flow-proportional
• time-proportional

For a flow-proportional sample, there are 2 approaches, a:

• fixed amount of sample, taken for each pre-defined volume of effluent (for example, every 10m³)

• variable amount of sample volume in proportion to the flow, taken at constant time intervals, for example once per hour

For time-proportional samples, a fixed amount of sample is taken from the effluent for each time unit (for example every hour).

The analysis of a composite sample gives an average value of the determinand during the period over which the sample has been collected. It is normal to collect composite samples over 24 hours to give a daily mean value. You can use shorter times if you get prior agreement from the Environment Agency.

You should use flow proportional sampling when the volume of effluent discharged varies significantly throughout the sampling period. When the volume of discharge is relatively constant then time proportional composites are appropriate.

We recommend that you use automatic sampling equipment, see automatic sampling equipment and MCERTS. However, you must consider whether the target substances will remain stable over the total sample collection time as samples may deteriorate while sitting in the automatic sampling device.

Spot sampling

These are discrete samples taken from a discharge at random time intervals and are not related to the volume of discharge. They are best suited where:

• the composition of the waste water is relatively constant
• the discharge contains mineral oil or volatile substances
• the target substances are not stable in the sample due to decomposition, evaporation or coagulation
• separate phases are present (for example an oil layer floating on water)
• you need to check the quality of the discharge at a particular moment, normally to assess compliance with permit conditions
• the discharge is not continuous (from batch or hold-up tanks) when the effluent is well mixed
• you are collecting larger object and floating matter that is not representative of the discharge

Bulked composites

When you are preparing a bulk sample by manually compositing a series of samples in a laboratory, you must consider the stability of the determinand being measured. For example, BOD will start to deteriorate significantly after 24 hours.

It may not be appropriate to collect composite samples for periods over 24 hours due to the stability of some determinands, even when auto samplers are refrigerated. For example, BOD, pH, chemical oxygen demand (COD), volatile organic compounds and ammonia.

Automatic sampling equipment and MCERTS

Automatic sampling devices used for self-monitoring must be tested and certified to the MCERTS performance standard: Performance standards and test procedures for continuous water monitoring equipment – part 1 automatic sampling equipment.

Find:

• more information about the MCERTS scheme
• a list of certified equipment

The MCERTS standard only covers sampling from non-pressurised channels and vessels. You may install break tanks or other such devices to allow monitoring to take place.

Access, facilities and services

The access, facilities and services you will need for sampling will vary depending on your approach to monitoring, and the equipment you use. However all will need:

• a safe means of access to the sampling position
• a safe place of work at the sampling position
• shelter and weatherproofing of equipment
• space for the equipment and personnel
• essential services such as electricity and lighting

Sample bottles, storage and transportation

If you need to refrigerate your samples to preserve them, you should store them at between 1°C and 8°C during transportation to the laboratory and whilst they are retained in an automatic sampling device. The average temperature over the period should be no more than 5°C. You can use ice packs, refrigerators or other appropriate methods to maintain the correct temperature. You should have appropriate procedures demonstrating you are doing this, for example using temperature loggers.

We recognise that you may need some time to bring the sample temperature to within this range during high ambient temperatures. The temperature range is to allow for the:

• cycling of the refrigeration devices
• opening and closing of refrigeration devices during normal operation
• effects of adding a number of warm samples

You should transport samples in sealed containers, which you should regularly clean and disinfect.

Examine the samples as soon as possible after collection – if possible within 24 hours of sample collection. Where logistics do not allow this, you may examine samples up to 48 hours after collection, provided they are kept:

• cool (between 1°C and 8°C, average not above 5°C)
• in the dark

If you store samples at a laboratory, you may need to use method-specific storage requirements. For example, laboratory storage temperatures may be 1 to 5°C. For most analytical purposes, best practice is to keep the samples at a constant temperature of not more than 5°C.

Sample containers should be appropriate for the determinand and analytical system used. Wherever possible they should be supplied by the analysing laboratory. Sample containers should not be rinsed with the sample before being filled, unless specified in the sampling procedures.

It is very important to note any laboratory requirements about filling sample containers. For example, some tests require no air space be left after filling, to stop the loss of volatile components. Others need space left in the bottle to allow extraction solvents to be added when reaching the laboratory.

If you do not follow laboratory instructions on using sample bottles, it may lead to invalid analytical results. Laboratories accredited to ISO 17025 for an MCERTS standard should reject improperly presented samples. If the customer insists on analysing the samples then a disclaimer should accompany the results stating they may be invalid.

Where appropriate, add preservatives to make sure that there is no material change in the concentration of the determinands before analysis.

Preservatives are often added to sample containers before they are dispatched from the laboratory, and these containers should not be rinsed. Details of best practice in storage and preservation are given in EN ISO 5667-3.

Sampling procedures

Operators should fully document their sampling procedures. This should include details of the:

• precise location of the discharge, spot sampling point and automatic sampler installation (if appropriate)
• sampling process
• storage conditions and transport of samples
• types of bottles or containers and their closures
• cleaning procedure for each type of bottle, container and closure
• sample preservation measures (if used)
• calibration and maintenance of automatic samplers, timers and thermometers
• records such as training records of those who take the samples, and the automatic sampler installation and testing record
• actions to be taken in the event of automatic sampler failure
• procedure for notifying a sampling event failure to the Environment Agency
• quality assurance procedures for sampling activities

These procedures should be part of the management system and be made available to all staff that carry out sampling. If requested, submit them to the Environment Agency for approval.

The procedures should also include examples of:

• the sampling event record sheets to be presented to the analysing laboratory
• sample results forms returned from the laboratory plus the format of any computer generated data
• sample event failure reports

Having a well-controlled sampling methodology is essential to make sure you minimise the contribution of sampling to the overall uncertainty of measurement.

Sampling quality control

The sampling process can have a substantial effect on the overall uncertainty of a measurement result. You can verify the performance of field procedures for each batch of samples taken. You should take and analyse control samples with the batch of samples they are associated with.

To monitor the variation of sampling control samples, you should record or plot results on control charts, see quality control and preparing and interpreting simple AQC charts. You can use the data collected as part of an estimate of sampling uncertainty.

The following types of control sample may be suitable.

Blank samples

These can be used to assess levels of contamination in the sampling process, and investigate contamination of different areas of the process, such as filtration or bottles.

Spiked samples

This is where you add a known quantity of a determinand standard solution to a sample. It can be used to assess different areas of the sampling process, such as sample bottles, pre-treatment (filtration, preservation) or transport. Standards used for spiking the sample should be from a different source or lot number to those used for calibration.

Duplicate samples

This is where you repeat (as far as is practical) the entire sampling procedure and submit 2 separate samples to the laboratory for analysis.

Reference materials or samples

These can be used to carry out quality assurance checks on field instruments used for on-site tests.

Laboratory analysis

This section outlines the general requirements your chosen laboratory should meet.

Choosing a method

Standard methods for laboratory analysis are developed by various national and international organisations. Regulation requires you to use the following hierarchy of measurement methods.

1. Standard methods required by relevant EU directives.
2. CEN standards for the relevant determinand.
3. ISO standards.

You can choose the following alternative methods with prior approval from the Environment Agency (we may impose extra requirements):

• SCA blue books
• in-house methods
• modified methods
• test kits

The degree of validation detailed in standard methods is variable, especially for the matrix the method is tested in. Therefore it is very important that you evaluate the measuring method to check that it is fit for purpose. You must also check that the laboratory using the method is able to verify any performance criteria stated in the appropriate standard, or specified by the Environment Agency, for example in an MCERTS standard.

If the sample you will analyse is of a matrix that has not been tested through suitable validation tests, then you will need to carry out further tests to ascertain the suitability of the method. We may require proof of method suitability and details of how matrix problems are addressed.

The index of CEN and ISO monitoring methods has a list of CEN and ISO analytical methods. A list of other methods is given in index of some alternative monitoring methods. These lists may not be mandatory and are not exhaustive. Other appropriate methods may be acceptable with satisfactory validation.

You can get CEN, ISO and British Standards from the British Standards Institute and SCA blue books from the Standing Committee of Analysts.

Laboratory equipment

Any equipment used in the analytical systems must be shown to be fit for purpose. You must:

• fully document instrument operation instructions, calibration procedures and performance checks and make them available to users as part of the management system
• carry out instrument performance checks and calibration procedures at appropriate intervals and keep a record showing that you have maintained calibration
• correctly maintain all instruments and keep records of this maintenance, even if it is carried out by a third party, such as the instrument manufacturer
• demonstrate that the calibration of equipment such as balances, thermometers, timers, auto-pipettes, ovens and heating blocks can be traced to national measurement standards, and you must make available any corresponding certificates or other records
• clearly label calibrated equipment so the user can see it is calibrated and within date

Test kits

The majority of appropriate test kits involve colorimetric methods. They come in 2 main formats – those using visual comparators and those using portable or bench top spectrometers.

Generally, using visual comparators is not acceptable, as these systems are very dependent on the operator and environmental conditions. They often lack the accuracy required for assessing regulatory permits. If you use these, you must demonstrate that they are fit for purpose.

Test kit methods using spectrometers have improved, and many are based on standard laboratory methods. You still need to achieve traceability of the results.

Test kits offer some advantages, such as:

• reagents being pre-packaged
• ease and convenience of use
• built in calibration routines

However, you should carry out a full evaluation before using them to make sure they have the appropriate performance characteristics and matrix suitability. You should treat them in the same way as a standard method in terms of documentation and quality assurance or quality control procedures.

Test kits are used for many determinands, including COD, ammonia, phosphate and iron. You can find comprehensive guidance on using test kits in BS 1427: Guide to on-site test methods for the analysis of waters.

Validation

You must support analytical methods by assessing the following performance criteria in appropriate matrices to demonstrate that it is fit for purpose:

• selectivity and interference effects
• range of applicability
• linearity
• calibration and traceability
• bias (recovery)
• precision (repeatability, intermediate reproducibility)
• limit of detection
• uncertainty estimates

You should carry out these performance tests before you put a method into routine use, but only when the analytical system has been optimised. If an analytical system is modified, you may need to revalidate, for example when a piece of equipment is replaced.

You can find a typical protocol for validation in MCERTS: performance standard for organisations undertaking sampling and chemical testing of water.

Methods and procedures

You must fully document the measuring method. This must include details of the:

• scope and performance characteristics and an estimate of uncertainty
• principles
• hazards and disposal of waste materials
• reagents and standards
• equipment
• sample collection, preservation and preparation procedures
• calibration procedures
• quality control procedures
• calculation and reporting procedures
• references

Find more detailed guidance in the following documents:

• Fitness for purpose of analytical methods: a laboratory guide to method validation and related topics

• ISO 78-2:1999. Layout for Standards – part 2: Methods of chemical analysis

Quality control

Laboratories should have established and fully documented AQC procedures that are part of their quality management system. These procedures must provide a continuing check on the day to day performance of analytical systems. The laboratory must subscribe to an external proficiency scheme, where available and appropriate.

Internal quality control procedures ensure the quality of specific samples or batches of samples. These may include:

• analysing reference materials and measurement standards
• analysing blind samples
• using quality control samples and control charts
• analysing blanks
• analysing spiked samples
• analysis in replicate
• system suitability checks

As a minimum, using Shewhart control charts is recommended for routine analysis of effluent and water samples. These can easily identify when an analytical system is out of statistical control.

A method is in statistical control when the variability within the analytical system comes from a stable set of sources of random analytical variability. That is, the precision of the method falls within expected limits.

You can find these limits in MCERTS: performance standard for organisations undertaking sampling and chemical testing of water. These are the limits MCERTS accredited laboratories must achieve – they show what is achievable by a well-run laboratory.

The causes of variation are equally likely to result in analytical errors in a positive or negative direction and will affect all measurements. Loss of statistical control is characterised by either:

• introducing sources of systematic error (bias, trueness)
• a change in the size of the random error (precision) operating in the analytical method

Laboratories must have fully documented procedures for actions they will take when a system is identified as out of control. They should keep records of breaches, and also of remedial measures taken. They should also have a procedure for re-evaluating and updating the control limits.

An example detailed procedure can be found in the guidance preparing and interpreting simple AQC charts. You can find more information in the following documents:

• NS30 – A manual on analytical quality control for the water industry, R.V.Cheeseman and A.L.Wilson revised M.J.Gardner, June 1989, ISBN 0902 156853
• ISO TS 13530:2009 Water quality – guidance on analytical quality control for chemical and physicochemical water analysis
• Quality control charts in routine analysis, WRc report CO4239, M J Gardner, 1996
• BS ISO 7870-1: control charts – general guide and introduction
• BS ISO 7870-2: control charts – Shewhart control charts

Proficiency testing (PT) schemes are inter-laboratory comparisons of analytical performance. They consist of regularly distributing homogeneous samples to a number of participating laboratories, by an independent organisation accredited to EN ISO 17043.

The concentration of target determinands in the samples should reflect limits set in permits and consents. Participating laboratories should not be told what these are.

The matrix of test samples should be as near as possible to that analysed by participating laboratories. The results of the analysis are statistically evaluated – they give an assessment of the analytical performance of participants. Laboratories should have a fully documented PT procedure, including methods for investigating and recording the actions taken when poor PT performance occurs.

See examples of PT schemes which comply with EN ISO 17043:

• LGC website (LGC AXIO, formerly Aquacheck)
• EPTIS website

Measurement uncertainty

Measurement uncertainty is the range of values in which you will find the true value of an analytical result, with a specified level of confidence.

Every measurement has an uncertainty associated with it. This is because of errors in sampling and analysis and from imperfect knowledge of the factors affecting the result. For measurements to be fit for purpose you need some knowledge of these errors. The Environment Agency may request a statement of the uncertainty associated with a reported result.

You can use 2 general approaches to estimate measurement uncertainty.

Approach 1

In the first approach you must identify and list all individual sources of uncertainty. Sources include:

• random and systematic errors
• volumetric equipment
• balances and weights
• calibration
• sample pre-treatment
• temperature effects
• interferences

Estimate each independent contribution. Some data will be already available, some may be available from literature (certificates, equipment specifications) and some may require further experimental studies.

This approach may lead to:

• underestimating measurement uncertainty as it is difficult to assign all causes
• overestimating measurement uncertainty as it is difficult to be sure all identified contributions of uncertainty are fully independent

Approach 2

The second approach is more appropriate for routine analysis of discharges. In this approach, you estimate measurement uncertainty using an overall estimate of precision which you get from validation studies and long term AQC data (intermediate reproducibility).

Consider all possible sources of uncertainty. You may need to incorporate a contribution to uncertainty from:

• bias (recovery) measurements
• sample homogeneity
• matrix
• concentration

If precision and bias studies identify a source of uncertainty and find it accounted for, you may not need to carry out a further evaluation of uncertainty. For example, if data is drawn from a whole year then the variations in the laboratory’s environmental temperature will be adequately represented. Or, if a variety of different volumetric apparatus has been used their effects on calibration will be taken into account.

Bias (recovery) is usually studied by analysing certified reference materials or spiked samples during method validation or longer-term evaluation, or both. You can also use PT results. You should make every effort to eliminate or reduce the bias effect.

You only need to include bias in estimates of uncertainty if it is considered to be significant. If an analytical procedure is considered empirical, then you only need to evaluate bias for laboratory performance, and not for the method. This is because the result is defined by the method applied, and depends solely on it.

Once you have identified and estimated all the important sources of uncertainty you should convert them to standard uncertainties, which are expressed as standard deviation. If it is based on single measurements then intermediate reproducibility data is usually already a standard deviation. But if it is based on replicate determinations then you should calculate standard deviation of the mean.

You should then use individual standard uncertainties to calculate the combined standard uncertainty using an appropriate method. You should express uncertainty as an expanded uncertainty, by multiplying the combined uncertainty by a coverage factor (k), which is derived from student t values. This gives an appropriate level of confidence to the uncertainty estimation. The value of k should be 2, giving a 95% confidence in most cases. This will not be true when the combined degrees of freedom of the estimate is small, but this situation should not arise.

You should report results in the form R ± U, where R is the result and U is the expanded uncertainty. If k has a different value than 2 then you need to state this with the result.

See further details and worked examples:

• In-house method validation – a guide for chemical laboratories, LGC, ISBN 094892618X
• M H Ramsey, S L R Ellison and P Rostron (eds.), Eurachem/EUROLAB/ CITAC/Nordtest/AMC, Guide: measurement uncertainty arising from sampling: a guide to methods and approaches. Second edition, Eurachem (2019). ISBN (978-0-948926-35-8) – available from Eurachem
• EURACHEM/CITAC Guide to quantifying uncertainty in analytical measurement EURACHEM / CITAC Guide CG 4 – available from Eurachem
• Evaluation of measurement data – Guide to the expression of uncertainty in measurement, JCGM 100:2008, GUM 1995 with minor corrections
• Handbook for calculation of measurement uncertainty in environmental laboratories, Nordtest report TR 537 – available from Nordtest

Collecting and reporting data

The routine test report must contain the following information:

• the name and address of the laboratory where analysis took place
• a reference to the method or standard used
• any deviations from the standard used, or options employed
• a full identification of the sample, including date and time taken, date and time received
• the results of the determinations and expanded uncertainties if requested
• any factors which may have affected the results including recovery factors

Continuous water monitors

Many of the topics covered in the sections on sampling and laboratory analysis apply to CWMs.

Location of the sensor

You can locate the sensor in the effluent flow, or remote from the effluent flow with the sample pumped to it. Read the guidance on choosing your sampling point before you position the sensor or the sample intake.

The guidance in access, facilities and services also applies. You should also clean sampling pipes and tubes where necessary.

Type of CWM

The method used by the CWM must be appropriate for the exact determinand specified in the permit. It must be measured using an appropriate technique. Where available and suitable you should use instruments with MCERTS certification. Here are some examples of analytical techniques that can be used by CWMs:

• direct electrochemical, for example pH, dissolved oxygen, conductivity
• specific ion electrodes, for example nitrate and ammonia
• anodic stripping voltammetry for metals
• colorimetric (spectrometry), for example ammonia, phosphate, total phosphorus, iron
• total organic carbon analysis
• turbidity analysis

Calibration and maintenance

Regular calibration and maintenance of CWMs is vital to make sure they produce quality monitoring data with minimum breakdowns. CWMs should be installed, commissioned and validated by their manufacturers or manufacturers’ agents.

You should document maintenance procedures. Maintenance should be carried out following the manufacturer’s instructions and at recommended frequencies. Major servicing should be carried out by manufacturers or specialist companies. Suitably trained staff can undertake interim (daily or weekly) calibration and maintenance checks.

You should follow a written schedule of maintenance and calibration tasks and keep records of all maintenance and calibration activities.

You should hold on site an appropriate quantity of spare parts and consumables to make sure the CWM is in continuous operation. A call out contract is recommended for emergency repairs.

Flow measurement

The uncertainties associated with flow measurement can have a significant effect on the calculation of emission loads. Small fluctuations in flow measurements can lead to large differences in load calculations.

Two MCERTS standards cover the inspection of effluent flow monitoring arrangements including the monitoring installations and the associated quality assurance systems.

Minimum requirements for the self-monitoring of effluent flow

The standard for minimum requirements for the self-monitoring of flow is for permit holders who discharge sewage or trade effluent (or both) to controlled waters or public sewers. The standard covers performance requirements for measurement and management systems. It also includes the relevant quality systems and the collection and reporting of monitoring data.

Competency Standard for MCERTS inspectors – effluent flow monitoring

The competency standard for inspectors is for independent technical specialists who will undertake the assessment process on behalf of the consent holder. The scheme operates as follows.

MCERTS inspectors are appointed by the Sira Certification Service. They operate on our behalf and are part of the CSA Group. The scheme is delivered through a number of companies operating in a commercially competitive market. Operators place a contract with one of the companies employing MCERTS inspectors. Find details of companies employing MCERTS inspectors.

MCERTS has set a total daily volume target of better than +/- 8% uncertainty for effluent flow monitoring systems. MCERTS inspectors will check this during their inspection. This uncertainty target covers the whole monitoring system, not just the performance of the flow meter. For example, deviations in the construction of V-notch weirs and other installation issues.

Following the inspection, the MCERTS inspector prepares a report, including a recommendation of whether the flow monitoring arrangements meet the MCERTS requirements. This includes an assessment of the flow application, type of flow measurement device and maintenance.

The QMS (quality management system) for flow monitoring also needs to be assessed. The assessment is performed by a UKAS accredited certification body that has MCERTS for flow included in its scope. This can be the Sira Certification Service or an operator’s existing ISO 9001 or 14001 auditor if they have appropriate accreditation.

The Sira Certification Service will then check the MCERTS inspector’s report and the QMS auditor’s report. If the MCERTS requirements are being met they will issue an MCERTS Site Conformity Inspection Certificate, valid for 5 years.

The QMS auditor must also carry out an annual QMS surveillance visit. The frequency may be reduced if the auditor is satisfied that the QMS is able to guarantee performance.

Operators must also use MCERTS certified flow meters. Meters that comply with this performance standard can produce results of the required quality and reliability when operated within the MCERTS flow scheme.

Find detailed guidance on what is expected in the MCERTS standard minimum requirements for the self-monitoring of effluent flow.

Find additional guidance in MCERTS bulletins at the CSA Group website.

Calculating results as averages and specific load

You may need to report results to us as an average or specific load.

Averages from continuous and periodic measurements

For continuous monitoring, you may need to calculate averages at different time periods, for example 1 hour, 2 hours or 24 hours, depending on sector requirements. You should use the same time period for averaging the results of the flow measurements.

If you collect flow-proportional composite samples for periodic monitoring, the results are already representative averages, as the sampling method already considers the flow. This is the preferred sampling method for estimating average and load.

Time-proportional composite samples are not suitable if you collect a single composite sample for the period. This is because there will be no relation to the flow, unless the flow is constant.

If you collect discrete samples (say hourly samples over 24 hours) you can make additional calculations to relate to the flow at the time of the sample. By doing so, it becomes representative of the discharge of a pollutant for a specific time period. You must calculate the total discharge volume for each time interval.

To calculate a representative concentration over a longer time period, you need to average and weight the individual measurement results by the related waste water flow. See equation 1 in calculations used in monitoring discharges to water (ODT, 32.5 KB).

To calculate a yearly average concentration based on 24 hour flow-proportional composite samples, do the following:

1. Take the measurement result of each 24 hour flow-proportional composite sample obtained during the observed year.
2. Multiply them with the corresponding daily average flow.
3. Sum these up and then divide by the sum of all the daily average flows.

Specific load

You can calculate a monthly or yearly average load based on a daily measurement frequency using equation 2 in calculations used in monitoring discharges to water (ODT, 32.5 KB).

If a limit is given as a specific yearly load, based on 24 hour flow-proportional composite sampling, do the following to calculate the specific daily load:

1. Take the measurement result of each 24 hour flow-proportional composite sample obtained during the year.
2. Multiply them with the corresponding daily flow.
3. Divide them by the daily production output.

The specific daily loads are summed up and divided by the number of measurements to calculate the specific load for the year. You can take the same approach for monthly loads.

You can calculate specific loads based on a measurement frequency of less than once a day in a similar way, but you need to make sure that the measurement results are representative for the examined time period. However, you should report the average as an average over the samples obtained during the period, to avoid any confusion.

For some example calculations see appendix 6 of the JRC reference report on Monitoring of emissions to air and water from IED installations 2018.

Definition of terms

This section explains how the Environment Agency defines the terms used in this guidance. We have provided these explanations for clarity, as some of these terms can be interpreted in different ways.

Bias and recovery

Bias is the systematic error of an analytical system. It can be expressed as the difference between the mean of a significant number of determinations and the true or accepted value. You can use certified reference materials to estimate bias.

Sources of bias include:

• sample instability
• interference and matrix effects
• calibration
• blanks
• inability to recover all forms of the defined determinand

You can carry out recovery tests using real samples to estimate bias from some sources. You do this by:

1. Spiking a sample with a known amount of a determinand.
2. Analysing portions of the sample and the spiked sample a number of times (11 batches of 2 duplicates is typically used, which will guarantee 10 degrees of freedom).
3. Calculating the percentage of spiked determinand recovered - see equation 3 in calculations used in monitoring discharges to water (ODT, 32.5 KB).

Recovery or bias for general and metals determinations is usually acceptable where the true mean recovery lies between 95% and 105% with 95% confidence, or 90% to 110% for organic determinands.

Outside of this, it may be necessary to correct results, for example with trace organic analysis. You should report the correction factors applied with the results.

Calibration and traceability

Calibration is the process that relates the output from an analytical system to the concentration of the substance being measured.

This is usually done by analysing a series of samples of known concentration that are prepared from, or relate to, the substance being measured.

You can then relate the output from the analytical system to the concentration of the substance being measured, for example by using a calibration curve.

Most laboratory analytical procedures rely at some stage on measuring properties such as:

• weight
• volume
• temperature
• time

For example:

• volumetric flasks
• analytical balances that are used to weigh out materials to prepare calibration standards
• thermometers

You must calibrate these devices and document all calibrations. Calibrations must be traceable to national or international reference standards through an unbroken chain of comparisons with known uncertainties.

Linearity

An analytical system should (within its working range) provide a test result that is directly proportional to the concentration of the determinand being measured.

You can check this by following these steps:

1. Measure a blank and a range of standards (minimum of 6) spread evenly across the calibration range.
2. Plot a calibration curve to show outliers and general shape.
3. Use the calibration curve to calculate regression coefficients.

If the response is non-linear then it may be possible to employ a suitable non-linear calibration function.

Limit of detection (LOD)

An important aspect of monitoring is the LOD of an analytical system. The uncertainty associated with a measurement increases the closer the result is to the LOD.

For good analytical practice the LOD of a method should not exceed 10% of the concentration of interest. This is usually the emission limit defined in the permit.

This should not be confused with the working range of the method, as a LOD can often be below 1% of the analytical system’s range of applicability.

Estimating the LOD required provides a guide for selecting an appropriate method – it helps minimise the uncertainty associated with a measurement result that is close to the emission limit.

There are several methods for calculating the LOD. The most appropriate for discharges is a statistically based approach.

LOD can be defined as the concentration at which there will be a 95% probability that you will detect the determinand. This means that the probability of failing to detect it is very small.

Precision

Precision is the distribution of a number of repeated determinations, expressed as the standard deviation of results. It estimates errors of a random nature.

Total standard deviation is calculated from a combination of within analytical batch, and between analytical batch standard deviations. It is measured under a number of conditions.

Under repeatability conditions, you analyse a sample by the same method, equipment, laboratory and analyst within a short time interval.

This precision data can be produced during method validation studies. It can be used to derive internal quality control charts.

Typically, you use 11 batches of 2 duplicates, to guarantee 10 degrees of freedom.

An in-house or intermediate reproducibility condition is where the method has been put into routine use by a variety of analysts using different equipment, over a longer time. This would reflect variations caused by environmental conditions (for example laboratory temperature).

Precision may vary across the concentration range. You should test it at a minimum of 2 different concentrations. One of these should be at the level of interest, such as the permit limit. Laboratories often use 20% and 80% of the highest concentration determined by the method.

Range of applicability (working range)

This is the range of concentrations at which the method can provide analytical results that meet accuracy and precision requirements.

Selectivity and interference effects (matrix effects)

A method must not only measure the determinand specified, but the measurement must not be affected by the presence of other chemical species in the sample.

You can investigate this by analysing standards with a range of potential interferences added at varying concentrations. These should include the highest possible concentration that may be found in the sample. Because the extent of interference may depend on determinand concentration, you should study at least 2 different concentrations of the determinand.

Published 11 June 2020
Last updated 18 May 2021 + show all updates

1. 18 May 2021

   Replaced images of equations with a ODT attachment that gives the following calculations as text: 1. representative concentration over a longer time period, 2. monthly or yearly average load based on a daily measurement frequency, 3. percentage of spiked determinand recovered.

2. 11 June 2020

   First published.