Assessment of limited downstream nitrate dosing for sulphide control in pressure sewers: case study in Germany

Wastewater discharges from intermittently operated pressure sewers used for collection services in rural and sub-urban areas are ubiquitously known to cause severe odour and corrosion problems. In central and eastern Europe, pressure sewers are commonly used to convey wastewater from rural and suburban areas to a centralised wastewater treatment plant. With the pressures of the European Water Framework directives in 2000 to expand the scope of water protection and achieve good chemical and ecological quality for all waters (surface and groundwater), the member state countries needed to rethink their sanitary services in their rural regions. In response to this mandate, extending the sewerage system was one option implemented by several water service providers (Gerend 2019). This transformation resulted in the installation of long pressure sewers to collect and transport wastewater in rural and suburban areas to a centralised system where the wastewater is directed to a treatment facility.

The wastewater transported in pressure sewers found in rural areas is typically characterised by long residence times and is under anaerobic conditions for most of the time. Under these conditions, sulphate-reducing bacteria (SRB) are active and play an important role in the sulphur and carbon cycle by coupling the production of hydrogen sulphide with the oxidation of organic compounds (Villahermosa et al. 2016). Sulphate reduction by the SRB residing in biofilms on the sewer walls is mainly responsible for the build-up of H₂S concentrations in the bulk wastewater (Li et al. 2017). H₂S is subsequently released to the gas phase at points in the sewer network associated with turbulent conditions and is influenced by the liquid phase H₂S concentration, pH and temperature. In the gas phase, H₂S is well-known for causing microbial-induced corrosion in concrete sewers, obnoxious odours, and its toxicity to sewer workers and residents when released to the sewer atmosphere and its surrounding areas.

To avoid corrosion, an H₂S concentration of less <3 ppm is necessary to mitigate corrosion risk to a rate of 10 mm a⁻¹ (Weissenberger 2002), and odour is perceptible from 0.5 ppm H₂S (Hvitved-Jacobsen et al. 2013). To achieve these H₂S control targets, mitigation strategies aim to inhibit the production of sulphide by inactivating SRB bacteria, reducing the H₂S concentration in wastewater, or treating the H₂S gas released. However, the problems are often detected in ‘already built’ sewer systems, so construction measures to avoid odours are difficult to implement. Possible countermeasures in an existing sewer system include the introduction of oxidising compounds, e.g. hydrogen peroxide (H₂O₂) or oxygen (O₂) (Saračević 2009; Urban 2010), the dosing of alternative electron acceptors like nitrate, the precipitation of sulphides, e.g. by iron salts or the prevention of H₂S emission by increasing the pH level (ATV-DVWK 2003; Zhang et al. 2008).

In Germany, mainly calcium nitrate, flushing with pressurised air and iron salts are used (ATV-DVWK 2003; Ott 2004 in Barjenbruch 2007). A chemical dosing survey in Australia revealed that iron salts and oxygen are mainly used for sulphide control (Ganigue et al. 2011). They are easy to implement and cheap, and many companies specialise in implementing such dosing systems. The most straightforward implementation is to continuously dose the chemical into the wet-well (Feldhaus et al. 2005; Frey 2008). For both these strategies, chemical dosing is made at the upstream point. Therefore, enough chemical is required to prevent anaerobic conditions throughout the entire length of the pipe or precipitate the sulphide produced by the sewer biofilms. Hence, for the nitrate dosing, enough nitrate has to be added to account for the consumption in the wastewater, biofilms and sediments (Friedrich et al. 2004; Frey 2008). The disadvantages of nitrate dosing at an upstream point such as the wet-well are the consumption of easily degradable COD, the possible accumulation of polysulfides leading to higher sulphide production once the dosing is stopped, the adaptation of the denitrifying bacteria increasing the amount of chemical and thereby costs over time, a thickening of the biofilms and the possible stimulation of SRB downstream of the dosed sections (Feldhaus et al. 2005; Frey 2008; Mohanakrishnan et al. 2008; Saračević 2009; Auguet et al. 2015; Liang et al. 2016). Some of these problems might be reduced by a different dosing strategy: the downstream dosing of nitrate, as suggested in (Auguet et al. 2015) and in (Friedrich et al. 2004). In this approach, less nitrate is dosed. Only enough nitrate for oxidation of the sulphide that is already generated and the consumption in the remaining pipe section is required. Thereby chemicals can be saved, and the adaptation of biofilms is also reduced due to the limited amount dosed (Auguet et al. 2015).

This study aims to assess the effectiveness of limited downstream nitrate dosing for sulphide control on a 9.0 km pressure sewer pipe. Year-long monitoring of the liquid phase H₂S concentration was made to capture the seasonal effects in response to the downstream dosing strategy. Furthermore, in this study, we also assessed the impact of residence time on the performance of downstream nitrate dosing in an attempt to determine the period that was dosed ineffectively. As part of understanding the influence of the residence time on the dosing strategy, we examined how the pump operation schedule at the transfer station and hydraulic conditions could influence the sulphide formation, which can provide valuable insight into optimising the pumping schedule for other sewer systems with similar characteristics. Finally, we demonstrated how wastewater flow data could help optimise the dosing strategy and reduce chemical input. In the end, a cost comparison of upstream and downstream dosing is performed to showcase the value of applying a limited downstream nitrate dosing.

At the first dosing station, installed at the Leopoldshagen vacuum pump station, FeCl₂ is used to suppress the sulphide produced in the Mönkebude sewer section (L = 4.9 km, DN = 200 mm) and thereby aims to alleviate the release of H₂S when discharged in the Mönkebude transfer station. One of the significant contributors to the sulphide problem of this sewer section is the abnormally long travel time through the pressure sewer system. Daily wastewater volumes ranging from 3 to 50 m³ per day are periodically pumped into this section resulting in residence times ranging from 16 to 48 hours. Even with the ferrous chloride dosing of 66 L d⁻¹ applied upstream, relatively high H₂S concentrations persist and are only partially effective in controlling the sulphide formation in this sewer section. Furthermore, the chemical dosing in this section is plagued with several operational challenges, for example, the formation of iron precipitates at the dosing point, which was assumed to affect the magnetic field of the wastewater IDM flow meter resulting in false wastewater flow. Consequently, this results in disproportional FeCl₂ concentrations (chemical delivery per volume of wastewater).

Given the challenges of the upstream ferrous chloride dosing and that there were ongoing plans to reconfigure the upstream dosing, we focus our efforts on reducing the sulphide concentrations in the Grambin sewer section (GSS). The wastewater from the Mönkebude sewer section is transferred into the GSS (L = 4.9 km, averaged DN=190 mm) at the Mönkebude transfer pump station (Figure 1). At the discharged point, wastewater from the pressure pipe enters the gravity sewer, which runs through the populated areas in the town of Ueckermünde before making its way to the treatment plant. At this point, the wastewater has its first contact with air and due to its septic status, it causes serious odour problems along the first 200 m from the point of exit. Furthermore, the hydraulic residence time in GSS ranges from 8 to 20 h, depending on the time of day. To reduce the intense odours at the end of the pressure sewer and its surroundings, a chemical dosing station was installed near the downstream end of the pressure sewer section, as shown in Figure 1.

To identify the dry weather flows, we followed the guidelines in ATV-DVKW A 198E (Standardisation and Derivation of Dimensioning Values for Wastewater Facilities 2003). According to the standard, the daily dry weather flow can be determined either by using the number of dry days based on the weather log of wastewater treatment plants (Chap. 4.2.2.1 Para. 1) or by using the mathematical derivation based on Fuchs et al. (2003) (Chap. 4.2.2.1 Para. 4). Both methods above were applied in this study.

⁠: H₂S produced when the wastewater slug arrives at the end of the sewer section.

⁠: residence time of the wastewater slug when it arrives at the end of the sewer section.

Flow proportional nitrate addition is made by a Memdos LP 4 (Lutz-Jesco GmbH) dosing pump which is operated in analogue input mode. At maximum operating capacity, the dosing pump can dose up to 4 L h⁻¹. When operated in analogue input mode, an external 0/4–20 mA signal controls the stroke frequency of the dosing pump. In this case, the external signal comes from the non-invasive ultrasonic flow meter (FLUXUS F501, SebaKMT), in which a wastewater flow rate of 90 m³ h⁻¹ is used as the current value (20 mA) for 100% delivery capacity whereas, for 0% delivery capacity, a current value of 4 mA is used for a flow rate of 0 m³ h⁻¹. Based on the measurements obtained from the 2018 to 2019 monitoring campaign, the amount of Ca(NO₃)₂ solution delivered per cubic meter of wastewater is indicated in Table 1.

For the evaluation of the chemical dosing trials, we calculated the H₂S removal efficiency of the dosing and classified the H₂S measurements under dosed conditions according to defined evaluation criteria. Baseline periods (reference period where no chemical dosing was administered) were established by switching off the chemical dosing pumps. Periods selected for baseline monitoring (without dosing) lasted for 5–8 days, with the first two days not included in the evaluation dataset. Leaving out the first 2 days ensured that the nitrate residuals in the dosed section were completely flushed out. The evaluation criteria were based on four risk levels for sewer corrosion and odour. The risk is associated with the measured H₂S concentration and is grouped into either negligible, medium-low, medium-high, or high (CH2M 2017). The evaluation criteria for classifying the H₂S measurements into respective risk groups are presented in Table 2.

The unionised form of sulphide, H₂S (referred to as liquid phase H₂S in this article), was the chosen parameter for optimisation of the downstream nitrate strategy applied in this study. We chose this parameter because (1) only the unionised H₂S can be emitted to the sewer atmosphere, causing odour and corrosion problems; (2) the sensor used for optimisation directly measures the unionised H₂S. The SulfiLogger^TM S1/X1-1020 (SulfiLogger A/S, here forth referred to as SulfiLogger^TM) were installed in the Mönkebude transfer station (inlet) and at the end of the Grambin sewer section (Figure 1) for monitoring the liquid phase H₂S. Measurements were made in 1-minute intervals between 11.10.2019 and 03.03.2020 using the SulfiLogger^TM.

To characterise the activity of the SRB during the baseline monitoring periods, the liquid phase H₂S measurements were converted to total dissolved sulphide concentration following (Despot et al. 2021). Since the pH values recorded during the baseline monitoring were not made continuously, the median of measurements obtained was used in the conversion procedure.

Grab samples were taken at the inlet and outlet of the treated sewer section during the baseline and chemical dosing monitoring field campaigns. The samples were immediately stored in a cooler box and taken to the laboratory, where they were either measured immediately or cooled to 4 °C and analysed the next day. For the sulphide measurements, 250 ml of the sample was preserved on-site by adding aluminium and sodium hydroxide for flocculation and zinc acetate to avoid losses during transportation to the lab. Total dissolved sulphide measurements were made using a modified version of Saračević (2009) with the sample preservation method adapted from the APHA Standards (APHA 2018). Temperature and pH measured on-site using HQ-D40 electrodes (Hach-Lange) were used for conversion between the sulphide species. For the soluble COD, sulphate and nitrate measurements, cuvette tests by Hach-Lange were used (LCK 514, LCK 153, LCK 339).

To understand the wastewater composition entering the pressure sewer and its subsequent transformations relative to the sulphide build-up in pressure mains, sulphide and sulphate, COD, dissolved oxygen, pH and temperature were monitored both at the inlet and outlet of the pressure sewer. The wastewater composition at the inlet of the pressure sewer can be described as medium strength (Hvitved-Jacobsen et al. 2013). At the inlet (Mönkebude transfer station), the soluble COD ranged from 373 to 413 mg L⁻¹, with the highest values being measured during the winter. Generally, the soluble COD for this study site accounted for 40% of the total COD. Sulphate concentrations ranged from 62 to 183 mg SO₄²⁻ L⁻¹, with the highest values, also detected in winter. The pH values measured for the different seasons were in a similar range, between 7.8 (winter) and 8.3 (summer). Regarding the sewage temperature at the study site, a 10 °C difference between summer and winter temperatures was recorded. Differences between the inlet and outlet sewage temperature were minor (see Figure S2). Under dry weather flow conditions, the dissolved oxygen concentration in both the inlet and outlet of the sewer section was <1 mg L⁻¹.

Using the calculated anaerobic residence time and the area to volume ratio of 40 m⁻¹, the maximum area sulphide production rate of the biofilm was calculated to be between 0.43 and 2.1 g S m⁻² d⁻¹ for the baseline monitoring periods of the different seasons (Figure 3(b)). The sulphide production rates measured during summer had the highest values, which coincides with a temperature increase of 10 °C in reference to the wastewater temperature in the winter. Therefore, the seasonal variation in the sulphide production rates indicates the changing SRB activity throughout the year and the need to implement a season-factor dosing scheme. It is important to realise that the transition periods from winter to spring and from summer to autumn may consequently result in periods of underdosing and overdosing and, therefore, should be carefully investigated. The effects caused by the transition periods were observed in our study, with noticeable differences between the area sulphide production rates in autumn and spring. Finally, it is important to note that sulphide formation persists during the colder months of the year. Consequently, it is necessary to implement chemical dosing throughout the whole year.

Figure 4(c) shows the anaerobic residence time, which corresponds to the time the wastewater slugs spend in the section from the Mönkebude transfer station to the Grambin DS. The anoxic residence times, shown in Figure 4(d), represent the duration the wastewater slugs stay in the anoxic or dosed section of the pressure sewer main from the DS to the end of the pressure pipe at the entrance to Ueckermünde town. It is visible that the anoxic residence time follows a distinct daily pattern. The slugs arriving at the end of the pressure sewer between 4:00 and 6:00 am have the longest travel time. Also, a clear pattern with the highest retention times between 5:00 and 7:00 am is visible for the anaerobic residence time, although it is not as defined as the anoxic retention time pattern. A reason for that could be the undefined pressure sewer connections from small pump stations and houses on the way from Mönkebude to Grambin. In contrast, there are no additional connections in the last part of the sewer. The lowest absolute residence time for both anoxic and anaerobic sections can be observed in summer and spring 2020, where stay-at-home measures due to the COVID-19 pandemic were in place. The reason for this is the higher flow in summer and during the first month of COVID-19 lockdown, shown in Figure 4(a) and 4(b). Abu-Bakar et al. (2021) and Lüdtke et al. (2021) demonstrated how the UK and Germany's COVID-19 lockdown measures drastically altered water consumption patterns. Both studies indicated that the lockdown measures instigated a significant increase in water consumption, directly translating into increased wastewater production. The lowest flow in autumn corresponds to the longest retention times in the anaerobic and anoxic sections.

Table 3 shows the summary of the residence time analysis. Considering both wet and dry weather flows, the mean daily flow of the Grambin sewer section is 191 m³ d⁻¹. Although the sewer system is a separate sewer system, excess water during wet weather events increases the mean daily flow to 214 m³ d⁻¹. The higher flow directly leads to a lower retention time. On average, the wastewater spends 16.5 h in the pressure sewer from Mönkebude to Grambin, the last 3 hours under anoxic conditions in the dosed section of the pipe. The residence times in our study are relatively long when compared to other field studies, which recorded highly septic wastewater after more than 3–7 hours (Frey 2008; Jiang et al. 2013a). With downstream nitrate dosing, an ideal residence time in the dosed section of less than 90 min is recommended due to the depletion of the oxidising agent and continuation of sulphide formation (Friedrich et al. 2004). The 3 hours recorded in our study are therefore longer than recommended for this dosing strategy. The residence time is dictated by the daily wastewater flow, which depends on the water consumption. Adapting the chemical dosing based on the system characteristics (defined by the consumption and storage capacity) must be considered when planning chemical dosing measures for sulphide control.

The inability to achieve the reduction target during the autumn and winter monitoring periods is the result of the higher upstream H₂S loads resulting from the insufficient sulphide removal of the upstream FeCl₂ dosing. It was found that when the upstream sulphide concentration was less than 0.5 mg L⁻¹, the H₂S measured at the end of the pipe was <0.2 mg L⁻¹. Furthermore, the long residence times in the anoxic section promote the regeneration of sulphide production (Friedrich et al. 2004). When the nitrate added is utilised for the oxidation of the sulphide load coming from the wet-well as well as the sulphide generated in the anaerobic section before the dosing point, the residual nitrate falls to the typical nitrate concentration in domestic wastewater (<1 mg NO₃⁻-N L⁻¹). At this point, the artificial anoxic conditions no longer prevail, and SRB activity resumes. The sulphide oxidation process occurs within the first 2 hours after being dosed (Friedrich et al. 2004). It is assumed that the wastewater slugs are exposed to a transition from anoxic to anaerobic conditions between the 2nd to 3rd hour after entering the dosed section. The nitrate concentration was 0.8 mg NO₃⁻-N L⁻¹±0.1 (n=14) after the dosing was switched off for one day, which is typical for wastewater in pressure sewers. During dosing, the nitrate concentration downstream was, on average, 1.7 mg NO₃⁻-N L⁻¹±0.3 (n=8). These samples were taken between 9:30 and 11:00 am for technical reasons and do not represent the critical areas in the early morning where nitrate is completely depleted (see red shaded regions in Figure 5(a)).

When a residence time of 3 hours is exceeded, anaerobic conditions prevail once more, favouring the regeneration of sulphide in the pressure main. This effect is shown in the shaded regions (in red) in Figure 5(a), where the SulfiLogger^TM was able to detect the regeneration of the sulphide because of the long residence times in the anoxic section. Another essential point contributing to the sulphide peaks highlighted in Figure 5(a) and 5(b) is the increasing residence times of the wastewater slugs arriving at the dosing point. At this time of the day, the residence times of the wastewater slugs approach their maximum, thereby increasing the sulphide concentration beyond the expected limit. This indicates that an additional amount of nitrate is needed to effectively reduce the sulphide concentration of the slugs arriving at the dosing point between 4:00 and 8:00 am. Furthermore, since the nitrate is also used as an electron acceptor by heterotrophic bacteria to oxidise organic matter, supplying an additional amount of nitrate for this process must be considered (Yuan et al. 2015).

Figure 6(b) displays a typical departure–arrival time curve for the wastewater parcels transported through the dosed sewer section. Considering the interquartile range, departure times that had anoxic residence times >3 hours and corresponded to higher H₂S concentrations were identified to be in the time window between 9:30 p.m. and 3:30 a.m. This time window marked the arrival of the wastewater parcels from the Mönkebude pump station, which was likely to cause sulphide peaks between 4:00 and 8:00 a.m. Therefore, targeting the corresponding wastewater slugs entering the dosing point will improve the limited downstream nitrate dosing performance.

So far, the limited downstream nitrate dosing application has proven to be very effective given the low chemical cost per year (1,853 € yr⁻¹) and relatively high reduction percentages. A nitrate dosing range of 1.5–5 mg-N L⁻¹ was shown to reduce the H₂S concentration to 0.60–0.26 mg L⁻¹. Despite the remarkable performance of the limited dosing strategy, we found that the success of the dosing at this point was strongly dependent on the residual sulphide concentration in the upstream wet-well, the sulphide build-up in the sewer section before the dosing point and the residence time in the dosed section. The anaerobic residence times of the wastewater slugs entering the dosing point are strongly influenced by the sulphide build-up in the un-dosed section and are responsible for the sulphide peaks occurring between 4:00 and 8:00 am. Moreover, when the residence times in the dosed section exceeded 3 hours, sulphide production resumed, and higher H₂S concentrations were detected by the SulfiLogger^TM.

To address the concerns of the upstream sulphide concentration and the wastewater slugs remaining in the dosed section for periods >3 hours, we propose the following measures:

• 1.

  Adjust the amount of nitrate added to account for upstream sulphide loads.

• 2.

  Implement residence-time-based dosing by defining the time window for the wastewater slugs residing in the anoxic section for >3 hours, targeting the sulphide peaks typically produced during 4:00–8:00 a.m.

For the upstream dosing, approximately five times more nitrate solution is needed than the downstream dosing as it is applied currently, hence almost five times the costs for chemicals. With the current dosing, sulphide peaks in the morning are still noticeable. However, the suggested profile dosing addresses the sulphide peaks (4:00–8:00 a.m.) only at an operating cost of around 750 € more per year. One reason is that the time window that requires higher nitrate dosages is the period of the lowest flow. The main reason for explaining the difference between the applied downstream dosing (Scenario 1) and the profile dosing (Scenario 2) is the difference in sulphide production rates between autumn and spring. Indeed the difference is a direct result of the autumn having higher residence times than spring. Furthermore, we used a slightly higher N-NO₃/H₂S ratio to compensate for the consumption of heterotrophic organic matter oxidation during the transport in the dosed section (Yuan et al. 2015). It is important to emphasise that no special or additional equipment is needed for implementing Scenario 2 since the current dosing pump in Grambin can deliver different dosing rates considering a pre-defined pattern.

The nitrate demand is reduced by 73% when considering Scenario 2, which is expected to result in the complete abatement of H₂S produced in the Grambin sewer section. Similarly, Gutierrez et al. (2010) showed that nitrate consumption was reduced by 42% when a modified RT-based dynamic downstream dosing was performed on a sewer laboratory reactor. Overall, both limited downstream nitrate strategies (Scenario 1 and Scenario 2) offer potential savings and are significantly cheaper when compared to the theoretical upstream requirements.

According to Mohanakrishnan et al. (2009), upstream nitrate dosing increases the sulphide production capability of the downstream biofilm; this, in turn, increases the overall capability of the sewer section for sulphide production when anoxic conditions no longer prevail. If upstream nitrate dosing continues over time, higher nitrate dosages will be required, increasing chemical costs. The upstream nitrate dosing becomes increasingly expensive in the long run and for long pressure sewers. These findings should be considered when implementing a new dosing concept in any location where long pressure sewers are utilised.

Regarding undesirable effects of nitrate, especially when dosed in excess, nitrous oxide (N₂O) formation and emissions, especially if trace amounts of oxygen are present, are almost certain to occur (Zhang et al. 2021). N₂O is a severe greenhouse gas with a global warming potential of 298 times the one of CO₂ (US EPA 2019). According to Short et al. (2014), past research shows that the wastewater type (e.g., C/N ratio) and organic loading rate, aeration regime (aerobic–anoxic cycling) and environmental parameters such as dissolved oxygen (DO) concentration or pH influences N₂O generation in wastewater systems. In addition, the concentration of substrates (NH₄⁺-N, NO₃-N) and intermediates (NO₂-N, NO and free nitrous acid), as well as the abundance and activity of N₂O-producing microorganisms (i.e. nitrifying and denitrifying microbes), are also linked to N₂O generation. Therefore, minimising the amount of the residual nitrate concentration of the dosed sewer sections is an important measure that must be considered when using N-based salts such as nitrate. Auguet et al. (2015) indicated that limited nitrate dosing downstream produces negligible N₂O emissions since the added nitrate was mostly utilised. However, in the case of upstream nitrate dosing, excess dosing is likely to result in high residual nitrate concentrations; therefore, N₂O formation is more likely. To evaluate the depletion of nitrate, nitrate was measured at the end of the pressure sewer, with the majority of the nitrate measurements (85%) having a concentration <1 mg N L⁻¹, indicating that added nitrate was mostly depleted. Also, the wastewater exiting the dosed sewer section is expected to be mixed and diluted with wastewater entering the main gravity sewer system at downstream points, further decreasing the possibility of N₂O formation.

Besides the possibility of contributing to the GHG, high nitrate dosing rates increase the consumption of readily biodegradable carbon sources during denitrification, which is likely to affect biological processes at the treatment plant (Liu et al. 2015). Furthermore, nitrate, among other oxidants like ferric iron and hydrogen peroxide, was found to cause the production of other volatile odorous sulphur compounds like dimethyl trisulfide when dosed intermittently (Gu et al. 2019). Furthermore, the presence of nitrate increased sulphide production compared to the original state when dosing is interrupted due to polysulphide accumulation during dosing (Liang et al. 2016).

According to the findings of Mathioudakis & Aivasidis (2009), temperature significantly impacts the denitrification kinetic profile. The percentage of nitrate that is completely denitrified to dinitrogen gas increases, and the nitrite accumulation in wastewater is suppressed. This observation suggests that the dosed sewer section will remain anoxic for longer periods under lower temperatures <15 °C. Therefore, nitrate dosing is better suited for lower temperatures, for example, from October to April in central Europe. At higher temperatures, the consumption increases, making nitrate dosing costly and ineffective. Therefore, careful consideration to avoid high nitrate dosages for extended periods should be made, especially since the nitrate uptake capability of the biofilm increases with repeated exposure to nitrate (Mohanakrishnan et al. 2009).

Given the potential undesirable effects of nitrate dosing, it is worth exploring alternative technologies for sulphide control that can be applied downstream. An alternative to nitrate dosing for downstream sulphide control could be ferrous chloride. The main mechanism here is not oxidation but sulphide precipitation, which was proven to have fast reaction kinetics. However, studies point out that this reaction also needs considerable time (Kiilerich et al. 2017), and ferrous will not have an inhibitory effect on methanogens like nitrate (Jiang et al. 2013b). Also, there have been trials of dosing pressurised air downstream with a perforated tube around 500 m before the transition to the gravity pipes. With this strategy, good ventilation of the pressure pipe and a positive slide slope of the pipe have to be ensured for a safe operation (Urban 2010; Baxpehler & Urban 2020).

Liquid phase H₂S measurements for monitoring chemical dosing made it possible to closely examine the diurnal sulphide pattern and influence seasonal changes and other flow events. Thereby it was possible to define a profile dosing strategy with a specified time of higher dosages without drastically increasing the overall chemical dosage. However, this dosing profile needs to be constantly updated and checked for changes regarding shifts in seasonal and daily residence time and sulphide profiles. Therefore, installing online H₂S liquid phase sensors such as the SulfiLogger^TM is a valuable tool for continuously optimising and fine-tuning the chemical dosing and ensuring that an effective dose-response relationship is achieved. Integrating these sensors in an online dosing control system can regulate the delivery of appropriate chemical dosages according to the measured H₂S, offering an opportunity for water utilities to reduce chemical addition for H₂S control. Given that the SulfiLogger^TM only measures unionised sulphide, H₂S, a conversion to total dissolved sulphide is required for characterising the SRB activity (sulphide production rate/ sulphate reduction rate) in the sewer section being dosed. Therefore installing a corresponding pH measurement at the dosing point that is made continuously will improve the accuracy when estimating the sulphide production rates and ultimately improve the optimisation of the control strategy presented in this study.

The limited downstream nitrate dosing applied performed exceptionally well given the low nitrate concentrations used. The sulphide loads in the upstream wet-well significantly influenced the performance of the downstream nitrate dosing. When the upstream sulphide concentration was less than 0.5 mg L⁻¹, the H₂S measured at the end of the pipe was <0.2 mg L⁻¹. Besides the upstream sulphide concentrations, the residence time of the wastewater slugs arriving at the dosing point and the time these slugs stay in the dosed section were identified as the main reasons for the sulphide spikes frequently occurring between 4:00 and 8:00 a.m. When residence time in the dosed section is >3 hours, sulphide regeneration was likely to occur. The wastewater slugs arriving at the DS between 9:30 p.m. and 03:30 coincide with residences that were >3 hours. The slugs arriving between 4:00 and 8:00 a.m. had the highest residence times, which also contributed to sulphide peaks. To remove the sulphide peaks usually found between 04:00 and 08:00 hours, we recommend increasing the nitrate dosing concentrations for the slugs arriving between 9:30 p.m. and 7:00 a.m. at the DS. This study has demonstrated how continuous monitoring of liquid phase sulphide can be used for optimising chemical dosing. Furthermore, the advantages of downstream nitrate dosing compared to upstream dosing, including chemical savings and odour relief, were clearly shown.

The authors are grateful for the support from Gesellschaft für Kommunale Umweltdienste mbH Ostmecklenburg-Vorpommern (GKU) during the field measurement and SulfiLogger A/S for providing the online liquid phase H₂S sensors for optimising the chemical dosing.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.