Improving microbial water quality in WWTP effluent through post-treatment in a constructed wetland

(This is a translation of a Dutch article that was published in H2O-online / 18th May 2021)

Aleida Hommes-de Vos van Steenwijk, Marc van Bemmel (Orvion), Henk Tamerus, Oscar van Zanten (Dutch waterboard de Dommel)

The constructed wetland ‘waterpark Groote Beerze’ provides a natural post-treatment of the effluent of the municipal wastewater treatment plant (WWTP) de Hapert. The effect of this post-treatment step on the microbial water quality was not known. Using a combination of DNA techniques (NGS and qPCR) this study showed that the microbial composition of the WWTP effluent significantly changed as a result of post-treatment and that as a result the impact on the receiving water is minimal.

The waterpark Groote Beerze at WWTP de Hapert was constructed in 2001 and was the first natural post-purification (constructed wetland) of the Dutch waterboard ‘Waterschap de Dommel’. In 2018, the waterpark was optimized to improve its performance and make it more environmentally friendly. The waterpark is a system of ponds, ditches, open waters with aquatic plants and swamp forest, that converts the effluent from the WWTP into natural freshwater. The Groote Beerze stream (where the WWTP discharges its effluent) is one of the most natural streams in the management area of the Waterboard . It is  a Natura 2000 area, a large EU network of protected natural areas. The stream valley is home to valuable plant and animal species such as bluegrass meadows, water plantain, water annelids and the spined loach. The main purpose of the waterpark is to better protect this vulnerable ecosystem from the WWTP discharges. In addition, the waterpark also fulfils functions as (rainwater) buffer, nature development and recreation and education.

Part of the naturalisation of effluent water in the waterpark is the removal of microorganisms from sewage that enter the wastewater treatment plant. After all, these are not part of a natural stream. However, the extent to which this is achieved is unknown. Concentrations of culturable indicator bacteria, such as Escherichia coli, decrease in constructed wetlands as a result of natural mortality, through sedimentation, biological filtration and by UV irradiation [1]. However, analyses based on culturing provide too limited a picture. The effluent also contains many other non-culturable microbes for which E. coli is not a good indicator. Moreover, birds and other animals also make use of the waterpark. These can introduce E. coli back into the water, which gives a distorted picture of the removal of this faecal indicator.

To determine if the waterpark is capable of effectively removing microorganisms from the WWTP effluent, Waterboard De Dommel and Orvion chose an alternative approach. Instead of monitoring just a few culturable indicator bacteria, the entire microbiome was measured using ORVIdecode NGS analyses at various points in the waterpark and in the stream. This provided valuable insight into the effect of the natural post-treatment on the microbial composition of the WWTP effluent and the impact it has on the stream. Also, a first insight was obtained in the effect of the different steps of post-treatment on the microbial composition of the water. In addition, measurements on specific faecal indicator bacteria were performed with qPCR to further assess the source of encountered faecal contamination in the water (human/WWTP or other).

Waterpark Groote Beerze
Waterpark Groote Beerze consists of two parts, North and South. Both parts consist of eight compartments. A schematic overview of the water park is given in Figure 1, the compartments are briefly explained below.

Water tap

Figure 1. Overview of the waterpark with different compartments (1-8). Sampling sites are indicated with red dots.

Buffer/distribution ditch (1 & 2): is used for buffering and even distribution of the water to the settling pond. The retention time is 0.2 days (5 hours).

Flea Pond/UV Plateau (3 & 4): the effluent is cleared of settling particles (sedimentation), suspended solids and (some of) the bacteria in 1.8 days (43 hours). The large numbers of water fleas (daphnia) in the pond cause, among other things, a decrease in bacteria (biological filtration). The last six meters of the flea pond is shallower over the entire width (10-20 cm water depth), enabling UV irradiation to kill-off bacteria.

Reed ditch/pond with submerged aquatic plants (5 & 6): The ditch and pond form an ecological filter (through filter action, plant uptake and biodegradation) and through submerged aquatic plants there is oxygen rhythm (day/night O2 rhythm) in the water. The retention time is 0.8 days (19 hours).

Swamp forest with fast course (ditch) (7 & 8): This ditch is shallow and has a sandy bottom. UV irradiation may also kill-off bacteria here. Water only flows into the swamp forest during high rainwater inflow. The residence time in the rapid loop is 0.1 day (2 hours).

The total average hydraulic retention time (HRT) is about 3 days underdry weather conditions.

The following samples were taken in July 2020, at the locations indicated in Figure 1:
WWTP effluent: sample of the effluent. This sample contains the microorganisms to be removed in the waterpark.
Waterpark: the outflowing water from both North and South. These samples show the extent to which the microorganisms have been removed from the WWTP effluent. Two more samples were taken in waterpark South: at the end of the settling pond and after the pond with aquatic plants. These samples provide insight in the microbiome in the various compartments of the waterpark.
Stream de Groote Beerze: a sample from the Groote Beerze was taken upstream as a reference sample that is not influenced by the WWTP treatment plant. Furthermore, a sample was taken after the discharge point of South and a third after the discharge point of North (influence of both discharges). This provides a first insight into the extent to which the discharge has an impact on the microbial composition of the stream.

DNA analyses

To provide insight into the microbial composition ORVIdecode-NGS analyses were performed on the eight water samples. This method decodes all the DNA present in a sample using next generation sequencing (NGS). Algorithms are used to identify the decoded fragments of DNA using databases containing known microbial DNA sequences. In this way the biodiversity (microbial composition/microbiome) of the WWTP effluent, the waterpark and de Groote Beerze have been elucidated the first time. The data was been filtered and visualized in various ways to gain insight into the operation and effectiveness of the waterpark.

An ORVIdecode analysis provides insight into the relative numbers (percentages) of microorganisms in a sample. An increase in the proportion of a bacterial group does not necessarily mean that the absolute numbers have also increased (or vice versa). Therefore, to obtain information on absolute concentrations, qPCR analyses were also performed for ‘universal bacteria’ and ‘universal Archaea’. To gain further insight into the removal of faecal bacteria in the waterpark, additional qPCR tests were performed on the faecal indicator organisms E. coli and Bacteroides dorei (a faecal indicator more specific to human feces).

Bacterial counts of the water

Bacterial concentrations in the eight water samples were analysed by qPCR to determine the variance between the samples. The results show that the number of bacteria in the eight samples is relatively similar and differs by up a factor of 4.2: the WWTP effluent contained the lowest concentrations of bacteria at the time of sampling (2.6×106 genetic units/ml) and the sample after the settling pond/UV plateau contained the highest concentrations of bacteria (1.1×107 genetic units/ml). The North and South outflow water samples contained slightly lower numbers of bacteria than the upstream sample of the stream. This shows that natural stream water does not necessarily contain lower concentrations of bacteria; rather, the microbial composition determines the quality of the water.

The WWTP microbiome

Many different bacterial groups were found in the WWTP effluent with ORVIdecode-NGS. A number of these are dominant and distinctive for the WWTP effluent compared to the other seven samples. They have a clear origin from the treated sewage or from the treatment plant itself (the activated sludge). These include faecal bacterial groups such as Bacteroides and Faecalibacterium, but also the bacterial group Ca. Accumulibacter which plays a central role in the purification process in the WWTP activated sludge. The bacterial group Arcobacter has also been found to be dominant. It has recently become known that this potential pathogen is not effectively removed by a WWTP [2]. It is therefore valuable to assess whether the water park can improve this removal efficiency.

The results of the ORVIdecode analyses show that many of the bacterial groups dominantly present in the WWTP effluent have virtually disappeared after the water park. Such as Bacteroides, which makes up 7.4 percent of the microbial biodiversity in the WWTP effluent got down to 0.04 percent in the outflow water from North. Ca. Accumulibacter shows a similar trend. This specium constitutes 5.6 percent of the microbial biodiversity in the WWTP effluent and only 0.08 percent in the North outflow water (see Figure 2).

Water tap

Figure 2. Percentages of three bacterial groups in samples from the WWTP effluent, waterpark, and stream. Bacteroides is the most numerous intestinal bacterium in mammals and thus an important faecal indicator bacterium in water. Arcobacter is a potential pathogen that is not efficiently removed by WWTPs. The group Enterobactereriaceae is a collection of many known enteric bacteria such as E. coli, Citrobacter and Salmonella.

The relative abundance of two bacterial groups that are dominant (>1%) in the WWTP effluent hardly changed in the discharged water. This indicates that they may not be removed efficiently in the post-treatment process. For example, in the WWTP effluent, the proportion of Acinetobacter (2.0%) and Flavobacterium (3.5%) is similar to the proportion in the outflowing water from South (1.3% and 3.1%, respectively) and North (1.7% and 2.7%). Another notable bacterial group is Pseudomonas, which is dominant in the WWTP effluent (3.6%), but has proportionally increased in the discharged waters of South (9.2%) and North (19.3%). At this time, the extent to which species-level composition has changed within these bacterial groups has not been determined. Further exploration of the microbial data will have to reveal whether these bacterial groups are not being properly removed. Or whether the species composition did change significantly as a result of post-treatment. These species would then not originate from the WWTP effluent, but would be introduced or grow during treatment in the waterpark.

The waterpark microbiome
The waterpark is effective in far-reaching removal of microorganisms from sewers and WWTPs. But which bacterial groups take their place? It was determined which bacterial groups have become dominant (>1%) in the water park effluent compared to the WWTP effluent (Figure 3). This gives insight in which bacterial groups are growing in the waterpark. These include Stenotrophomonas and Polynucleobacter, which are mainly known from environmental ecosystems such as soil and freshwater [3], [4].
It is therefore likely that these bacterial groups have become dominant in the waterpark ecosystem. The microbial composition of waterpark North and South are broadly similar: the same groups have become dominant. This indicates that the natural post-treatment process is similar for both sections.

Water tap

Figure 3. Bacterial groups that have become dominant in the outflowing water as a result of post-treatment in the waterpark.

During sampling, it was noted that a bypass was in operation at South (orange arrow in Figure 1). As a result, a relatively small portion of the WWTP effluent skipped a large part of the treatment process (settling pond, UV plateau, reed ditch, and part of the aquatic plant pond) and passed almost directly to the end of the aquatic plant compartment.
The microbial data clearly show the effect of the short-circuit flow. The outflowing water from South has a less efficient removal than North for a number of WWTP microorganisms (Figure 4). In particular, the removal of Arcobacter seems to benefit from a longer residence time and the additional treatment steps. Its relative removal is 16 times higher in North than in South.

Water tap

Figure 4. Bacterial groups from WWTP effluent less effectively removed in South (with short-circuit flow) than in North. (n.d. = not detected).

This is confirmed by qPCR counts of the faecal indicator bacteria Bacteroides dorei and E. coli (Figure 5). The largest decrease in these numbers occurs after the settling pond and UV plateau. Because some of the water did not pass through these steps, but flowed directly to the end of the pond with aquatic plants, concentrations are higher in this sample. Therefore, in the outflowing water of South B. dorei was still found in increased numbers, while in the outflowing water of North it was no longer detected.

Water tap

Figure 5. Numbers of two faecal indicator bacteria in the different samples. (n.d. = not detected).


Not all available data can be discussed in this article. For example, it has been found that specific bacterial groups become dominant in the different compartments. Such as the groups Porphyrobacter, Pseudanabaena and Oscillatoria that become dominant in the settling pond, or Nitrospira in the aquatic plant pond. Such information can eventually be used to better understand the role of the various treatment steps at a microbial level and to support management and maintenance accordingly.

Impact on the stream ‘de Groote Beerze’

Ultimately, the waterpark should minimize the impact of the WWTP on the Groote Beerze. To assess how effective this is on a microbiological level, an upstream sample (no effect of the WWTP) was compared to a downstream sample after the discharge points of both  North and South. It was assessed whether the proportion of bacterial groups that are dominant in the WWTP effluent increased after the discharges (Figure 6).

The results show that a low background proportion of these bacterial groups are already present in the reference sample (0.01% to 0.09%). It can be seen that the proportion is slightly increased after the discharges, except for Faecalibacterium. The proportion of Arcobacter increases from 0.09% in the upstream reference sample to 0.2% in the downstream sample. Clearly, however, the increase would be  much higher should the WWTP effluent discharge directly onto the stream. The proportion of a number of other bacterial groups also increased after the discharges, such as Stenotrophomonas and Massilia (both primarily associated with plants [3], [5]). These bacterial groups originate from the natural environment of the waterpark and are therefore expected to have less (or no) negative impact on the stream when compared to the WWTP effluent.

Water tap

Figure 6. Left: shares of some bacterial groups in the Beerze River that are dominant (>1%) only in the WWTP effluent sample. Right: counts of faecal indicator bacteria. (n.d. = not detected).


The qPCR counts show that B. dorei is no longer detected downstream sample (Figure 6, right). This indicator is more specific for human faecal contamination than E. coli and is therefore less affected by faecal contamination from birds or other warm-blooded animals. This again shows that the impact of the WWTP effluent on the stream after further treatment in the water park is minimal.

Post-treatment in waterpark to prevent microbial contamination of the environment

Despite the efficient operation of the WWTP, the effluent still contains high concentrations of bacteria from the treated sewage, which is also the case with other WWTP. This is reflected, for example, in the high proportion of faecal bacteria and potential pathogens such as Arcobacter.

DNA techniques have shown that post-treatment of the WWTP effluent in the Groote Beerze constructed wetland waterpark significantly improves the microbial quality of the water. The microbial impact of the effluent on the receiving water, the stream ‘de Groote Beerze’, is hardly measurable. It is important to realize however that these measurements provide a snapshot. The effect of seasonal influences or other natural variations on the microbial removal efficiency has not yet been assessed.

The microbial composition of the water changes significantly during the post-treatment process. Especially bacterial groups associated with freshwater, plants and soil become dominant. Measurements in the different steps of the post-treatment have shown that the main removal of the WWTP microbiome takes place in the settling pond and the UV plateau.

The data clearly shows the effect of skipping a large part of the constructed wetland in waterpark South. This also demonstrates the importance of management and maintenance for proper functioning of the post-treatment system. The short-circuit flow has now been remedied and it would be valuable to assess whether the microbial treatment efficiency has improved as a result and the impact on the stream has been further reduced.

The data showed that Arcobacter in particular is removed less efficiently when the waterpark performs sub-optimally. A more routine monitoring of Arcobacter, B. dorei and ‘universal bacteria’, would therefore be useful to quantitatively assess how the natural post-purification performs and if/how it can be optimized.

Using DNA methods it is possible to assess the removal of antibiotic resistance genes (ARG) by the waterpark. WWTP effluents are ARG hotspots and contribute to an increase in ARG in the aquatic environment [6]. Minimizing this influx would help prevent the spread and development of antibiotic resistance. ARG can be exchanged among bacteria and therefore the removal of WWTP bacteria may not be directly related to the removal of ARG. Data on ARG in the eight samples is available in the generated ORVIdecode results and could be further investigated.

The analytical methods applied have provided the necessary data to make the effectiveness of the natural post-treatment of WWTP effluent understandable and measurable. The invisible world of underwater microbes has been made visible and with that the added value of waterpark Groote Beerze in improving the microbial water quality.


  1. Stichting Toegepast Onderzoek Waterbeheer (2012). Waterharmonica, onderzoek naar zwevende stof en pathogenen, deelstudierapport 4 (report in Dutch), deelstudierapport 4
  2. Kristensen, J.M., Nierychlo, M., Albertsen, M., Nielsen, P.H. (2020). ‘Bacteria from the Genus Arcobacter Are Abundant in Effluent from Wastewater Treatment Plants’. Environ. Microbiol. 86.
  3. Ryan, R.P. et al. (2009). ‘The versatility and adaptation of bacteria from the genus Stenotrophomonas’. Rev. Microbiol. 7, 514–525.
  4. Jezberová, J. et al. (2010). ‘Ubiquity of Polynucleobacter necessarius ssp. asymbioticus in lentic freshwater habitats of a heterogenous 2000 km2 area’. Microbiol. 12, 658–669.
  5. Ofek, , Hadar, Y., and Minz, D. (2012). ‘Ecology of Root Colonizing Massilia (Oxalobacteraceae)’. PLOS ONE 7, e40117.
  6. Sabri, N.A. et al. (2020). ‘Prevalence of antibiotics and antibiotic resistance genes in a wastewater effluent-receiving river in the Netherlands’. Environ. Chem. Eng. 8, 102245.

Comments are closed