In 2021, England created the Environment Act (EA 21), a landmark piece of legislation aimed at improving environmental protection. Section 82 (S82) of EA 21 aims at introducing rigorous regulations to preserve water quality, safeguard aquatic ecosystems and lessen the risk to public health. Section 82 outlines mandatory monitoring criteria for the water bodies which receive either fully treated, partially treated or untreated wastewater, from sewerage and water treatment undertakers. It mandates regulatory agencies to actively monitor and manage pollution sources, ensuring that all effluent discharges comply with established contaminant permits, as set by the Environment Agency. It also affords the chance to understand more about the impact of these discharges on the wider river environment, and in the context of other pollution sources.
1. Environment Act
In 2021, England created the Environment Act (EA 21), a landmark piece of legislation aimed at improving environmental protection. Section 82 (S82) of EA 21 aims at introducing rigorous regulations to preserve water quality, safeguard aquatic ecosystems and lessen the risk to public health. Section 82 outlines mandatory monitoring criteria for the water bodies which receive either fully treated, partially treated or untreated wastewater, from sewerage and water treatment undertakers. It mandates regulatory agencies to actively monitor and manage pollution sources, ensuring that all effluent discharges comply with established contaminant permits, as set by the Environment Agency.
It also affords the chance to understand more about the impact of these discharges on the wider river environment, and in the context of other pollution sources. The key water quality parameters addressed in S82 include dissolved oxygen, pH, temperature, turbidity, ammonia and, more recently, electrical conductivity. The Secretary of State has also stated that ‘systems must be able to be upgraded to accommodate at least two additional parameters without requiring redesign or reinstallation’ a, this not only future-proofs instruments but allows a level of flexibility in monitoring capacity.
To maintain compliance, water companies must continuously monitor both upstream and downstream of all Combined Sewer Overflows (CSOs) and Wastewater Treatment Works (WWTWs) discharge points. The parameters set forth are descriptive measures of water quality however, in terms of determining the ‘polluting ability’, the burden of responsibility clearly falls on one particular parameter: Ammonia. All the other parameters reflect the general health of the receiving water body but do not specifically quantify the extent of any pollution event.
2. The Elephant in the room: Ammonium ISEs
Ammonium is a critical parameter to monitor as it can indicate industrial, agricultural, and wastewater runoff. Elevated concentrations can promote algal blooms and are toxic to aquatic organisms. However, monitoring ammonia in water presents several challenges. In environmental water, ammonia can take two forms, gaseous NH3 and/or as aqueous NH4; the concentration of each determined by the water pH (Figure 1). As it is complex to measure ammonia directly in water, the most common method is to use an Ion Selective Electrode (ISE) to measure ammonium directly. Ammonia is then calculated by utilising the ISE with simultaneous pH, conductivity, and temperature readings
Figure 1: Graph showing the concentration balance between Ammonium (NH4) and Ammonia (NH3) with changing pH.
Traditional ISEs utilise a silver/silver chloride electrode which are mounted behind a selective membrane and within a custom reference fluid; the selective membrane allows ammonium ions to pass through. The ammonium ISE cannot be used in saline waters typically above 2000µS/cm due to interferences from other ionic species. Most ammonium ISEs, including that of Proteus, will utilise pH, temperature, and conductivity to provide an accurate measurement of ammonia.
Traditionally, ammonium ISEs have been touted as short-term, discrete measurement sensors, a fact repeated throughout manufacturer literature, manuals, and product specifications. S82 has provided a new challenge, demanding continuous monitoring from a sensor which was never designed for long-term, continuous deployment. A somewhat concerning trend has seen retailers of these technologies, reframe or restrict to small print the legitimate shortcomings of their use in a continuous monitoring scenario. ISEs undoubtedly have their place in water quality monitoring but it is not in a continuous, long-term format. The determination to overlook the genuine shortcomings of ISE technology in order to meet ‘on paper’ requirements of the S82 legislation is concerning. Attempting to use ISE technology within sondes for large scale continuous monitoring could be nothing short of disastrous with consequences that will go far beyond each individual monitoring project.
As a manufacturer that uses ISEs, we are aware of the simple truths surrounding their use that everyone should know before using them:
- Accuracy Limitations: ISEs typically offer an accuracy of ±2 mg/L, which may not be sufficient for detecting low-level pollution events. For a typical English river, the background level of ammonium will be <0.5 mg/L; then add to that a river discharge of 2500 L/sec and a CSO with 10 mg/L ammonium which discharges into the river at 10 L/sec. In this broad scenario, the increase in ammonium to the river would be 0.04 mg/L, which would be undetectable based on the degree of accuracy stated by various ISE manufacturers. Table 1 demonstrates the range of scenarios within which an ISE would be a useful tool (green) compared to the actual likely scenarios where background levels of ammonium in rivers typically reside (red); the bottom line is, the real-life scenarios where ISEs can operate accurately are very rare
Table 1: A dilution-scenario matrix defining the potential levels of ammonia based on the discharge from both the CSO and receiving water course. The colour coding depicts whether the resulting concentration of ammonia/um can be detected by an Ion Selective Electrode (ISE). Additionally, these data are unable to account for the complication of ISE error, defined as +/- 2 mg/L or +/- 10%.
- Sensor Drift: ISEs are prone to drift (Blaen et al, 2016; Castrillo and Garcia, 2020), necessitating frequent recalibration to maintain data integrity. With a single ISE on an instrument there is no way of determining when the ISE starts to drift; not unless a redundant sensor is used in tandem. Neither S82 nor other manufacturers have stipulated how or if this should be addressed. Blanket recalibration periods are problematic as both the timing and the degree of drift in these sensors can change, depending on a range of environmental and physical factors. It also cannot be assumed that the rate and magnitude of drift will remain the same for a particular installation. For the Proteus multiprobe, we already employ tandem monitoring to overcome the drift issue by utilising our optical ammonium measurement or using a secondary ISE sensor.
- Maintenance Demands: regular maintenance, including monthly calibrations and bi-annual membrane replacements, are required, leading to high operational expenditures (OPEX). Some manufacturers even recommend replacing membranes every few weeks! The cost burden for this aspect is immense; the following article explores some of the cost savings associated with optical ammonium measurement either as the primary form of measurement or as the secondary form. [Article Link]
- Cleaning: ISE membranes are very sensitive, typically only allowing submersion <15m depth. In line with the sensitivity, some manufacturers state outright that it should not be wiped with an integral wiper at any point during deployment. The chief concern being that wiping with a brush may risk forcing material into the membrane, causing the fouling it is attempting to mitigate. This accidental fouling can be observed in existing datasets, specifically in post-storm conditions. This is a major cause for concern as arguably post-storm is one of the most critical times for tracking ammonium in a river. However, being unable to guard the sensors against external fouling equally a large concern when ammonium is so crucial to the success of S82 monitoring.
- Data Reliability: Due to the aforementioned issues, data from ISEs can be unreliable, especially for compliance monitoring under EA 21. As shown through publicly available data from the Environment Agency, data issues cannot be corrected because it is either impossible or inappropriate to do so. The current technical guidance even states that the technology is not at the level of accuracy and reliability expected for reliable measurement. Clearly, innovation, like that of optical ammonium, must be employed to make this form of measurement more reliable.
- Responsibility: As it stands, there has yet to be any formal comment on the degree to which ammonium ISEs are unsuited to fulfil the S82 requirements. It appears it has just been universally accepted that the existing technology will be used, with no regard to the very real shortcomings. Manufacturers have a degree of responsibility to provide unequivocable evidence that they can meet or exceed these standards both in terms of accuracy and over time. Manufacturers should not mislead users in suggesting that greater accuracies can be achieved than their specification datasheets. If these accuracies are true, then datasheets should be updated globally, and where applicable, that evidence should be provided.
ISEs in practice
It is possible to see these issues with ISE technology in existing EA monitoring datasets. We highly recommend viewing through the DEFRA Hydrology Explorer to really gain an understanding of how these factors interact to compromise the data integrity and interpretation. Unfortunately, the limitations of ISE technology means that ammonium datasets contain significant amounts of unreliable data; sensor drift is a significant issue in almost all datasets. For example, Figure 2 demonstrates how this data can be difficult to interpret and use; there is clear drift in the ammonium at the beginning of the dataset followed by erroneous data in late November/early December. From this point there would appear to be downward drift but, it’s almost impossible to distinguish if this is drift, actual trends or a combination of the two, and to what extent each is influencing the signal. Even with a monthly calibration and prompt tip replacement, there is no guarantee that data would be affected in the same way with each drift period. This highlights a big issue for interpretation, comparison, and application of these datasets; the ammonium data cannot be trusted as an accurate standalone indicator of pollution.
Figure 2: Example of drifting ammonium data in a pristine river, compared to dissolved oxygen level – data is from Hydrology Explorer. Dashed line indicates the data is ‘unchecked’.
Figure 3 displays a different dataset over 12 months demonstrating the difficulty of using ammonium ISE data; there are at least four significant jumps in the data where recalibration has been conducted, and the resulting impact on the signal compared to the period before calibration. This seriously calls into question the validity of the data as to how can drift be distinguished? Additionally, the measurement gaps between the before and after of each recalibration do not vary by a consistent magnitude. Therefore, attempts at retrospective data correction are exceedingly unlikely without a significant amount of investigation; additionally, that would still only be applicable to that specific location. Regrettably, the data shown in Figures 2 & 3 are not isolated either; this is likely to happen wherever an ammonium ISE is deployed.
Figure 3: Graph demonstrating difficulty of using Ammonium ISE data to provide usable insights into long-term water quality trends.
Setting aside the ISE technology specific problems, there is another to consider in the form of pH. Accurate ammonium/ammonia measurement is heavily dependent on accurate and reliable measurement of pH, temperature, and . Unfortunately, it is somewhat overlooked but accurate pH measurement is a crucial component for the calculation of ammonia; a drifting (upwards) pH sensor will effectively report an erroneously high level of ammonia, whilst a downwards drifting pH sensor will under-report the level of ammonia. At a pH of 8.0, the proportion of ammonia is around 10% whilst at pH of 9.0 that proportion approaches 50%! Hence, even if the ammonium sensor is highly accurate and well maintained, it also requires an extremely reliable pH sensor. As such, we recommend that anyone wanting to get close to an accurate ammonium/ammonia measurement must ensure that the pH sensor is also extremely accurate and reliable; and doesn’t need regular replacements. It is our experience that this sensor’s performance is significantly overlooked. With regard to its importance, the Proteus pH sensor has been designed to a robust specification, with a minimum operational life of five years without requiring any replacements or spare parts. It only requires a reference fluid change and a calibration every 3 months, therefore reducing OPEX costs for pH measurement much lower.
Such is the lack of assurance with ammonium ISEs, all manufacturers have limited their accuracy statements to ±2 mg/L and therefore it is extremely difficult to use these sensors to provide defensible data; admittedly they can be used to identify trends but is this limited ability really what we want for Section 82? One would argue not and that what we need is wholly reliable, drift free and accurate data to enable informed decisions upon; not sensors that drift and provide subjective data. Figures 2-3 effectively demonstrate the difficulty in using ISEs as defensible sources of data! Some manufacturers even recommend applying a site-specific offset to their data; doesn’t this simply undermine the stated accuracy and trustworthiness of the sensor?
On a side note, drift correction techniques have been a very poignant topic. However, the very idea of correcting drift for an ISE is not for the faint hearted. It is never possible to predict when drift occurs, and it will certainly not apply to the whole dataset between two calibration points. As such, applying a linear drift factor (or linear regression models) should be avoided as it would simply attempt to adjust some data that may already be correct. Using less invasive methods such as smoothing functions tend to run into issues when genuine peaks appear in the record and with a habit of pushing data into the negative at the lower end. Furthermore, wastewater discharges generally produce more drift in NH4 signals compared to those upstream of WWTWs. By utilising optical ammonium technology as a redundant sensor, it is possible to pinpoint exactly when drift starts occurring; as such operators can then adjust maintenance programmes accordingly based on drift threshold alerts (~0.1mgl difference over 3 hours).
3. Use of Ammonium ISEs for Mixing Point Surveys
Another rather amusing if not slightly concerning situation is the use of ammonium ISEs to determine the point of full-mixing downstream of a CSO with the receiving water. Logically, this approach is full of potential problems:
- Mixing surveys are never done in storms due to safety. The full mixing location during a storm will be very different to that under normal conditions. Flow conditions within the river and bankside would make this almost impossible to achieve; so why are we seeing articles/posts showing that it can be done? In reality, without the ‘marketing magic’, it will be extremely difficult and at very best would require major co-ordination and storm-chasing!
- True mixing surveys can only ever be conducted when CSOs spill and sometimes that might be years in between spills; which clearly isn’t practicable.
- Mixing surveys have only ever been demonstrated for final effluent discharges from WWTWs (and not CSOs) because they are continually flowing. Unfortunately, WWTW discharges only make up a fraction of all the 20,000+ monitoring locations required for Section 82.
- The limited sensitivity and low-end inaccuracy of ISEs means that the ISE will pinpoint the full mixing zone point much closer to the discharge point than it is; purely because of the accuracy and resolution of ISEs. Several manufacturer examples are being cited that ISEs can be used in this way but, in effect, they are simply demonstrating their poor accuracy/resolution to measure ammonium accurately. Modelling/CFD approaches, or more sensitive sensors will certainly provide more accurate data; the use of fluorescence sensors would present a more appropriate means of determining the mixing point as their measurement is in ppb and not mg/L.
Although this approach hypothetically has some merit, the stark reality is that it’s almost impossible to implement; and even if you could, the accuracy, the minimum detection and reliability of ammonium ISEs will not provide the true picture required to determine the mixing point.
4. Proteus Instruments Optical Ammonium Sensor
To address these challenges, Proteus Instruments has developed the world’s first instantaneous optical ammonium measurement technique for a sonde. Research has long shown that fluorescence can be used for measuring ammonium in the field. Baker et al., (2004) demonstrated strong correlations between ammonia and tryptophan-like fluorescence. However, all measurements were based on laborious grab samples and benchtop fluorescence spectroscopy to provide the data. With the latest R&D and innovation, Proteus has pioneered the field-based optical ammonium approach utilizing multiple fluorometers, turbidity and temperature to provide superior accuracy and reliability:
- Enhanced Accuracy: High precision in ammonium measurement, suitable for detecting low-level pollution events. With local calibration, our sensor can achieve accuracies <0.1 mg/L and can measure far lower than any ISE sensor. Our underlying measurement is measured in parts per billion and is instantaneous (<1 sec).
- Stability: our optical ammonium sensor experiences minimal drift, unlike their ISE counterparts, which ensures consistent data quality over 12 months (and potentially beyond). Furthermore, optical measurement is not prone to the same interferents that ISEs suffer from.
- Low Maintenance: Sensors require calibration only once per annum, not every 4-8 weeks. This can reduce OPEX by 90-95% compared to traditional ISEs. Furthermore, our sensors are designed to last >10 years.
- Application range: optical measurement is not limited to purely freshwater or <1bar depth applications. Proteus optical ammonium measurement can be used in marine or freshwater applications, deployed at up to 100 m depth.
For a full understanding of how optical ammonium measurement exceeds that of ISEs, please read this Article.
5. Long-term deployment testing
In 2023, a Proteus sonde was installed immediately downstream of a CSO to measure the impact of the overflow during storm events. The Proteus was equipped with pH, optical dissolved oxygen, conductivity, temperature, turbidity, tryptophan, CDOM, ammonium-ISE, optical ammonium and E. coli. Using a proStation Lite telemetry system all 15-minute data was transmitted hourly to our Wildeye cloud telemetry server.
The Proteus was independently calibrated and maintained by a third party to ensure there was no bias. Pre- and post-calibration measurements were also made with a different calibrated water quality meter. Although the ammonium ISEs provided generally reliable data based on monthly calibrations, the optical ammonium data did not suffer from any accuracy or drift issues commonly associated with ISEs. The standard deviation of the ammonium ISE against all of the reference measurements was 0.093mg/l whereas the optical was 0.054mg/l. Although this might not at first appear as a significant improvement, it demonstrates the strength in optical ammonium. On this occasion, the ISE was calibrated monthly, well maintained, and proven to be working at its optimum accuracy, in comparison, the optical received a single calibration at the start of the monitoring period and was maintained by the probe wiper and anti-foul guard without further intervention. The data collected clearly demonstrates that even when an ISE was performing at its highest, the optical ammonium was able to out-perform it, confirmed by a third ammonium instrument in an unbiased trial. The optical ammonium remained without intervention for the entire 18-month deployment period, as opposed to an ISE which would need monthly calibrations and ~3 tip replacements, due to their six-month expiry.
Figure 4 demonstrates the stability and accuracy of the optical ammonium sensor compared to well maintained and relatively accurate ISE sensor.
Figure 4: 4-month trend showing NH4 outputs for ISE & Optical
The optical sensor utilised nearly 2000 individual data points from the onboard ammonium ISE providing a RMSE of 0.025 mg/L (Figure 5). The ammonia measurements were typically very close to zero due to pH being 8.0-8.1. What this does demonstrate is that optical can measure an order of magnitude lower for both ammonium and ammonia.
Figure 5: Optical NH4 Calibration using 1980 values with inset graph showing readings <0.25mgl
The criteria as to what is ‘acceptable data’ for an ISE seems to vary from one user to another, with little indication as to the reasons behind it. This is where the use of ISEs becomes particularly concerning. By way of example, cited examples stating that any laboratory readings are to within ±0.25 mg/L of a continuous dataset (which averages 0.5 mg/L) but at the same time demonstrates no corresponding agreement in their trends, is misleading. It is therefore essential not to accept misleading claims about ISE performance or accuracy. Due to all the issues that ISEs suffer from and with over-arching performance accuracy of +/-2 mg/L, these sensors are limited to only being used as a yardstick for measuring ammonium; it should be noted that, in the majority of rivers/bathing waters, ammonium levels reside between 0.2- 1mgl (Environment Agency, 2014). Anyone that uses ISEs beyond their specifications, such as in continuous monitoring, is likely to make ill-informed, unsupported decisions.
Figure 6 is an example of a combined dataset using ammonium ISE sensors publicly viewable from the DEFRA Hydrology Explorer alongside a Proteus; the stability and sensibility of the Proteus data provides significantly more insightful . As is immediately obvious, the data from the EA sondes is difficult to interpret when coming to assessing conditions and ammonia levels instream. The optical ammonium is far more stable and gives a clearer indication of the ammonium dynamics. The optical ammonium visibly detects the smallest increases in ammonium, related to influences upstream of the sonde’s location.
Figure 6: 12-month dataset for Upstream (U/S) & Downstream (D/S) of CSO for Optical and ISE sensors
With no pre-filtering, reagents or regular calibration, this technique is ideal for challenging applications. The optical technique is technically capable of measuring more than 200 mg/L; and any one Proteus sonde can measure freshwater, marine, wastewater, industrial or potentially even drinking water applications.
6. Facing the Truth: Ammonium ISE Accuracy Claims
It doesn’t need an expert to state that ISEs are not ideal for long-term deployments. Manufacturers have a burden of responsibility to inform users of their proper use cases.
Unfortunately, the language surrounding ISE marketing across the industry relies on unsubstantiated language, to give the suggestion of “high accuracy” without ever stating whether the “high accuracy” is accurate enough for the application. Terms such as ‘long-lasting’, ‘high accuracy’, ‘long-term deployment’, ‘extended calibration intervals’, ‘high stability’ are bandied about carelessly, ignoring the actual implications of such statements. Low variation datasets are used to justify “low drift” without consideration of the far more variable conditions that these ISEs will be placed into for S82. A scan of literature from a range of manufacturers and sellers promotes ISEs for short-term deployment within their literature for good reason, as that is more where their usefulness lies; how can it be accepted that they have been put forward as a key component in S82 monitoring?
Ultimately, we as an industry need to move on from this language, be honest and simply accept that this technology is not currently fit for purpose. There is one very easy way to prove this; given all the concerns around ammonium ISE measurement, could anyone take a polluter to court and prove that they were polluting the river based on this data if the readings were below 2 mg/L. Furthermore, would any company willingly stake their reputation on their experiences of using ISEs? Given that almost all rivers and even most pollution events will see a dilution below 2 mg/L and that ISEs have an accuracy of +/-2 mg L, how can this data be enforceable; the measurement error of the sensor exceeds the measurement range of the application! We must move forward, embrace innovation, and develop new techniques. Unfortunately (to date) the technical guidance is not facilitating the use of innovation but instead holding on to outdated technology; there is little doubt that this will change.
Thankfully a trial by WRC later this year will evaluate field performance of the main providers of multiparameter sondes including Proteus. This will be independently assessed and reviewed; we thoroughly support these initiatives along with the opportunity to allow newer innovative technologies to be a part of the trial. Until now, this opportunity to demonstrate innovation at this level has been restricted; it’s certainly an opportunity that the water companies will embrace. The underlying message here is that we can all help here by simply embracing innovation and accepting that we need change.
Stable & Accurate: The future with optical ammonium
It is widely accepted that optical sensors provide excellent performance for a wide range of parameters. This has been proven the world-over with optical sensors providing: (a) higher precision and accuracy; (b) instantaneous direct measurement; (c) wider application ranges; (d) greater robustness with no moving/mechanical/replaceable parts; (e) lower OPEX costs and (f) much lower measurement ranges for some parameters. The optical ammonium sensor ticks all these boxes and, for the first time, can provide data just as robust as your turbidity or dissolved oxygen sensor. This is, for anyone with experience of operating ammonium sensors, a game-changer!
7. AI-Driven Insights
It is impossible to ignore the fact that artificial intelligence (AI) is changing our world almost by the second. The assumption of AI insights is that it’s built on accurate and reliable data. There is little doubt that AI-data will be incredibly useful but given that the industry is not currently comfortable with the performance and appropriateness of ammonium-ISE technology, a whole range of issues can arise:
- Propagation of Misinformation – AI data can simply amplify errors or biases within the data.
- Compounded Errors – data models trained on inaccurate data will produce inaccurate insights and hence poor decision making.
- Bias Reinforcement – the subtlety of AI is that it can reinforce the biases created leading to discriminatory outcomes.
- Lack of Trust – poor AI insights will only lead to users becoming sceptical of AI capabilities
- Legal and Ethical Issues – using inaccurate data, especially in bathing water applications, could have serious consequences, potentially even legal liabilities.
- Garbage In, Garbage Out (GIGO) – the golden rule of any modelling and prediction, if the input data is flawed, so will the output data.
We’re not just talking about ISE data either; installation, service, calibration, maintenance and other sensors all have a direct effect on overall data accuracy. Unfortunately, the focus on AI-driven insights has led to the neglect of ammonium ISE sensor performance, despite their crucial responsibility in providing accurate data. ISE drift, interference, and degradation, can result in flawed AI predictions. Many AI systems also lack mechanisms to detect faulty sensor readings, leading to misleading trends or incorrect actions. AI models cannot make up for poor sensor performance; only accurate and reliable sensors will provide the building blocks of good AI models. Therefore, we should not rely on AI as a solution for poor sensor performance. To do so could lead to critical errors and disastrous decision making with wide-ranging consequences. Doing so could lead to critical errors and disastrous decisions.
8. Technical Standards & Real-World Applicability
RS Hydro have supplied the largest network of water quality monitors in the UK for Northern Ireland Water (NIW). Although NIW do not have to comply with Section 82, they have already deployed over 150 outstations. You will not be surprised to understand that maintaining a high degree of accuracy/reliability with ISEs is very challenging. The data on the EA/DEFRA’s Hydrology Explorer is not a world-away from the issues that NIW will have to contend with. Even at this relatively small number of installations, using ISE technology is likely to be unsustainable. 20,000 of these devices will, in effect, leave operations teams beyond overstretched and overwhelmingly resource poor. The discussions of how to better manage the scenario will be a recurring management dilemma for many AMPs to come, wasting significant amounts of time in both meetings and site visits.
It has been stated within the current technical guidance that the calibration intervals could be extended based on evidence. The apparent unanswered question is what would classify as evidence; some of the suggestions of would not be suitable and there is a clear requirement for a performance standard, such as MCERTS, to provide a common standard. However, the roll-out of this standard could be many years away and so the question then becomes, how will this process of providing evidence-based data to extend calibrations be regulated? Surely the only way would be to provide ISO17025 assured data until that point? The technical guidance, legislation and this story clearly has some way to go before providing the framework that we all need; some would argue that as the guidance has some way to go, so shouldn’t we get this right first before spending £billions on new monitoring installations and collecting rafts of unusable data?
9. The Future . . .
Innovation, collaboration and logical thinking! At a recent SWIG event (January 2025) regarding Section 82 compliance, a very clear message was made: the current guidance is not robust enough to deliver the comprehensive monitoring and data reliability that both water companies and the public will need. There was clear recognition that operational technology must advance beyond where it is now; and that we should not hold on to the past. Optical ammonium from Proteus is certainly one technique but the need for truly innovative new techniques across the board is required urgently if the existing technical guidance under Section 82 is to be met.
Afterall, what will all this data mean to the average member of the public?
What will happen to the masses of data collected?
How will water companies handle the flood of enquiries that will relate to poor datasets?
How do accuracy, calibration, maintenance and chain of custody affect reliability in legal cases? Lots of questions…
We know that the data will end up on some form of public data interface but how will anybody make sense of all this data, especially if it’s interlaced with all the highlighted issues that surround ammonium (and potentially other) measurements. Surely, we need something more reliable and more meaningful. Given that river health is the priority here, most people can interpret and make immediate decisions based on parameters such as E. coli and Enterococci far quicker, perhaps even BOD, COD, phosphate and blue-green algae. How will those people be able to make the same decisions based on some of the ammonium-ISE data cited within this article…The fact remains that sewage is simply one component of a far larger puzzle when it comes to UK river and marine pollution. If we are going to monitoring, we must have absolute confidence in the data, eliminate subjectivity using mediocre parameters and then provide the transparent overview that the public will require without the need to patch up poor datasets with millions of man-hours of data cleansing.
Although we (Proteus Instruments) are privileged as a company to know what good data looks like, optical ammonium being just one example, we do understand that it cannot necessarily replace the current approach immediately. However, Proteus is the only fully compliant instrument available in the market-place due to its ability to meet the full specification, provide a minimum of 10 sensors and have an additional 3 spare ports, over and above the 2 spare parameters required. This allows us to utilise sensor redundancy (ISE or optical) to provide accurate, reliable, and dependable data using the most consistent sensors on the market.
At the very least optical ammonium can provide the absolute reassurance to water companies that their ISE data is accurate (or not). It will pin-point exactly when there is an issue with the sensor, exactly when it starts drifting and hence give the ability back to the water company to provide predictive maintenance and improve monitoring. This will avoid water companies populating public dashboards with erroneous data and being inundated with RFIs. Surely, our approach needs to build trust between water companies and the public not the reverse. As we all know just a single false dataset event could completely undermine all this investment overnight!
It is hoped that the trial run by WRC and supported by seven water companies later this year will provide the benchmarked performance that we all need to understand the performance of the EA21 sensor technologies more fully.
Ultimately, our goal is to protect the environment, not to drive instrument sales, create unnecessary jobs, or increase carbon emissions. By prioritizing accurate data and well-informed decisions, we can reduce waste, minimize our carbon footprint, and restore our rivers into thriving ecosystems—getting it right the first time for a truly sustainable future.
For a more detailed overview of the benefits of Optical Ammonium please read the following Article:
Authors:
Rob Stevens, Managing Director, Proteus Instruments, RS Hydro
Hannah Gunter, Project Research Manager, Proteus Instruments
References
Baker, A. and Inverarity, R. (2004), Protein-like fluorescence intensity as a possible tool for determining river water quality. Hydrol. Process., 18: 2927-2945. https://doi.org/10.1002/hyp.5597
Blaen PJ, Khamis K, Lloyd CEM, Bradley C, Hannah D, Krause S. Real-time monitoring of nutrients and dissolved organic matter in rivers: Capturing event dynamics, technological opportunities and future directions. Sci Total Environ. 2016 Nov 1;569-570:647-660. doi: 10.1016/j.scitotenv.2016.06.116. Epub 2016 Jul 2. PMID: 27376920.
Castrillo, M, García, Á.L (2020) Estimation of high frequency nutrient concentrations from water quality surrogates using machine learning methods,Water Research, Volume 172, 115490, ISSN 0043-1354, https://doi.org/10.1016/j.watres.2020.115490.
DEFRA (2025) Hydrology Data Explorer. [online]. Available at: https://environment.data.gov.uk/hydrology/landing [last accessed 21/03/2025]
UK Government (2025) Environment Act 2021 Section 82 [online]. Available at: https://www.legislation.gov.uk/ukpga/2021/30/section/82 [last accessed 21/03/2025]
a CWQM Interim Technical Guidance