The island of Texel in the western part of the Dutch Wadden Sea in north-western Europe is short of fresh water. Drinking water is transported through a pipeline from the mainland. The five sewage treatment plants on the island discharge to a relatively small body of surface water. This necessitated a higher quality of the effluents. Therefore, after preliminary tests on pilot scale, the waterboard has constructed a full-scale wetland system to improve the quality of the 3500-m³/day effluent from the sewage treatment plant Everstekoog.

The constructed wetland is a combination of a pond and ditches, with helophytes and submerged aquatic plants. After passing a presettling basin, the water flows through nine parallel ditches with a length of 150 m. The first half of each ditch is only 0.2 m deep and has vegetation with reed (Phragmites australis) or cattail (Typha latifolia). The deeper (0.5 m) part contains submerged aquatic plants. One ditch serves as a control without vegetation. The total surface of the pond and ditches is 1.3 ha, which corresponds with 0.5 – 1 m² per population equivalent.

The total hydraulic retention time (HRT) of the whole system was just over two days. In the second part of the project, the ditches had different HRT´s: between 0.3 and 10 days.

The specific costs are US $ 0.02 – 0.10 per m³ treated water depending on the price of land and the HRT. These costs are low compared with techniques as disinfection with UV, sand filtration and other tertiary treatment steps. The energy requirement of the wetland system is very low; the process "runs on solar energy".

The constructed wetland acts as a buffer area between the sewage treatment plant and the surface water. It fits well in the Dutch polder landscape. The main result is that the treated sewage is converted to a different kind of water, without activated sludge flocs from the aeration tank. The odour has disappeared. The most important indication for improvement of the water quality is the diurnal change in oxygen concentration. In the compartment with submerged plants, it becomes low at night, but very high, up to 20-30 mg/l, during daytime. The water starts to "live"; it becomes more like a natural surface water. Nitrification and denitrification depend very much on the hydraulic retention time. At longer HRT´s, we have found a substantial N-removal. In addition, disinfection is important in the constructed wetland. E.coli-levels were under 10/ml for most of the time in 1995-1996, and around 1 per ml at the longer HRT´s in 1997.

The research project taught us, that we should consider treated sewage as a source of good clear water suitable for various applications, instead of a burden. The origin of sewage is clean drinking water and rainwater. After converting the bulk of the pollutants into sludge, the effluent is clear water with a high potential for further use, because of its mineral content. Despite the high levels of nutrients, the water in the presettling basin in the wetland system remains clear. The level of chlorophyll-a stays well under 10 mg/l for most of the time, probably because of grazing by the usually high numbers of Daphnia magna. During the summer of 1998 tests were carried out, together with TNO, in order to understand more about the system and its processes. One of the goals is to use the nutrients in the effluent of the oxidation ditch to "produce" Daphnia to improve the food situation on Texel for fish and subsequently birds; like Spoonbills, which feed on small zooplanktivorous fish.

The waterboard "Uitwaterende Sluizen" operates over twenty sewage treatment plants (STP’s) in the north-west of the Netherlands. Five of them are located on the island of Texel (Figure 1). Since the early 1970’s the policy of the water board has been to keep the (fresh) effluent inside the surface water system of the island, rather than discharge it into the North Sea (Wadden Sea), because of the shortage of fresh water on the island in summer. The discharge of the effluent into the surface water on the island causes however particular problems, because of:

• The surface waters on Texel have a characteristic and diverse flora and fauna. Therefore more attention on the quality of the treated wastewater is needed;
• The volume of waterbodies on the island that receive the cleaned effluent is small. Any residual pollution in the effluent will hardly be diluted.

On Texel, a great variety of water types exists. From west to east, a salinity gradient in the surface waters results in many differences in plant and animal life. During summer, water losses from the surface through evaporation are replenished with groundwater, rainwater and effluent. The latter source can be substantial: in dry years up to 90% of the water in the surface water system is effluent. Drinking water is transported to the island through a main pipe, is the only external supply of fresh water.

During the last decades water treatment in the Netherlands has evolved to a high standard. From an ecological point of view however, the treated water is still "dead" water. Originally it is beautiful water; it has been rainwater and potable water. The bulk of the water is only used for rinsing and as a transport medium. In an activated sludge plant, the treatment process is a biological process, but it only involves organisms, common in a polluted environment. The result is an effluent with a rather low content of organics (most of the carbon pollutants have been converted in sludge) and a high level of minerals and nutrients. Most effluents in the Netherlands are very clear, with only a yellowish colour. Normally the levels of suspended solids are low, below 5 mg/l. Only after heavy rains it can contain up to 1000 mg sludge/l . This sludge can have a considerable impact near the discharge point. The sludge particles contain high numbers of human bacteria and viruses. Per population equivalent sewage contains 10¹⁰ to 10¹¹ E.Coli per day: or about 10⁸ to 10⁹ per ml. After biological treatment, the numbers diminish to 10³-10⁴ per ml. This make effluent unsuitable for swimming. Further odour (effluent smells), foam on the water surface and the sludge particles in the water may negatively effect the aquatic ecosystem receiving the effluent. Ideally, effluent quality should be as close as possible to that of the receiving surface water.

Many technological solutions exist to improve the quality of treated wastewater, for example sand filtration and membrane filtration. These processes are expensive, and above all the water is still not a "living water". We have been looking for a system between the treatment plant and the surface water to improve the "ecological" water quality. The basic idea is pictured in figure 2.

This model uses a "buffer" between the sewage treatment plant and the surface water. In this "buffer" the different fields of engineering and ecology meet each other to improve the quality of the STP effluent into more natural "living" surface water.

For Texel, a constructed wetland has been chosen to fulfil this task. The results of the pilot study of a constructed wetland treating 240 m³/d effluent from the Everstekoog STP on Texel from 1988 through 1992 has been described by Schreijer et al [in Haberl, 1997]. The positive results of this test have led to a full-scale constructed wetland to convert the effluent from the Everstekoog oxidation ditch into a more natural "living water". Some of the considerations were:

• The constructed wetland is a simple system like the oxidation ditch;
• It runs on solar energy "like surface water";
• There is enough space available;
• It fits in the rural Dutch landscape.

The four-year research project was a demonstration of the possibilities of a full-scale surface flow constructed wetland after an oxidation ditch. It should provide knowledge of:

• The separate processes in the constructed wetland;
• Give insight in maintenance, control and costs of the system. Does the combination with the constructed wetland lead to savings in the STP, for instance in the construction of settling tanks because the sludge will be retarded in the wetland or in nitrification/denitrification?;
• Is it a stable system, also in winter time or at high flows during rain?

This summer the field work of the Everstekoog project has been followed up by exploratory research to test whether the effluent could be used as a source of water to "produce" biomass in order to improve the food situation of fish and birds, especially Spoonbills, on the island of Texel in near future.

The sewage treatment plant Everstekoog is an oxidation ditch with a load of 45.000 p.e. in summer. The dry weather flow is 3000-4000 m³/day, the maximum flow is 10.000 m³/day. Phosphorus removal takes place simultaneously with FeSO₄.

The constructed wetland is a surface flow system constructed in 1994. The system consists of a presettling basin, nine parallel ditches with a length of 150 m and a discharge ditch. The first part of each ditch is only 0.2 m deep and has a vegetation with reed (Phragmites australis) or cattail (Typha latifolia). The deeper (0.5 m) part is planted with submerged aquatic plants. One ditch is a control without plants (Figure 3 and Figure 4). The total volume is 7140 m³ (Table 1). The mean hydraulic retention time (HRT) is 2.1 day at dry weather flow. In the first research period (1995-1996) all ditches received the same flow, in 1997/1998 four different flows through the ditches resulted in HRT´s of 1.6 up to 11.3 days (the retention time in the ditches alone were 0.3, 1, 3 and 10 days).

Constructed wetlands can have considerable values for aquatic birds (Knight, in Haberl 1997). Though the constructed wetland is situated in the agricultural part of the island, it attracts quite high numbers of birds. Spoonbills (Platalea leucorodia) come to feed on small fish. Table 2 contains the number of breeding birds in the wetland in 1997.

Pressure sensors in the presettling basin, the ditches and the discharge ditch measured the flows through the wetland. The instrumentation also included nine oxygen probes with thermometers, two redox sensors and a weather station. The results of the continuous measurements (15 minute averages) of water levels, oxygen concentration, temperature, wind speed and direction, precipitation and light intensity were stored in data loggers and transferred automatically to the Edam office.

Some aspects of the monitoring and research programme are:

• fortnightly routine water analyses (suspended solids, N, P, macro-ions and faecal coliforms);
• study of the contribution of the separate compartments on nutrient removal:
• storage in helophytes and periphyton;
• microbial conversions (nitrification/denitrification);
• accumulation in the soil (sedimentation, adsorption and precipitation processes);
• compilation of water and mass balances.

The preliminary test programme, carried out this summer to assess the possibilities to "produce Daphnia" has been carried at TNO, Department for Ecological Risk Studies by a student from the Vrije Universiteit van Amsterdam (VUA). Her work included laboratory and field studies at the Everstekoog plant for measuring growth of algae, Daphnia, food chain and toxicity studies.

The quality of the effluent of the STP Everstekoog is typical for a well functioning oxidation ditch (very low loaded activated sludge plant), results of 1997 - 1998 are summarised in table 3.

After treatment in the oxidation ditch, the sewage already has a much better quality. Sewage is grey, smelly. Effluent has become clear water, but still with an odour. Despite the removal of particles in the settling tank, the effluent still contains fine activated sludge particles, with a variety of bacteria. Already in the presettling basin in the constructed wetland the water starts to "live", it is going to resemble euthrophic surface water. It becomes a home for water life; regularly the water turns red because high numbers of waterfleas (mainly Daphnia magna). See also the recent publication in Water Research, describing growth tests of Daphnia in diary waste stabilization ponds (Kennedy, 1998). The number of different species of plants and animals in the wetland grew each year. The presettling basin also plays a role at storm water flows to catch the sludge losses from the STP.

It is difficult to quantify the changes in water quality. The suspended solid content of the effluent from the wetland is higher then of the STP effluent, but it is a different kind of suspended solids. In stead of activated sludge flocs the water contains algae, Daphnia and rests of water plants.

A first sign in the change of water quality is the diurnal oxygen fluctuation. (Figure 5). The oxygen level in the effluent of the oxidation ditch is stable and low, in the presettling basin the oxygen level is around 3 mg/l. In the part of the ditches with submerged aquatic plants the daily oxygen pattern starts to resemble the pattern of a normal surface water. During the daytime, the submerged aquatic plants and algae produce such an amount of oxygen that the levels do rise well above the saturation value. In the afternoon, the high oxygen levels will help the oxygen to penetrate deeper into the sediment. At the end of the day the oxygen levels drop sharply (Figure 6). On sunny days (nr. 178 and 179) the oxygen maxima are higher than on less sunny days (e.g. day 177) This "solar energy process" is also stable during longer periods. During the summer of 1997 the submerged water plants became covered by Duckweed (Lemna spp.) and floating algae. (Figure 7). During the summer, the oxygen production came to a complete stand still, but after the removal of the Lemna in the end of August the production started again quite quickly.

At the rather high loading in the first part of the project (HRT just over 2 days, hydraulic loading around 25 cm/day) both the nitrogen and the phosphorus removal were rather low. This occurred partly because of the already low levels in the effluent of the oxidation ditch. In the summer of 1996 NH₄ was were removed with 20 %, both NO₃ and PO₄ with 50 %. Other periods gave a negative P-removal through release from sediments and dying plant material. Even at low temperatures in the winter of 1995-96 the constructed wetland lowered the, temporarily, high level of 30 mg NH₄/l to 10 mg NH₄-N/l.

During the second part of the research program the ditches had different hydraulic retention times, the total HRT varied from 1.6 to 11.3 days. The HRT had a clear effect on both N and P removal (Figure 8, Figure 9 and Figure 10)

The literature figures for removal of P and N through harvesting of biomass vary. On an average the removal through standing crop, incl. the litter will be around 125 kg N/ha.year (63-220) and 15 kg P per ha/year (3-19) in the Netherlands [Toet, 1995]. This project will give more accurate figures as soon as the data are analysed. The load of N and P with the STP effluent to the constructed wetland is around 6800 kg N/year and 800 kg P/year. Denitrification is probably the most important pathway for N-removal. In general, denitrification in periphyton shoots of submerged helophytes was much higher than in the soil, due to diffusion limitation in the soil top layer.

Already the pilot-study in a small ditch like constructed wetland [Schreijer et al 1996] proved that the combination of ponds and helophytes was effective for disinfection of the effluent of the oxidation ditch.

In this pilot study the relation between E.Coli: (numbers per ml) and the hydraulic retention time (days) was10log E.Coli = -0.65*time + C. Table 4 gives the k-values in the different seasons during the last research year 1997/1998. Possibly, due to wildlife in the system, the E.Coli numbers were rarely below 1 per ml. For E.Coli values of less then 10/ml throughout the year, the HRT must be at least 4 days. For only disinfection to a level of E.coli 10/l , a HRT of 2 days will be sufficient. Two of the ditches had the same retention time (HRT of the complete system 2.1 days) during the whole research period: Figure 13. To minimise the influence of storm water flows it is important to buffer as much water in the system as possible, for a surface flow system this can be done by means of a clever design of the weirs.

One of the surprises in this research project is the augmentation of the turbidity of the water, especially at higher hydraulic retention times (Figure 11). An important feature of the presettling basin is that it acts as a buffer when during storm water flows the STP effluent contain high contents of dry solids. Retaining these high peaks in discharge of sludge will lead to savings in settling tanks.

The investment costs of the constructed wetland alone were less than US $ 250.000, excl. the extensive instrumentation for the research project. This leads to capital costs of US $ 25.000 per year. Maintenance and supervision costs are about US $ 25.000 per year. At a flow of 1.200.000 m³/year the specific cost are about US $ 0.05 per m³ at an HRT of 2 days and US $ 0.10 at a HRT of 4 days. To put this in perspective the costs to transport the waste water to the STP are estimated at US $ 0.10 per m³ and of the treatment of the waste water in the oxidation ditch, including sludge treatment US $ 0.50/m³.

The constructed wetland increased the water quality of the effluent from an oxidation ditch:

• The oxygen regime improved, the effluent became a clear diurnal pattern of oxygen as long as the sunlight can reach the submerged plants. This pattern also existed even in wintertime at low temperatures;
• The hydraulic retention time has a distinct effect on the fate the output of the wetland. Especially the removal of ammonium, nitrate and E.Coli becomes effective at longer HRT´s even in wintertime. The turbidly increases at longer HRT´s due to the production of biomass in the wetland.

An HRT of two days is sufficient for a reliant disinfection in summer, at a HRT of around 4 days the disinfection is also effective in wintertime. At this HRT, also a substantial removal of nitrogen has been found. A stable P-removal in all seasons needs either a much longer retention time or, in most cases more practical, chemical precipitation.

The Everstekoog constructed wetland proves the values and the possibilities of a simple semi-natural/constructed ecosystem for the improvement of effluents from sewage treatment plants. Currently the waterboard "Uitwaterende Sluizen" is studying for which of the other plants of the waterboard constructed wetland for this purpose is feasible. Another waterboard in the Netherlands, De Maaskant, is constructing an "Everstekoog" type wetland with a capacity of 35.000 m³/day.

An application of this type of wetland we are studying currently, is the combination of the discharge of polished effluent and a fish ladder. Near the STP De Cocksdorp, a siphon fish ladder has been constructed. Due to the high dikes, fish like Three Spined Stickelback (Gasterosteus acculeatus) and Eel (Anguila anguila) can not reach the fresh surface waters of the island anymore. To lure the fish to the fish ladder a flow of 350 m³/day fresh water is pumped in to the North Sea. When enough fish has gathered, the fish is transported by means of a siphon over the top of the dike into the surface waters of the island. A couple of problems exist:

• During drought the fresh water used for the lure flow is needed by the farmers;
• The STP De Cocksdorp, which is overloaded in summer, discharges its effluent close to the point where the fish is released.
• The fish ladder therefore functions not as good as possible and the fish’s spawning-bed will be polluted.

The basic idea is whether a surface constructed wetland can:

• improve the surface water quality near De Cocksdorp and;
• produce a water quality, good enough to be used as lure flow for the fish ladder and;
• be used to grow food for the Three Spined Sticklebacks that arrives on the island through the fishladder, thus improving the food situation of Spoonbills and other birds on the island.

This wetland could function as a trapped system where nutrients are turned into biomass (Figure 14). Algae as a food source for zooplankton will grow on nutrient rich effluent. The zooplankton is grazing the algae and they serves as a food source for the Stickleback.

During a six-month pilot study, the possibility of using effluent from a wastewater treatment plant for the improvement of the food situation of Sticklebacks and Spoonbills was studied. Some pictures of the experiments.

Questions that needed to be answered were:

• Are there restrictions to cultivate algae and zooplankton on effluent due to water quality problems (e.g. oxygen levels, pH etc.)?
• Is it possible to breed zooplankton on organic matter from the STP (active sludge)?
• Is it possible to maintain a breeding process in a surface flow system on effluent?
• What is the risk of biomagnification of copper and zinc in the food chain from zooplankton to Spoonbills?

The growth of Chlorella pyrenoidosa on effluent of STP De Cocksdorp and STP Everstekoog and OECD was followed for a 20 day period in duplicate systems. Nutrients were added to avoid nutrient limitation in order to allow for the assessment of toxic effects.

C. pyrenoidosa is able to grow on effluent, but the growth rate is lower than in the algal growth medium (OECD) (Figure 15). The inhibition of algal growth may result from unfavourable water quality conditions. Conclusion is therefore that low algal numbers in the presettling basin is not only due to grazing.

D. magna was cultured on effluents from STP De Cocksdorp, STP Everstekoog and a growth medium (OECD) to compare its performance (determined by survival and reproduction).

In most systems D. magna numbers increased (from the initial 20 individuals at the start) as a result of reproduction (Figure 16). Although reproduction was observed in the effluents, reproduction was hampered in comparison with the standard medium. Growth on effluent of STP De Cocksdorp was more limited than growth on effluent of STP Everstekoog.

From an experiment with plankton enclosures (plastic bags) in the presettling basin, supplied with effluent it appeared that zooplankton grows faster when the concentration of activated sludge increased. Therefore, it can be concluded that D. magna is able to grow on sludge particles (activated sludge) in the effluent of the oxidation ditch.

When it became clear that phytoplankton and zooplankton where able to grow on effluent, research was done to find out if it was possible to cultivate phytoplankton and zooplankton in a flow-through system (Figure 17)

Originally, the effluent consists of rain and drinking water. Because this water flows through either copper drinking water pipes and zinc gutters, the effluent may contain higher copper and zinc concentrations. Therefore, biomagnification of copper and zinc in the system may cause a risk. Both copper and zinc concentrations in Daphnia, grown in effluent were not higher than usual in Dutch surface water. The risk of biomagnification of copper and zinc was studied by feeding Three Spined Sticklebacks for a three week period with zooplankton caught from different locations; the presettling basin, a field location, and from a culture. After three weeks, the Sticklebacks will be analysed on copper and zinc concentrations. Results are not available yet.

• Cultivation of phytoplankton and zooplankton on effluent of STP De Cocksdorp en STP Everstekoog was possible at suboptimal rates. A difference in growth-rate between the STP effluents was observed.
• Cultivation of zooplankton on algae in a flow-through system is possible as long as the algal concentration can be kept above 40 µg chlorophyll-a/L. Zooplankton can be cultivated on active sludge instead of algae.

A surface flow constructed wetland, like the Everstekoog system, is a simple attractive system. It is also cheap as long as land costs are not too high. It looks like a Dutch polder landscape; the maintenance of the system resembles the maintenance of our ditches and canals. Near populated areas one of the features of the constructed wetland is that it can be combined with recreation areas and wildlife.

A combination of open ponds, helophytes and submerged aquatic plants is a cost-effective way to change sewage in to a "living" water suitable for various purposes. The use of the minerals in effluent of STP´s to "produce food" to improve the food situation on Texel of fish and subsequently birds, which feed on small zooplanktivorous fish, like Spoonbills seems feasible. A trapped ‘food-chain type’ system seems a promising option to increase the ecological value of effluents from oxidation ditches.