wetpol 8 - Ecole des mines de Nantes

52 downloads 2142 Views 2MB Size Report
17 Oct 2013 ... waters that are exported to a main collecting ditch called Le Ru des Gouffres before to sink ... sluice gate has been installed at the inlet of three of them and ..... Métropole not only as client, but also as a main stakeholder and ...
Wetland's role in pollutant management at the catchment scale

ABSTRACTS - WETPOL 2013 - October 13-17, 2013 - Nantes - FRANCE 211

Evaluation of agricultural pollutants (pesticides and nitrate) removal performance in a constructed wetland (O.50) Roux Laetitiaa, Tournebize Juliena, Chaumont Cédrica, Hocine Héninea, Ginzburg Irinaa a

Hydrosystems and Bioprocesses Research Unit - IRSTEA, 1 rue Pierre Gilles de Gennes, Antony, 92761, FRANCE ([email protected])

INTRODUCTION The European Water Framework Directive (EWFD) enjoins the EU member states to protect aquatic ecosystems so that their rivers and ground water achieve good chemical and biological status by 2015. The implementation of this directive involves both technical solutions and incentives for changing social practices. For the technical part, previous research showed the potential of constructed wetlands as a management practice for pesticides removal (Braskerud & Haarstad, 2003; Maillard et al., 2011; Passeport, 2010). The Rampillon’s wetland has been monitored by the team TAPAHS of IRSTEA during 2 years. The quality of water has been analysed for nitrate and pesticides at the inlet and outlet of the wetland to evaluate the performance of the system. METHODS Description of the catchment The catchment is located in the city of Rampillon (03°03’37.3’’ E, 48°32’16.7’’ N, 70 km south-east of Paris, France). The 390 ha agricultural watershed receives an annual mean rainfall of 689 mm. The average annual air temperature is 10.5°C, and the annual mean potential evapotranspiration is 679 mm (Tournebize et al., 2012). Sinkholes are present at the outlet of the watershed. They allow the direct infiltration of 800 000 m3/year in the Champigny aquifer. This amount corresponds at the yearly water consumption of 15 000 inhabitants. Farmers mainly grow winter wheat, sugar beet, corn, beans and rape. Pesticides are for the most part applied in November, March, April, May and June. The whole catchment has been artificially drained by 90-cm-deep and 10-m-spaced subsurface drains which cover about 450 ha instead of the 390 ha of the natural watershed. Rainfall events generate pipe drainage waters that are exported to a main collecting ditch called Le Ru des Gouffres before to sink into the aquifer. Innovative strategy for water pollutant management On the whole agricultural catchment, four wetlands have been constructed in 2010. A sluice gate has been installed at the inlet of three of them and allows the closing during winter time (December – March) when no pesticides are applied and rainfall events are more frequent. This innovative solution permits to choice the period of the year when the drained water is the most polluted. The entire volume of these four wetlands is based on a measured spring flood of 9000 m3. The aim was to obtain a volume of 7 m3 per drained ha. The technical (topography) and human (farmer acceptability) constraints do not allow to reach this objective. The three smaller are managed by the farmers themselves. They are connected to about 100 ha for two of them and 15 ha for the latter one. A collective wetland of 3156 m² has been constructed at the outlet of the catchment and equipment has been installed in December 2011 to allow the evaluation of the system performance. The depth varies from 0.1 m to 1 m. The wetland receives the drained water of the whole drained catchment, namely 450 ha. The inlet and outlet pipes have a diameter of ABSTRACTS - WETPOL 2013 - October 13-17, 2013 - Nantes - FRANCE 212

500 mm. The flow which enters the wetland varies between 0 and 120 l/s. The wetland is partially covered with vegetation, mainly Typha latifolia, Phragmites vulgaris and Juncus conglomeratus. Equipment Analysis of water samples Permanent monitoring equipment has been installed at five locations: 1. inside the inlet pipe, 2. inside the outlet pipe, 3. in the wetland next to the outlet, 4. in the ditch, 5. beside the inlet of the wetland. The measurement station in the ditch was installed downstream from the outlet of the wetland, using a wooden controlled section adapted to natural ditch cross section (Birgand et al., 2010). Three height and velocity Doppler sensors were set up at the inlet, outlet and in the ditch. Quality was measured using three automated water samplers at the same preceding locations. Water samples of 100 mL were collected in a 10 L glass jar every n*60 m3, with n an integer which depends on the season. If we expected high rain events, n could be equal to 12. The composite water samples were collected every two weeks on average and were placed on ice waiting for future laboratory analysis. Water samples have been then sent to an external laboratory to determine pesticide concentrations in water. About 70 molecules are analysed with 10 different analytical methods. Ammonia (NO3-) concentrations have been also recorded every 30 minutes at the inlet and the outlet with two spectrometer probes. A rain gauge measured the rainfall next to the inlet every 2 mm. In addition, water level and temperature measurement device was installed inside the wetland, next to the outlet. RESULTS AND CONCLUSIONS The laboratory analyses of pesticides are in process, as well as the analysis of the nitrate data. The comparison between the concentration of the pollutant at the inlet and the concentration of pollutant at the outlet, will allow the evaluation of the wetland mitigation performance. The results will also serve to the calibration and the validation of a numerical model, which is under construction. ACKNOWLEDGEMENTS This work is mainly supported by IRSTEA and l’Agence de l’Eau Seine-Normandie. We would like to thank the AQUI’Brie association for their involvement in the project. REFERENCES Braskerud, B. C. and Haarstad, K. (2003). Screening the retention of thirteen pesticides in a small constructed wetland. Water Science & Technology, 48:267–274 Maillard, E., et al. (2011). Removal of pesticide mixtures in a stormwater wetland collecting runoff from a vineyard catchment. Science of the Total Environment, 409:2317–2324 Passeport, E. (2010). Efficacité d’une zone humide artificielle et d’une zone tampon forestière pour dissiper la pollution par les pesticides dans un bassin versant agricole drainé.AgroParisTech Tournebize, J., et al. (2012). Co-design of constructed wetlands to mitigate pesticide pollution in a drained catch-Basin: A solution to improve groundwater quality. Irrigation and Drainage, 61:75–86 Birgand, F., et al. (2010). Guide d’application de la technique du Doppler continu pour une section calibrée en bois.

ABSTRACTS - WETPOL 2013 - October 13-17, 2013 - Nantes - FRANCE 213

Floating marshes for heavy metal removal in highway storm water pond (O.88) S. Ladislasa, C. Gérentea, F. Chazarenca, J. Brissonb and Y. Andrèsa a

L’UNAM Université, Ecole des Mines de Nantes, CNRS, GEPEA, UMR 6144, 4 rue Alfred Kastler B.P. 20722 F-44307 Nantes Cedex 3, France. (Email: [email protected])

b

Institut de Recherche en Biologie Végétale ; Département de sciences biologiques, Université de Montréal ; 4101 Sherbrooke Est, Montréal, QC H1X 2B2, Canada

INTRODUCTION Road/highway stormwater runoff has been considered as a major source of contaminants and one of the most frequent causes of surface water pollution. The heavy traffic flows have provided numerous sources of contaminants, such as motor vehicle emissions, drips of crankcase oil, vehicle tyre wear, asphalt road surfaces, etc. The contaminants are deposited on the road surface, and during rainfalls, they are carried by the road stormwater stream, flowing ultimately to rivers, lakes and oceans. Among these contaminants in highway stormwater, heavy metals and polycyclic aromatic hydrocarbons are two significant pollutants of concerns due to their high toxicity to the ecosystem and their persistence in the environment. Floating marshes can be considered as effective systems for the treatment of the dissolved metal fraction present in the urban runoff but only few studies are interested in fullscale conditions (Headley 2006). This work is intended to provide new data to: - verify the technical feasibility of an implementation of floating marsh directly to the surface retention basins runoff, - assess the accumulation of metals (Ni, Cd and Zn) in real conditions by the two plants, Juncus effusus and Carex riparia, - measure the microbiological activity and metal accumulation in the root biofilm. METHODS The studied site is a detention pond located in the northeast of Nantes, in France. It receives the storm water runoff coming from a highway section (27 000 vehicles/day, basin surface 375 m2). Three floating marshes have been installed on the detention pond and each raft has a surface area of 1.5 m2 (125 cm x 120 cm) and contains 16 plants (Figure 1). After 4 months, plant samples were harvested, mineralized and analyzed by atomic spectrometry absorption to determine cadmium, nickel and zinc concentrations.

Figure 1: Floating marshes scheme

ABSTRACTS - WETPOL 2013 - October 13-17, 2013 - Nantes - FRANCE 214

RESULTS AND DISCUSSION The Ni concentration in the leaves is between 23 and 31 µg/g dry matter and between 113 and 131 µg/g in the roots. For Zn, the results show an accumulation of 45 to 80 µg/g in the leaves and 168 to 210 µg/g in the roots. The root/leaves ratio is between 2,6 et 5,7 for Ni and Zn, highlighting the important function of roots in heavy metal accumulation. Regarding cadmium, no accumulation has been detected in plants (Table 1). These metals have also been measured in the biofilm present onto the roots as well as the microbial respiration activity. Moreover, we observe that the roots present a particulates filtration capacity

Table 1: Metal accumulation in aerial part (PA), roots (R) and translocation factors (FT) for Juncus and Carex plant.

The root biofilms study in test experimental conditions showed that the microbes of root biofilms can easily degrade one benzene ring aromatic hydrocarbon benzoic acid and toluene (in 5 days, 10.1 mg benzoic acid was degraded 93%; in 15 days, 40 mg benzoic acid, 10mg toluene and 40.4mg toluene was degraded 83%, 84% and 64%, respectively), whereas they cannot degrade two-ring PAH (the derivative of naphthalene tetralin), three-ring PAH (fluoranthene) and four-ring PAH (pyrene). Not appropriate PAHs concentrations used in the BOD tests and the relative short duration conducted might be the main reason why no PAHs degradation was found. CONCLUSIONS This study shows the feasibility of floating marsh for the metal removal of storm water pond. ACKNOWLEDGEMENTS This research was conducted within the regional and cooperative POLESUR project (POLlution des Eaux et des Sols en milieu URbain), co-funded by the region of Pays de la Loire (France). The authors want to thank Tom Headley for his advices, and the technical staff of Ecole des Mines de Nantes (France) for its technical support. REFERENCES Headley T. (2006). Application of Floating Wetlands for Enhanced Stormwater Treatment: A review, ARC Technical Publication No. 324, Auckland Regional Council, Auckland.

ABSTRACTS - WETPOL 2013 - October 13-17, 2013 - Nantes - FRANCE 215

Efficiency of Hybrid Subsurface Constructed Wetland for the Treatment of Mixture of Point and Non-point Source Load (O.111) a

Takashi Inouea, Hiroaki Sakuragia, Kunihiko Katob, June Haradaa, Wu Daa

Graduate School of Agriculture, Hokkaido University, N9W9, Kita-ku, Sapporo, Hokkaido, 060-8589, JAPAN ([email protected]) b NARO Tohoku Agricultural Research Center, Shimo-Kuriyagawa, Morioka, Iwate, 0200198, JAPAN ([email protected]) INTRODUCTION Hybrid subsurface flow constructed wetland (CW) system has been built for a treatment of mixture of point and non-point source load in the Town of Toyotomi, Hokkaido, northern Japan (N45.112, E141.703). Within the total catchment area of 10.3 ha, influents consist of wastewater from a small-scale single dairy farm and its barn (8.4% of total area), treated water from the wetland visitor’s observatory and its rest facility, and the parking lot (9.7%). Part of the runoff water from the asphalt-paved road (2.4%), grassland (43.0%) and forest (36.5%) are also treated by this system under normal flow conditions. The purpose of building this hybrid CW system is to reduce anthropogenic impact on the wetland ecosystem located downstream of the same watercourse. The system started its operation in April 2010. This paper describes the outline of the system and its efficiency on water purification. METHODS System outline The hybrid CW system consists of two vertical flow beds (VF1 and VF2) each equipped with auto-priming siphon for intermittent water dousing, and two horizontal flow beds (HF1 and HF2). The total area of the system is 470 m2 with 146 m2 for VF1, 87 m2 for VF2, 147 m2 for HF1 and 90 m2 for HF2. The sequence of water flow is shown in Fig. 1. Water from the barn, road, grassland and forest Fig. 1. Schematic diagram of water flow and come through the roadside ditch (MA) and sampling point enters the pump pit under normal weather conditions. Under high flow conditions, inflow to the system is limited by pump’s setting (on/off = 15/45 min.). This limit (peak-cut) avoids excess pump running. The concentration of various water quality parameters during high flow condition is rather low because of dilution thus the purification is not in required. The treated water from the septic tank of visitor’s observatory (MB) also flows into same pump pit. Both water from the roadside ditch and the septic tank flow into VF1 through siphon pit (M1), and then to HF1 (M2), HF2 (M3), and to VF2 through another siphon pit (M4). The drainage water from the parking lot as well as part of wastewater from the rest facility (MP) joins into the system at the downstream of HF2. The final effluent (M5) is released back to the roadside ditch. Measurement Water samples were collected and analysed almost once a month from April 2011 to Dec. 2012. Continuous flow volume through the system was calculated mainly from the amount of siphon movement in M1 and M4. Amount of rainfall and evapotranspiration was also taken into account of the flow volume. The volume of through-flow by roadside ditch (the amount of water that did not flow into the system) was also measured by a rectangular weir. ABSTRACTS - WETPOL 2013 - October 13-17, 2013 - Nantes - FRANCE 216

RESULTS AND DISCUSSION Volumetric inflow during the monitoring period was 0 to 267.3 m3·d-1 and last year’s average inflow was 44.4 m3·d-1. Water temperature at the outlet of the system was higher than 0 ºC even in the coldest time of the year, which suggests that this system is capable to operate throughout the year. Average concentrations of water quality parameters of the influent and effluent are shown in Table 1 along with the purification rate (PR). The average load and removal rate (RR) are also calculated. Despite the low value in average influent BOD5 (4.1 mg·L-1) the purification efficiency was still quite high (1.2 mg·L-1, PR=72%). The removed load of BOD5 was 0.29 g·m-2·d-1. When this is compared to the total influent load of 0.49 g·m-2·d-1, the RR became 59%. On the other hand, PR and RR for COD(Cr) were rather low (PR=37%, RR=26%). Reason for this low efficiency seems to be caused by high content of humic substances in the water from the forest. Suspended solid (SS) showed good decrease in both PR (94%) and RR (89%). PR and RR of total nitrogen (TN) were low but the concentration of influent itself is very low so its impact on the environment is negligible. Table 1. Average concentrations at each sampling point and their associated purification rates (PR). Influent

Influent

Influent

after HF1

Influent

after HF2

Effluent

* (mg·L-1) ** (#·mL-1)

after VH1

Purification Ratio

MA

MB

M1

M2

M3

MP

M4

M5

(%)

COD(Cr) *

80.4

29.7

74.7

69.4

60.3

62.7

47.0

45.8

39

BOD5 *

4.5

2.6

4.2

2.8

2.3

2.5

1.5

1.2

72

SS *

46.6

20.8

47.7

27.0

29.4

10.5

4.8

2.9

94

T-N *

2.0

3.2

1.4

1.7

1.8

0.9

1.3

1.2

16

NH4-N *

0.22

0.33

0.35

0.14

0.07

0.22

0.05

0.06

82

T-P *

0.38

0.13

0.24

0.19

0.13

0.02

0.06

0.05

81

PO4-P *

0.12

0.05

0.11

0.07

0.05

0.01

0.03

0.02

79

Total coli. **

12

303

31

17

13

16

20

9

70

DO *

2.0

4.7

5.9

7.3

5.0

4.1

6.8

9.6

pH

6.8

7.2

6.9

6.9

6.6

7.4

6.7

6.4

CONCLUSIONS This hybrid CW system achieved good performance both on reduction of concentrations and loads, even in the large range of flow rate. Due to concern on the consumption of electricity, the limitation has been set for the operation of this system. However, through this study, it became clear that the operation time could be doubled throughout the year, which enhance the efficacy of this system. ACKNOWLEDGEMENTS The authors of this article would like to acknowledge the staff of Rishi Rebun Sarobetsu National Park. Part of this study was supported by the research program granted by the Ministry of Agriculture, Forestry and Fishery of Japan.

ABSTRACTS - WETPOL 2013 - October 13-17, 2013 - Nantes - FRANCE 217

Constructed wetlands for urban runoff treatment from a residential watershed (O.114) a

N. Duclosa, A. Wankoa, J. Laurenta, M. Fischera, P. Molleb, R. Moséa

Icube, ENGEES, INSA, Université de Strasbourg, CNRS ; 2 rue Boussingault 67000 Strasbourg, FRANCE ([email protected]) b Irstea de Lyon, 5 rue de la Doua, CS70077, 69626 Villeurbanne Cedex, FRANCE INTRODUCTION Urban stormwaters are polluted by different substances [Lee & Bang, 2000; Zgheib et al., 2012]. Moreover, the strong variability of the effluents coming from the separated sewer network is a problem for an optimal design of stormwater treatment systems. This work aims to determine the treatment efficiency of a treatment system located downstream a separated sewer network from a residential watershed. This treatment system is made up with a sedimentation pond and a vertical constructed wetland. This abstract presents the methodology chosen, as well as the first results concerning hydrodynamic and water quality. METHODS The treatment system is presented in Fig. 1. The sedimentation pond has a volume varying between 2 and 10m3. The active layer of the constructed wetland is 30cm deep and composed of sand (0 to 4mm size, d10 = 0.16mm and uniformity coefficient = 8.62) with a saturated hydraulic conductivity equal to 2.2.10-4m.s-1. It has been planted with 2 years old potted Phragmites Australis in July 2012, at a density of 9 per m². A ventilation pipe located under the sand layer and a bottom drain provide passive aeration to favour the oxygenation of the media. Finally, a water reserve at the bottom of the filter prevents water stress of plants. The hydraulic load of the constructed wetland is equal to 30m3.m-2.year-1. It is capable of treating a raining event with a return period up to 2 years.

Fig. 1. The treatment system (in black, the elements of the system, in grey, the different probes)

The probes shown on Figure 1 allow precise monitoring of main physic-chemical parameters as well as the hydraulic behaviour of the system. Water samples are retrieved during raining events: the samplers are steered by the beginning of the raining event; water is then time-proportionally sampled. The composite sample is then reconstituted thanks to the hydrograph. Water physic-chemical parameters are then analysed. ABSTRACTS - WETPOL 2013 - October 13-17, 2013 - Nantes - FRANCE 218

RESULTS AND DISCUSSION The first treatment efficiency results are shown in Table 1. The corresponding raining event (return period: 3 to 6 months) occurred after a dry period of 15 days. Concentration of TSS

Concentration of BOD5

Concentration of carbon

15

60 40

TSS

30

10

20

5

20

1

2

0 1

3

Concentration of N

2

3

1

Concentration of Cu

4 N-NO3

2

N-NO2

1

N-NH4

0

12 8 4 0

1

2

3

2

3

Concentration of Pb 30

16 Organic N

3

Diss. OC

10

0

0

Part. OC

BOD5

Diss.Cu

20

Part. Cu

10

Part. Pb

0 1

2

3

1

2

3

Table 1. The first quality results - campaign of the 26/04/2013 (1: upstream, 2: intermediate, 3: downstream; values in mg.L-1 for TSS, BOD5, carbon and nitrogen, in µg.L-1 for Cu and Pb)

The concentrations of organic matter at each point of the treatment system are quite low. In terms of TSS and VSS, the results show that both the sedimentation pond and the constructed wetland are efficient. Concerning nitrogen, nitrates contribute deeply to the global nitrogen present. The sedimentation pond doesn’t contribute to the removal of nitrogen, contrary to the constructed wetland. This might be explained by the uptake of the reeds of plants during the growing season. Moreover, few metals are present in the influent, except for copper and lead. Copper is essentially in dissolved form. The difference between raw and dissolved samples can be explained due to the uncertainty of sampling and conditioning. For copper, the treatment is mainly attributed to the constructed wetland. Lead is retained by both sedimentation pond and constructed wetland. CONCLUSION The results show that the sedimentation pond and the constructed wetland both contribute to the treatment of the effluents, depending on the parameter considered. Further monitoring campaigns are still carried on. Treatment performances will be assessed as a function of the rain events types, the occurrence of long dry periods, etc. To determine optimal sizing of such treatment systems for the treatment of urban runoffs, the treatment efficiency is going to be compared to the ones of two other treatment systems set up in two other similar watershed. ACKNOLEDGEMENTS The authors would like to thank the Urban Community of Strasbourg, which provided the experimental installations, and the laboratory technician Marie-Pierre Ottermatte for her help in analysing the samples. REFERENCES Lee, J.H. and Bang, K.W. (2000) Characterization of urban stormwater runoff. Water Research, 34(6):17731780. Zgheib, S., Moilleron, R. and Chebbo, G. (2012) Priority pollutants in urban stormwater: Part 1 – Case of separate storm sewers. Water Research, 46(20): 6683-6692.

ABSTRACTS - WETPOL 2013 - October 13-17, 2013 - Nantes - FRANCE 219

Phosphorus and particle retention in constructed wetlands – a catchment comparison (O.118) Karin M. Johannesson1, Pia Kynkäänniemi2, Barbro Ulén2 and Karin S. Tonderski1 1

Department of Physics, Chemistry and Biology, Linköping University, 581 83 Linköping, SWEDEN ([email protected][email protected]) 2 Department of Soil and Environment, Swedish University of Agricultural Sciences, 750 07 Uppsala, SWEDEN ([email protected][email protected]) INTRODUCTION Constructed wetlands are used in the agricultural landscape as a measure to reduce the transport of both nitrogen (N) and phosphorus (P) to the sea, but the efficiency of wetlands as P traps in Swedish conditions has not been comprehensively investigated (Bergström et al., 2007). Studies of in- and out flowing masses of P in wetlands receiving agricultural runoff from different soil types often show a positive retention (e.g. Braskerud et al., 2005 and Koskiaho et al., 2003), but the variation is large. In this study, seven constructed wetlands, situated in agricultural areas in the south of Sweden, were investigated for sediment accumulation and P retention and factors affecting the efficiency. These wetlands were situated in areas with various proportions of clay soils, hypothetically resulting in various loads of particles and P to the wetlands. National monitoring has shown that agricultural areas dominated by such clay soils have among the highest P losses (Ulén, 2004). It has also been shown that a large proportion of P is transported in particulate form; hence, it is expected that sedimentation is the predominant P retention process in the selected wetlands. Sedimentation of clay may, however, be difficult to achieve in wetlands, and the aim of the study is to quantify the function of wetlands as sinks for the P lost from the catchments. METHODS Sedimentation and accumulation of particles were measured once a year for two years using sedimentation plates (0.4×0.4 and 0.25×0.25 m) placed on the bottom of the wetlands in transects perpendicular to the water flow. P and soil retention was estimated by extrapolating the amount of sediment accumulated on the plates, and the content of total phosphorus (TP), to the whole wetland area using ArcGIS 10.1. This paper presents annual P and soil retention after two years investigation. Hydraulic loads were estimated using discharge data from the Swedish Meteorological and Hydrological Institute (SMHI), which have modeled runoff from every large catchment in Sweden. P loads were estimated based on leakage from arable land in each leaching region using the ICECREAM model (SLU, 2007). P retention data will be used together with wetland and catchment characteristics in a principle component analysis (PCA) in order to identify if some of the collected data on catchment and wetland characteristics may explain the difference in P retention between these wetlands. The factors included were the ratio of wetland area to catchment area, modelled hydraulic and P load, and various catchment characteristics, e.g. soil type, average slope in catchment and land use.

ABSTRACTS - WETPOL 2013 - October 13-17, 2013 - Nantes - FRANCE 220

RESULTS AND DISCUSSION In all wetlands there was an annual net settling of particles on the plates, but there was a large variation, both between wetlands and between years. Sediment retention varied between 12 and 200 tons ha-1 yr-1. P retention also varied (between 25 and 200 kg ha-1 yr-1), as did the modelled P load (between 8 and 700 kg ha-1 yr-1). The difference between the wetlands was considerably larger than the difference between the two years but no relationship between modelled hydraulic load and the observed particle or P retention could be observed. This highlights the difficulties in using predictions of hydrological loads and nutrient losses from a large catchment area for a much smaller catchment. In some wetlands there was a clear relationship between amount of sediment settled and distance from inlet, with a significant decrease from inlet to outlet. Most of the wetlands with this sedimentation pattern were long and narrow, whereas no such distinct pattern could be observed in wetlands with a lower length/width ratio. Previous studies on how hydraulic conditions affect pond performance have shown that wetlands that are long and narrow have higher hydraulic efficiency compared to a wetland with a round shape (Persson and Wittgren, 2003). Preliminary results from the analyses of various catchment and wetland characteristics and their possible effect on sediment and P retention show that there are no factors in particular that effect retention efficiency, but further statistical analyses are needed. CONCLUSIONS In agricultural areas in Sweden, constructed wetlands do function as traps for sediment and P, but there is a large variation between wetlands. Preliminary findings suggest that the design of the wetland may be more important than its placement in the catchment. In order to obtain maximum P retention, one has to facilitate sedimentation in the wetland itself. REFERENCES Bergström, L., Djodjic, F., Kirchmann, H., Nilsson, I. and Ulén, B. (2007) Phosphorus from farmland to water. Report Food 21, no. 4. Swedish University of Agricultural Sciences, Uppsala, Sweden. Braskerud, B.C., Tonderski, K.S., Wedding, B., Bakke, R., Blankenberg, A.G., Ulén, B. and Koskiaho, J. (2005). Can constructed wetlands reduce the diffuse phosphorus loads to eutrophic water in cold temperate regions? J. Env. Qual. 34: 2145-2155. Koskiaho, J. Ekholm, P., Räty, M., Riihimäki, J. and Puustinen, M. (2003) Retaining agricultural nutrients in constructed wetlands - experiences under boreal conditions. Ecol. Eng. 20: 89-103. Persson, J. and Wittgren, H.B. (2003) How hydrological and hydraulic conditions affect performance of ponds. Ecol. Eng. 21: 259–269. SLU (2007) Leaching coefficients and results from calculations of the normal leakage of phosphorus. Swedish University of Agricultural Sciences. Division of Water Quality Management. Uppsala, 2007-12-19 Ulén, B. (2004) Size and settling velocities of phosphorus-containing particles in water from agricultural drains. Water, Air and Soil Pollution 157: 331-343.

ABSTRACTS - WETPOL 2013 - October 13-17, 2013 - Nantes - FRANCE 221

AZHUREV: A Constructed Wetland for Water Treatment and Biodiversity in Reims (France) (O.150) Etienne MORELa, Grégoire JOSTa, Dirk ESSERb a

SINBIO, 5 rue des Tulipes, MUTTERSHOLTZ, 67600, FRANCE ([email protected])

b

SINT, Chef-Lieu, 73370 La Chapelle du Mont du Chat, FRANCE ([email protected])

INTRODUCTION AZHUREV originated from a call for innovative demonstration projects in the field of ecological engineering, issued by the French Ministry of the Environment in July 2011, as part of the implementation the French national strategy for diversity 2011-2020. SINBIO and Council of Greater Reims (Reims Métropole), grouped with 3 other partners, proposed to construct a demonstration wetland of 5 to 10 ha to treat excess storm water from the inflow and to polish treated sewage at the outflow of the new Reims metropolitan waste water treatment plant (400 000 p.e.) on lands formerly used for landspreading of sewage and sludge, before discharge into a small creek. This proposal was one of three projects which were selected by the ministry in the category “treatment wetlands” (“phytoépuration”). The aim of the project is to demonstrate that FWS constructed wetlands can enhance water quality and biodiversity at low investment costs, by adapting, as far as possible, the design to existing topography and soil characteristics. If successful, the project will be upscaled to cover most of the 160 ha of available land area, thus allowing for treatment the total flow of storm water and effluent from the treatment plant. METHODS The multifunctional objectives of the project are reflected by an interdisciplinary design team and design approach. The team consists of SINBIO, the project leader, one of the leading consultants in France in the field of restoration of rivers and lakes as well as in the field of sustainable urban drainage and natural waste water treatment systems, Reims Métropole not only as client, but also as a main stakeholder and technical partner of the team, ESOPE as naturalists and specialists of biodiversity, METIS as sociologist working on questions of social acceptability and public access policy, and finally the ‘Laboratoire Réactions et Génie des Procédés (LRGP)” of Nancy for monitoring and modelling treatment performance. (One year monitoring in 2014-2015 is included in the AZHUREV project). The basic design and lay-out of the future constructed wetlands is elaborated in a collaborative manner in workshops with inputs from all team members, and in discussion with stakeholders, local politicians and the water agency. Preliminary studies include  Site analysis and meetings with major stakeholders,  Collection and synthesis of all available documentation and information about water quality, hydrology, historic and present uses of the site, etc.  Inventory of fauna and flora present on the site,  Topographical survey, soil and subsoil survey, hydro-geological survey,  Assessment of existing site pollution (heavy metals, hydrocarbons and pharmaceutical residues) and the risks of lixiviation of pollutants through the project  Studies of social acceptability problems and public access policy in other wetland projects  State of the art report on constructed wetlands, especially free water surface flow wetlands for tertiary treatment of nitrogen, phosphorous and emerging pollutants, as well as for stormwater management (“event driven wetlands”) ABSTRACTS - WETPOL 2013 - October 13-17, 2013 - Nantes - FRANCE 222

RESULTS AND DISCUSSION At the time of writing of this abstract, the project is in the end-phase of the preliminary studies and beginning of the design phase, so the definite design has not yet been decided upon. However, the area for the AZHUREV project has already been chosen according to criteria of topography and stakeholder acceptability linked to present land-use. A preliminary flow scheme has also been established, consisting of a first settling pond for sand and grid removal of storm water, followed by a permanent pond also receiving the treated wastewater with an extension allowing storage volume for stormwater, followed by predominantly surface flow wetlands optimised for treatment, with a final wetland and pond area optimised for biodiversity before discharge into the creek.

Fig. 1. Preliminary lay-out of the AZURHEV project

The soil and subsoil survey and the hydro-geological have shown that the topsoil layers are very permeable and that the groundwater table is high, with the existing ponds at ground water level. Present studies aim to establish what would be the risks of groundwater pollution through the project, if the intermediate soil layer can be made sufficiently watertight or if artificial liner will be required, which of course would be a major cost factor. Another question is whether the water should be allowed to filter through the permeable topsoil layers. Experiments from similar work (for example that of Ruehmland et al. 2008, with seven different pilots – CW and ponds- constructed at the former landspread site of Berlin) show that the physical filtration through a planted soil matrix can enhance treatment results for organic matter, nitrogen and some polar, persistent micropollutants, but they also bear the risk of lixiviation of pollutants and remobilisation of stored phosphorous. ACKNOWLEDGEMENTS The authors gratefully acknowledge the financial and technical support provided to the project team by the French Ministry of Environment via the Seine-Normandy Water Agency and by the Council of Reims Métropole via its Water and Sanitary Division. REFERENCES Ruehmland, S.; Barjenbruch, M.; Rustige, H., Heinzmann, B.; Duennbier, U. (2008): Comparison of seven CWS and Ponds for Advanced Wastewater Treatment, IWA, Presentation given at the 11th International Conference on Wetland Systems for Water Pollution Control, Institute of Environment Management and Plant Sciences, Vikram, University, Ujjain, India

ABSTRACTS - WETPOL 2013 - October 13-17, 2013 - Nantes - FRANCE 223

Effective Treatment of Combined Sewer Overflow (O.133) Katharina Tonderaa, Silke Rodera, Kassandra Klaera, Ira Brücknera, Johannes Pinnekampa a

Institute of Environmental Engineering of RWTH Aachen University, 52056 Aachen, Germany ([email protected])

INTRODUCTION Combined sewer overflow (CSO) can be a main source of surface water pollution in densely populated areas. The challenge is to treat several thousand cubic meters of combined sewer overflow during and after heavy or long lasting rainfall events. Up to now, only few technologies have been tested on large scale. Retention soil filters (RSFs) have proven to provide an effective retention and degradation of total suspended solids (TSS), ammonium and bacteria in the last two decades (Uhl and Dittmer 2005), but there is still need for optimisation. There is a growing interest for this technique in other European countries as for France and Italy (Meyer et al. 2013). In a recent research project, we focused on the retention of pollutants as TSS, bacteria and micropollutants in RSFs that have been in operation for several years. RSFs require a certain building area depending on the estimated amount of combined sewer overflow that needs to be treated. Thus, different research projects at our institute investigate other available technology: lamella settlers, treatment with ozonation, UV-radiation and performic acid. Lamella settlers help to improve the sedimentation of SS (Andersen at al. 2005). In our research project, a stormwater settling tank (SST) for the treatment of CSO was combined with lamella settlers to improve sedimentation. The effectiveness of the treatment with ozonation and UV-radiation in the special application of combined sewer overflow have only been evaluated in very few projects (EPA 2003; Wojtenko et al. 2001). There is no evaluation known for the treatment with performic acid. We evaluate the cleaning efficiency of these methods at the outflow of a SST for CSO. When it comes to costs, for RSFs and lamella settlers, the operating costs are almost independent from the amount of treated water whereas for the other treatment methods, operating costs such as utilities (especially chemicals) and energy costs correlate highly with the amount of treated water. The aim of this paper is to present the techniques and their application in the treatment of combined sewer overflow discharge and compare CAPEX and OPEX for these installations. METHODS A retention soil filter in the region between Cologne and Dusseldorf with a retention volume of about 4,200 m3 has been investigated in a one year project phase (Tondera et al. 2013a, 2013b). In this period, samples from the inflow and outflow were and analysed as to standard wastewater polluting substances, bacteria and bacteriophages and micropollutants. In a stormwater tank for CSO in the Ruhr area, a lamella settler is installed in a stormwater settling tank that experiences inflows with heavy first flush loads. There is a parallel tank as a reference without a lamella settler. The project will include ten events with samples taken at both inflow and outflow which will be analysed to TSS and COD. At the outflow of a stormwater tank in the Ruhr area with a retention volume of 2,350 m 3 two parallel set-ups from the company Xylem Water Solutions provide ozonation as well as UV-radiation for partial streams. In a reactor, less than 2 m3 of CSO-discharge are treated with ozone, 20 to 50 m3 via UV-radiation. Inflow and outflow are tested as to standard wastewater polluting substances, bacteria, bacteriophages and parasites.

ABSTRACTS - WETPOL 2013 - October 13-17, 2013 - Nantes - FRANCE 224

In summer 2013, a set-up with performic acid at the outflow of a stormwater tank is planned to be installed in the same region. RESULTS AND DISCUSSION At the RSF, the retention of total suspended solids was always at a level of more than 90%. The retention for Escherichia Coli was about 1.1 log grades in medium. Analyses from the first events give no clear results yet for the retention with the lamella settler. Since the beginning of the project, only few heavy rainfall events occurred. Thus, there will be samples taken at the next events presumably in summer 2013. Results for the treatment via ozonation, UV-radiation and performic acid are expected in summer 2013. CONCLUSIONS First results of the large scale experiments will help to evaluate the utilisability of CSOtreatment methods in future application. For RSFs and lamella settlers, the operating costs are almost independent from the amount of treated water whereas for the other treatment methods, operating costs such as utilities (especially chemicals) and energy costs correlate highly with the amount of treated water. For retention soil filters, the retention of bacteria even after several years of operation could be proved in a large scale project. Analyses from the first events give no clear results yet for the retention with the lamella settler. Results for the treatment via ozonation, UV-radiation and performic acid are expected in summer 2013. ACKNOWLEDGEMENTS We express our warmest thanks to the Federal Ministry of Education and Research, the Ministry for Climate Protection, Environment, Agriculture, Nature Conservation and Consumer Protection of the German Federal State of North Rhine-Westphalia for the support of the R&D projects. REFERENCES Andersen N.K., Hallager P., Laustsen A., Nielsen J.B., Kristensen N.T.D., Nordemann P., Rauch W., Nielsen G.G., Arnbjerg-Nielsen K. 2005 Full scale testing of enhanced local treatment enabling a cost-efficient implementation of the EU-Framework Directive. 10th International Conference on Urban Drainage, August 2126, Kopenhagen, Denmark. EPA 2003 CSO Disinfection Pilot Study: Spring Creek CSO Storage Facility Upgrade. Research Summary, Environmental Protection Agency, USA. Meyer D. 2013 Constructed Wetlands for Combined Sewer Overflow Treatment – Comparison of German, French and Italian Approaches. Water, 5 (1), 1-12 Tondera K., Koenen S., Pinnekamp J. 2013a Survey monitoring results on the reduction of micropollutants, bacteria, bacteriophages and TSS in retention soil filters. WST, in press. Tondera K., Koenen S., Pinnekamp J. 2013b Combined Sewer Overflow Treatment: Removal of oxygendepleting parameters through Retention Soil Filters. In Proceeding of the 8th International Conference Novatech, July 23 to July 27, 2013. Uhl M., Dittmer U. 2005 Constructed wetlands for CSO treatment: an overview of practice and research in Germany. Water Science and Technology, 51(9), 23-30. Wojtenko I., Stinson M.K., Field R. (2001) Performance of Ozone as a Disinfectant for Combined Sewer Overflow. Critical Reviews in Environmental Science and Technology, 31 (4). 295-309.

ABSTRACTS - WETPOL 2013 - October 13-17, 2013 - Nantes - FRANCE 225

Constructed wetlands as buffer zones between treatment plants and receiving water bodies: design and pre-implementation considerations (O.127) Petitjean, Aa, Forquet, Na, Choubert, Coquery, Ma, J-Ma, Boutin, Ca a

IRSTEA Lyon, 5, rue de la Doua, 69626 Villeurbanne - CS70077, France [email protected], [email protected])

INTRODUCTION Buffer zones between treatment plants and receiving water bodies have recently gained increased interest in France (Boutin et al., 2012). These planted discharge areas are similar to ponds, ditches, meadows, or to more conventional tertiary treatment systems (but without treatment requirements). The results of a French study (Boutin et al., 2012) show the diversity of the systems that are implemented, as well as the multiplicity of the intended outcomes. Research is needed to assess the efficiency of these systems: a comprehensive study will be carried out to understand the fate of water, conventional pollutants (suspended solids, ammonia, and phosphorus) and micro pollutants (substances that are refractory to up-stream biological treatment). To this aim, a five-year project has been launched, including the building of two kinds of semi-industrial scale pilots of planted discharged areas: meadows and ditches. This extended abstract shows the preliminary studies lead on the land, prior to the construction of the systems, and gives an overview of sizing and design choices. An original approach aiming at designing and implementing buffer zones on a given area is presented. PRESENTATION OF THE PILOT-SCALE PLANTED DISCHARGE AREAS Required information for planted discharged areas implementation Meadows and ditches mimic “natural” kinds of wetlands. Therefore, their implementation requires special attention given to the study of the field where they should be settled. Indeed the knowledge of the fate of the infiltrated treated wastewater is crucial for the conservation of the receiving water bodies. In this study, several geological and hydrological techniques were used to characterize the field: resistivity, radar, augers, permeability determination (“Porchet” test), piezometers, soil textures, presence of contaminants. These techniques provide the location of the different soil layers and the eventual groundwater of the land; and their physicochemical properties. Firstly, the consolidated results are decision tools regarding the sizing and placement of meadows and ditches. For instance, the study shows that the hydraulic working of the land is mainly driven by rainfall events and tides. Secondly, they are the basis for hydraulic and transport modelling, that will be used for qualitative and quantitative determination of the fate of pollutants. Design of planted discharged areas pilots The sizing of the pilot-scale meadows and ditches was set-up assuming that the whole effluent could be infiltrated in the most favourable infiltration capacity of the land. Figures 1 and 2 give the sketches of respectively meadows and ditches. Both of them are duplicated in order to study different operating conditions. Meadows and ditches are to be equipped with sensors for hydraulic measurements (TDR – Time Domain Reflectometry,

ABSTRACTS - WETPOL 2013 - October 13-17, 2013 - Nantes - FRANCE 226

ERT - Electrical Resistivity Tomography), and with porous cups for micro and macro pollutant characterisation of the pore water. These pilots will allow us to draw conclusions regarding the evolution of the hydraulic working of the systems, as well as their intrinsic soil remediation capacity.

Effluent system

equirepartition

Sensors Planted undisturbed soil Impermeable wall Manhole for sensor placement Discharge spout

Fig. 1. Design of the pilot-scale meadows

Pipe between ditches

2

Ditch Sensors

Manhole for sensor placement

Fig. 2. Design of the pilot-scale ditches

CONCLUSIONS The authors underline the importance of preliminary studies on the field, before the sizing and placement of planted discharge areas. From the research point of view, no relevant sensor positioning could be possible without these studies. This abstract gives the outlines of the sizing and design of pilot-scale planted discharge areas. The accuracy of preliminary fieldstudies and the evolution over time of the characteristics of the land still are at stake. ACKNOWLEDGEMENTS The authors are grateful to the ONEMA (Office National de l’Eau et des Milieux Aquatiques), and the Comunauté Urbaine de Bordeaux (CUB), for funding and support. REFERENCES Boutin, C, Prost-Boucle, S. (2012). Note de présentation – Les zones de rejet végétalisées. Sciences Eaux et territoires 09 (2012)

ABSTRACTS - WETPOL 2013 - October 13-17, 2013 - Nantes - FRANCE 227

N-P cycle in wetlands

ABSTRACTS - WETPOL 2013 - October 13-17, 2013 - Nantes - FRANCE 228

Organic matter and nitrogen removal in horizontal subsurfaceflow constructed wetlands planted with Cyperus articulatus (O.10) Hamer Vegab, Juan Camilo Lancherosab and Aracelly Caselles-Osorioab a

Centro de Estudios del Agua, Universidad del Atlántico

b

Research group: Wetlands Caribbean Colombian. Universidad del Atlántico, Km 7 Old way Puerto Colombia, Barranquilla, 08001, COLOMBIA [email protected]

INTRODUCTION One of the major problems causing pollution of freshwater and estuarine ecosystems is inadequate treatment of domestic, industrial and agricultural wastewater. Left untreated, these wastewater sources contribute prodigious quantities of organic carbon and fertilizer nutrients to receiving streams, which in the short term leads to oxygen depletions, fish kills, and the proliferation of algal blooms. The solution to this wide-scale problem is to treat and/or reuse the concentrated wastewater streams near their source. In urban regions this entails building and operating large sophisticated wastewater treatment plants, while in rural areas, it is necessary to deploy appropriate technologies, such as lagoons and constructed wetlands. One of the more promising technologies, horizontal subsurface-flow constructed wetlands (HSSF), entails use of shallow gravel beds populated with naturally occurring microbial communities and selected aquatic plants. It has been demonstrated that the use of aquatic macrophytes in HSSF constructed wetlands can enhance wastewater treatment, especially with respect to removal of nitrogen compounds such as ammonium and nitrate. There is little information available regarding the use of tropical plant species, for which there is a broad and rich biodiversity. The following paper evaluates the tropical emergent macrophyte Cyperus articulatus, a fast-growing species endemic to the tropical Colombian Caribbean region. METHODS This four month study was conducted at Laboratories of Centro de Estudios del Agua, Universidad del Atlántico, Barranquilla, Colombia. The mesocosm-scale systems consisted of four horizontal SSF, each measuring 0.61m2 and 0.5 m deep (water depth of 0.4 m). The systems were filled with granite gravel 10 mm in diameter, and with a void space of approximately 40%. Two treatment units were planted with C. articulatus at a density of 33plants/m2 while the other two units served as unplanted controls. The systems were batch loaded daily with university municipal wastewater which was pretreated in a 760 L septic tank. The loading rate was equivalent to a 3-day hydraulic retention time. Wastewater and plants analysis Influent and effluent water samples were collected twice a week and analyzed for COD, ammonium-N, nitrite-N and nitrate-N. Sample analyses were conducted following guidelines according to APHA -AWWA-WPCF (2005). Parameters including temperature, pH, redox potential and DO were measured in situ using hand held sondes (WTW Multi 3420) After four months of treatment, the aquatic plants were harvested, dried to a constant moisture content and weighed on an electronic balance (OHAUS Trooper model TR6RS). Normality and T-tests were performed to compare mean values of COD, ammonium, nitrate and nitrite between systems planted and not planted (Statgraphics Centurion XV.II).

ABSTRACTS - WETPOL 2013 - October 13-17, 2013 - Nantes - FRANCE 229

RESULTS AND DISCUSSION The average values of several physicochemical parameters showed similarities in both treatments (Table 1). However, pH, EH and electrical conductivity varied among treatments, possibly as a result of rapid plant growth Table 1. Averages (SD in brackets) of the water quality parameters in the effluent of the HSSF. Calculations based on 43 to 54 measurements depending on the parameter Parameter Temperature, oC Oxygen Dissolve, m/L EH, mV pH, Electric conductivity, uS/cm Total biomass, dry Kg/m2

Cyperus articulatus 29 .0 (1.3) 2.1 (1.2) 35. 0 (128.0) 7.0 (0.22) 781.1 (99.3) 6.4 (0.4)

No Plants 28.5 (1.1) 1.7 (1.2) 44.6 (85.3) 8.2 (0.14) 682.0 (114.0) -

There were also significant differences in ammonium removal rates among treatments, with the planted systems removing more ammonium-N that the unplanted control (P>005). Due to rapid growth and high biomass production of C. articulatus, it was necessary to collect biomass in two months of the growth period. COD removal averaged (75-81%), however there were only marginal differences among treatments, demonstrating that organic matter removal can occur with or without plants (Fig 1). These results are consistent with those reported by Caselles-Osorio et al. (2011). 100 Influent C. articulatus No plants

140

COD, mg/L

120 100 80 60 40

Afluente C. articulatus No Plants

80

NH+4, mg/L

160

60 40 20

20 0

0 Sep

Oct

Nov

Time, months

Dec

Sep

Oct

Nov

Dec

Time, months

Fig. 1. Average concentrations of influent and effluent concentrations of COD and ammonium-N as a function of treatment and time.

CONCLUSIONS C. articulatus is a rapid growing tropical macrophyte that can provide significant removal of nitrogen compounds in constructed wetlands treating municipal wastewater. ACKNOWLEDGEMENTS Centro de Estudios del Agua of the Universidad del Atlántico who facilitated facilities to advance research laboratories. REFERENCES

APHA-AWWA-WPCF. 2005. Standard Methods for the Examination of Water and Wastewater, 19 th ed. American Public Health Association. Washington, DC. Caselles-Osorio, A., P. Villafane., V. Caballero and Y. Manzano. 2011. Efficiency of Mesocosm-Scale Constructed Wetland Systems for Treatment of Sanitary Wastewater Under Tropical Conditions. Water, Air, & Soil Pollution 220. 161-171

ABSTRACTS - WETPOL 2013 - October 13-17, 2013 - Nantes - FRANCE 230

How to use in-situ UV/visible spectrophotometer for a better understanding of time-dependent constructed wetlands processes: a focus on nitrate (O.56) Nicolas Forqueta, Camille Provosta, Pascal Mollea a

Irstea - Freshwater system, Ecology and Pollution Research unit, 5 rue de la Doua, CS 70077, 69 626 Villeurbanne Cx, FRANCE ([email protected])

INTRODUCTION In order to better know processes occurring within constructed wetlands, mechanistic modelling is often presented as the most promising tool. However these models come with a large number of parameter to optimise. Therefore users often face an ill-posed problem: they have too little data to estimate the parameters either by direct or inverse methods. To overcome this issue, one can use regularisation technique but must moreover gain additional data to which models parameters are sensitive to or related. Among the different techniques to harvest more data, we focus in this paper on online measurements of pollutant concentrations at the outflow using an in-situ UV/visible spectrophotometer. UV/visible spectrophotometry is one of the three main online measurement techniques available, the two others being ion specific probes and online analysers. The main advantages of in-situ UV/visible spectrophotometer over the two others are that (i) it does not require any chemical supply, (ii) it also requires less operating care when equipped with automatic cleaning device and finally (iii) there is little drift of the emitting light over time. The main drawback of in-situ UV/visible spectrophotometry is that it does not provide direct measurements of the pollutant concentration. They can only be obtained after post-processing the collected spectra making this step critical. This paper presents our efforts to efficiently carry out calibration for measurements taken at the outflow of a first stage of a vertical flow constructed wetland fed with raw wastewater. We also decided to only focus on nitrate and nitrite measurements. MATERIALS AND METHODS An in-situ UV/visible spectrophotometer (S::CAN Messtechnik) was immersed at the outflow of a first stage of a vertical flow constructed wetland fed with raw wastewater (Evieu, France). The bed was fed during 3.5 days according to the normal operation. All along this period, 24 hours samples proportional to flow were taken as well as 26 grab samples. Rieger et al. (2008) distinguished two ways to perform the calibration of NOx-N measurements by spectrophotometer: the local and the global calibration. The local calibration is performed using the concentration results given by the spectrophotometer software and relies on the quality of the post-processing algorithm provided by the manufacturer which are not accessible to the user. Whereas the global calibration looks at relationship between the full spectra and pollutant concentration (in our case the NOx-N). Local calibration was performed using all 26 grab samples. We also performed a global calibration using partial least squares (Mevik and Wherens, 2007) to establish a relationship between the spectra and NOx-N concentrations. A particular attention was given to the estimation of uncertainties. Once the measurements were over, the spectrophotometer was taken back to the laboratory. There, we measured the spectrophotometer response to samples of demineralized water mixed with increasing known nitrate concentrations. ABSTRACTS - WETPOL 2013 - October 13-17, 2013 - Nantes - FRANCE 231

RESULTS A specificity of constructed wetland nitrate and nitrite outflows is the high values that can be observed especially after the first feeding that follows a rest period. Concentrations up to 80 mg/L of NO3-N were measured in the grab samples taken during the first four feedings. These values are significantly higher than the upper measurement range provided by the spectrophotometer manufacturer: 25 mg/L of NO3-N. We verified that linear relationship between NOx-N concentrations and spectrophotometer estimates breaks down for concentrations larger than 25 mgN/L. However, we demonstrated that using a second order polynomial approximation, we were able to predict concentration up to 35 mg/L. Contrarily to the local calibration, the global calibration was not limited in its range. Grab samples were separated between a training and a calibration dataset. A PLS was performed on the calibration dataset. 5 latent vectors were used explaining up to 99.9% of the total variance of the spectra and 99.6% of the variance of nitrate and nitrite concentrations. Global calibration performs better than the local one, residuals never exceed 1.36 mg/L. This online monitoring, well calibrated, allows to clearly point out the fluctuation of nitrate outlet concentrations with batches and feeding/rest periods. It is a necessary tool for modelling purposes. CONCLUSIONS Using an appropriate calibration procedure, on-line UV/visible spectrophotometry provide us with a detailed description of the NOx-N outlfow concentrations over time. Especially, it enables us to observe the nitrate wash-out that follows the first feeding after the rest period and to quantify it. ACKNOWLEDGEMENTS The authors would like to thanks researchers and technicians who participated to data collection, namely: Clément Crétollier, Olivier Garcia, Jean-Marc Choubert et Ania Morvannou. The authors would also like to thanks the chemists from the LAMA laboratory who analysed the samples. REFERENCES Rieger, L. and Langergraber, G. and Kaelin, D. and Siegrist, H. and Vanrolleghem, P.A. (2008) Long-term evaluation of a spectral sensor for nitrite and nitrate. Water Sci. Technol. 57(10): 1563-1569. Mevik, B.-H. and Wehrens, R. (2007) The pls Package: Principal Component and Partial Least Squares Regression in R. J. Stat. Softw. 18(2): 1-24.

ABSTRACTS - WETPOL 2013 - October 13-17, 2013 - Nantes - FRANCE 232

Effect of bed depth on nitrogen removal in a Vertical Flow Constructed Wetland planted with Vetiveria zizanioides (O.71) A.Almeidaa, M.Imaginárioa, I. Ferreira a, I. Castanheira a, C.Ribeiro a a

Departamento de Tecnologias e Ciências Aplicadas, IPBeja, Ap 158, 7801-902 Beja, Portugal; (E-mail: [email protected], Tel. +351284314300; Fax +351284388207).

INTRODUCTION Due to the presence of nitrogen compounds, most wastewater requires extensive treatment prior to environmental disposal. Constructed wetlands can be a good solution to nitrogen removal (Kadlec and Wallace, 2008). The processes that allow nitrogen removal are very dependent on several factors such as: pH, temperature, dissolved oxygen (DO), feeding mode, hydraulic load (HL) and hydraulic retention time (HRT).The Vertical Flow Constructed Wetland (VFCW) promotes higher atmospheric oxygen diffusion inside the matrix, which can boost nitrification (Cooper, 2005). Denitrification is known as an anaerobic process which can be inhibited by dissolved oxygen (DO) higher than 1 mg L-1, however it was observed for DO until 1.5 mg L-1 in systems with fixed and suspense biomass (EPA, 1993 in Kadlec and Wallace, 2008). Aerobic and anoxic zones always tend to coexist in constructed wetlands due to the presences of the plants, contributing to nitrogen removal. Also Zhang et al, (2010) found that the retention of NH4+-N and NO3--N increased with increasing species richness, and activities of enzymes such as protease and nitrate reductase, were strongly depended on the presence or absence of plants. Kong et al., (2009) noticed that the ammonium nitrogen removal efficiency was significantly correlated with both urease and protease activity in systems planted with Vetiveria zizanioides. METHODS Two VFCW microcosms (bed A (0.4  0.6  0.35 m) and bed B (0.4  0.6  0.70 m),filled with light expanded clay aggregates (Leca®NR 10/20) and planted with Vetiveria zizanioides were used. The system was constituted by two beds operated in parallel mode with a 56 (±SD 5) ≤ HL ≤ 358 Lm-2d-1 (±SD 13), and was fed with synthetic wastewater prepared as described in Almeida et al, (2009) with increasing concentration of ammonium nitrogen from a reservoir (125 L) equipped with two submersible pumps. Continuous feeding of equidistant sprinklers ensured a good distribution over the whole surface area. Dissolved oxygen (DO), NH4+-N, NO3--N and NO2--N were measured as described in Standard Methods (1998) in the influent and effluent collected daily from each bed. RESULTS AND DISCUSSION By the results obtained, toTN loads applied ranging from 4 gN m-2 d-1 (±SD 2) until 108 gN m-2 d-1 (±SD4) it can be said that ammonium nitrogen removal occurred in the two beds, (NH4+-Nremoved ≤26 gN m-2 d-1 (±SD 10) in bed A and NH4+-Nremoved ≤13 gN m-2 d-1 (±SD 2) in bed B respectively). The TN removed in bed A was TNremoved ≤22 gN m-2 d-1 (±SD 12) and TNremoved ≤9 gN m-2 d-1 (±SD 3) in bed B (Fig 1). Nitrification occurred in both beds, but lower in bed B (Fig 1). Differences between means obtained for each parameter in each bed were evaluated using ANOVA at the significance level of p10 mM cause leaf chlorosis and diminution in the wetland plants growth. It was not observed toxicity in the leaves of plants in both beds although was used [NH4+-N] 21mM (±SD 0.6). Bed A

Bed B

Fig. 1 – Ammonium ( ), total nitrogen ( ) removed and nitrate produced ( A (0.70 m deep) and bed B (0.35 m deep), fed in parallel; Mean (±SD).

) in bed

CONCLUSIONS Ammonium and total nitrogen removal and nitrification did occur in the two beds. Significative differences between means were obtained to loads applied of TN ≥ 29 gN m-2 d-1 (±SD 3.4), probably due to the different depths of the root systems of Vetiveria zizanioides in the two beds. Higher depth of root system seems to favor the creation of more zones with different oxidations conditions which allow a more intense removal of nitrogen compounds. Further work is needed to determine how depth of the bed influences these processes. REFERENCES Almeida, M. A., Ramalho V., Fernandes, R., Novais, J. M., Martins-Dias S. (2009) Ammonium Nitrogen Removal in a Vertical Flow Constructed Wetland Planted with Vetiveria zizanioides. 3rd Wetland Pollutant Dynamics and Control, Barcelona, Spain, pp 238-239. Britto, D.T.,Kronzucker, H. J.(2002) NH4+ toxicity in higher plants: a critical review. J. Plant Physiol.,159 (6), 567-584. Cooper, P. (2005) The performance of vertical flow constructed wetland system with special reference to the significance of oxygen transfer and hydraulic loading rates. Wat. Sci. Tech., 51 (9), 81-90. Kong, L., Wang, Y.B., Zhao, L.N., Chen, Z.H. (2009) Enzyme and root activities in surface-flow constructed wetlands. Chemosphere, 76 (5), 601-608. Kadlec, R.H., Wallace, S.(2008) Treatment Wetlands 2nd ,CRC Press, Taylor & Francis, 998. Standard Methods for the Examination of Water and Wastewater (2005)APAH (American Public Health Association), 21st Edition, Washington. Zhang, C.B., Wang, J., Liu, Wen, L.L., Zhu, S.X., Liu, D., Chang, S. X, Chang, J., Ge, Y. (2010) Effects of plant diversity on nutrient retention and enzyme activities in a full-scale constructed wetland. Bior. Techn., 101(6),1686-1692.

ABSTRACTS - WETPOL 2013 - October 13-17, 2013 - Nantes - FRANCE 234

Influence of influent wastewater composition on removal rates in multistage treatment wetlands (O.82) Magdalena Gajewskaa, Fabio Masib, Krzysztof Jóźwiakowskic, Ahmed Ghrabid

a

Gdańsk Universiuty of Technology, Faculty of Civil and Environmental Engineering Narutowicza st. 11/12, 80-233, Gdańsk, POLAND, ( [email protected]) b IRIDRA, Via Alfonso La Marmora 51, 50121, Florence, ITALY ([email protected]) c University of Life Sciences in Lublin, Faculty of Production Engineering Akademicka St. 13, 20-950, Lublin, POLAND, ([email protected]) d Research Center for Water Technologies – BP.273, 8020, Soliman, Tunis, TUNISIA ([email protected]) INTRODUCTION Wastewater can have various composition and, as a consequence, different biodegradability under treatment processes. It has been suggested that the level of easy degradable organic matter (measured by BOD5) and nitrogen (measured by TKN) to ensure effective nitrification could be like BOD5/ TKN < 1 in treatment wetlands (Crites et., al 2006; Kadlec and Wallace., 2009). An analysis of COD/BOD5 and BOD5/TN (or BOD5/TKN) ratios can provide useful information as: (i) rate of biodegradability of the studied wastewater (ii) nitrification capacity of the reactor basing on the quantification of the degradable dissolved organic matter, which is fundamental for effective removal of ammonium nitrogen in the conventional nitrification/denitrification process. The objective of the paper is to evaluate the influence of the influent wastewater composition, expressed by BOD5/COD and BOD5/TN ratios, on kinetics of organic and nitrogen removal in multistage treatment wetlands (MTWs). METHODS The investigations were carried out in five MTW in Poland: three for single family and two for local community and in one in Italy as well as one in Tunisia both for local communities (Table 1) (Gajewska and Obarska-Pempkowia, 2011; Jóżwiakowski 2009; Masi et al, 2013). Table 1. The operational conditions of the MTWs Plant and location

Flow [md-1]

pe

Configuration

Area [m2]

Janów, Poland

0.66

3

VF +HF

Σ 48

Dąbrowica I, Poland

0.3

3

VF+HF

Σ 48

Dąbrowica II, Poland

0.3

3

HF+VF

Σ 48

Darżlubie, Poland

56.6

650

HF I + Cascade bed +HF I+ VF+ HF III

Σ3350

Wikono, Poland

18.6

220

HF I +VF +HF II

Σ1902

Dicomano, Italy

525

3500

HF I + VF + HF II + FWS

Σ 6080

Chorfech, Tunisia

17.0

500

HF I + VF + HF II or Reservoir

Σ 1800

The samples of influent and effluent as well as collected with respect to hydraulic retention time after each stage of treatment in MTW were collected and analyzed for BOD 5, COD and TN according to the standard methods (APHA, 2005).

ABSTRACTS - WETPOL 2013 - October 13-17, 2013 - Nantes - FRANCE 235

RESULTS AND DISCUSSION In all analysed MTWs the quality of discharged wastewater was unstable in time what was expressed by very high coefficient of variation form 20% up to 800%. In spite of this the removal average efficiency of organics (BOD5 and COD ) and TN was very high (Table 2). The ratios presented a high fluctuation, the COD/BOD5 ratio varied from 1 to 4 and the BOD5/TN from 0.2 to 4.2. Relationships presented in Figure 1 do not confirm clear dependence between quality of wastewater represented by COD/BOD5 and BOD5/TN ratios on TN removal. Table 2 Efficiency removal in investigated MTWs Janów, Poland

Dąbrowica I, Poland

Dąbrowica II, Poland

Wikolno, Poland

Darżlubie, Poland

Dicomano Italy

Chorfech, Tunisia

COD BOD5 TN

n=11 93% 96% 68%

n=16 88% 92% 47%

n=16 90% 93% 59,5%

n=22 89% 96% 83%

n=23 90% 92% 91%

n=47 85% 92% 43%

n=7 95% 97% 71%

100,00

100,00

80,00

80,00

TN removal [%]

TN removal [%]

Paramiter/ Plant

60,00 40,00 20,00 0,00

60,00 40,00 20,00 0,00

0,00 2,00 y = 3,3315x + 57,517 R² = 0,0434

4,00

6,00

BOD 5/TN

8,00

0,00 1,00 2,00 y = -12,557x + 97,374 R² = 0,1372

3,00 4,00 5,00 COD/BOD 5

Fig1 Total nitrogen efficiency vs BOD5/TN and COD/BOD5 rations

Achieved results suggested that MTWs could ensure effective removal of nitrogen even when the concentration of organic is equal to TN and the organic matter is mostly present in less degradable forms (expressed as COD). In the paper the first-order areal rate constant for BOD5 and TN will be calculated and separate assessments for each stage of treatment will be conducted. CONCLUSIONS Achieved results suggested that multistage treatment wetlands are capable to ensure sufficient removal of both organic and nutrients even in unfavourable proportions of macronutrients (C and N). The usual assumptions for conventional biological treatment systems concerning adequate C/N ratios seem to be dubious in case of wastewater treatment in MTWs. REFERENCES Crites R.W., Middlebrooks E.J., and Reed S.C. (2006). Natural Wastewater Treatment Systems: New York, NY, Taylor & Francis Group, 552. Gajewska M., Obarska-Pempkowiak H. (2011). Efficiency of pollutant removal by five multistage constructed wetlands in a temperate climate. Environ. Protect. Engin.,37(3): 27-36. Jóźwiakowski K. (2009). Possibilities of protection of water resources with the use of constructed wetland systems: a review. Ecohydrology & hydrobiology. 9(2-4), 297-306. Kadlec R.H., Wallace S. (2009). Treatment Wetlands, Second Edition CRC Press Taylor & Francis Group, Boca Raton, London, New York, 1016, Masi F., Caffaz S. and Ghrabi A.. (2013). Multi-stage constructed wetlands systems for municipal wastewater treatment, Water Science & Technology, 67, 7, 1590-1598 ABSTRACTS - WETPOL 2013 - October 13-17, 2013 - Nantes - FRANCE 236

Characterization of the deposit layer of a partially saturated VFCW and influence of pH and redox on phosphorus release (O.67)

B. Kima,b, M. Gautiera, A. Bueta, A. Simidoffa, P. Michelb, R. Gourdona a

University of Lyon, INSA of Lyon, LGCIE, 20 av. A. Einstein, Villeurbanne cedex, F69621, FRANCE ([email protected])

b

SCIRPE (Society of design and production engineering for wastewater purification), 5 Allée Alban Vistel, Sainte Foy Lès Lyon, F-69110, FRANCE

INTRODUCTION Vertical-Flow Constructed Wetland (VFCW) has good efficiency for organic load removal and for nitrification. However, their phosphorus retention is still questioned (Brix et al., 2001). Indeed, the release of phosphates into sensitive aquatic ecosystems may promote eutrophication. The French company SCIRPE has developed a compact VFCW process called Azoé® (patent EP1857419A1 – Michel, 2007). In this process, dissolved phosphates are precipitated by injection of FeCl3 at the output of the trickling filter. The precipitates thus formed, mixed with suspended solids, thereby progressively accumulate as sludge on the surface of partially saturated VFCWs (ps-VFCWs). Before final evacuation, the sludge layer may undergo various bio-physico-chemical changes depending on the environmental conditions (pH, Eh, etc.). The objectives of the present study were to (i) characterize this deposit layer with a focus on phosphorus, (ii) to investigate the influence of the extreme conditions of pH and redox on phosphorus release. METHODS Samples The first Azoé® plant for domestic wastewater treatment, in operation since 2004 in Vercia (Jura, France, maximal capacity of 1,100 PE), was selected for sludge sampling. The average depth of sludge layer accumulated on the 1st stage of the plant since its implantation was around 16 cm, corresponding to a rate of accumulation of about 2 cm per year. Sludge samples were taken by manual shovelling of the deposit layer, stored at 4 °C until they were dried at 35 °C, sieved to 1 mm for further experiments and analyses. Physico-chemical and mineralogical sample characterization Organic Matter (OM) content of dried sample was estimated by calcination at a temperature of 550 °C for at least 1 h. Total element concentrations were determined by ICPAES analysis after alkaline fusion (LiBO2). Mineralogical characterization was carried out on dried and calcinated samples using X-ray diffraction (XRD) performed with a BRUKER® D8 Advance instrument and Fourier Transform Infrared (FTIR) spectrometry with a Perkin-Elmer Spectrum One FTIR spectrometer. Experimental set-up for leaching tests To study the influence of pH and redox on phosphorus release, successive leaching tests were performed in triplicates under different pH conditions by using aqueous solutions of different leachants (deionized water, HCl, NaO and sodium ascorbate). Ten grams of dried sludge were mixed with 100 mL of leachate at room temperature for at least 24 h under continuous rotary shaking. After centrifugation, each eluate was filtered through 0.45 µm on

ABSTRACTS - WETPOL 2013 - October 13-17, 2013 - Nantes - FRANCE 237

Minisart® syringe filters and the pellet was resuspended in a fresh leaching solution. The overall process was repeated. Each filtered eluate was analysed for pH, redox and dissolved reactive phosphate by molybdenum blue colorimetric method (Murphy and Riley, 1962) immediately after sampling. Complementary analysis of dissolved Fe, Ca, Al and Mn concentration in eluate were then performed by ICP-AES. At the end of the leaching tests, the pellets were lyophilized for characterizations. RESULTS AND DISCUSSION The mineralization of sludge accumulated at the surface of pf VFCW was clearly highlighted by comparing OM content in the sludge samples and in the suspended solids sampled in the feed influent. Moreover it was observed that phosphorus was principally concentrated in the sludge layer. The phosphorus content was about 2 % wt. The focus of the study was to determine whether phosphorus retained in the sludge layer could potentially be released under extreme conditions. Fig. 1 shows cumulated released phosphorus (% of total P content) and pH of eluate vs cumulated leaching volumes.

Fig. 1. Evolution of eluate pH and cumulated released phosphorus (% of total P content) over the leaching tests.

During the 1st phase of leaching test with mild pH conditions (pH 4 and pH 10), no pH variation of eluate was observed, indicating the strong pH-buffering capacity of the sludge. Released P in each eluate never exceeded 0.5 % of the total P content under mild conditions, showing that P retention within the sludge was not influenced by the pH variations which could be reasonably observed under field conditions. Under very extreme conditions, after 30 repetitions of the process, the percentages of P release compared to P content in the sludge were 6%, 25%, 40% and 50% for deionized water, acid, basic and reduced conditions respectively. CONCLUSIONS A good mineralization of the sludge from a deposit layer of a psVFCW was observed. Leaching tests showed a good pH-buffering capacity of the sludge and the notable resistance of the phosphorus retention within the sludge to the pH variations. Only continuous extreme conditions of pH and low redox lead to a significant phosphorus release from the sludge. REFERENCES Brix, H., Arias, C. A. and del Bubba, M. (2001) Media selection for sustainable phosphorus removal in subsurface flow constructed wetlands. Water Science and Technology. 44(11-12):47-54. Michel, P. (2007) Process and installation of effluent treatment. EP 1 857 419 A1. Murphy, J. and Riley, J. P. (1962) A modified single solution method for the determination of phosphate in natural waters. Analytica Chimica Acta. 27(0): 31-36.

ABSTRACTS - WETPOL 2013 - October 13-17, 2013 - Nantes - FRANCE 238

Steel slag filters to upgrade phosphorus removal in constructed wetlands - main factors controlling removal efficiency (O.98) C. Barcaa,b, D. Meyerc,a,b, S. Troeschb, C. Gérentea, Y. Andrèsa, F. Chazarenca a

L’UNAM Université, Ecole des Mines de Nantes, CNRS, GEPEA UMR 6144, 4, rue Alfred Kastler. B.P. 20722 F-44307 NANTES, France b Epur Nature SAS, 153 Avenue Maréchal Leclerc, 84510 Caumont sur Durance, France c IRSTEA centre de Lyon, 5 rue de la Doua - CS70077, 69626 Villeurbanne, FRANCE ([email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected]). INTRODUCTION Legislation on P rejects for municipal wastewater treatment plants (WWTPs) is becoming stricter in Europe (EU Directives 91/271/EEC and 2000/60/EC). Consequently, the research on low cost techniques to upgrade P removal has gained a growing attention. Many international studies have demonstrated that filtration through reactive materials with high affinities for P binding is a viable technique to upgrade P removal in small WWTPs such as constructed wetlands (CWs) (Vohla et al., 2011). Within the framework of the European research project “SLASORB” (SLAg as SORBent for P removal), our study aimed at investigating the use of steel slag (CaO-rich industrial coproduct) as reactive filter material to upgrade P removal in CWs. Two types of slag were tested: (a) BOF-slag, which originates from the refining of iron in a basic oxygen furnace, and (b) EAF-slag, which derives from melting recycled scrap in an electric arc furnace. An integrated approach was followed with investigations at three different scales: (i) batch experiments; (ii) column experiments; (iii) field experiments. The influences of several parameters, including slag composition, void hydraulic retention time (HRTv), temperature, and wastewater quality, on P removal mechanisms and performances were studied. METHODS Comparative batch experiments were performed to determine the P removal capacities (PRCs) of several samples of slag produced in Europe (Barca et al., 2012). Then, the samples presenting the highest PRCs were selected to perform column and field experiments. Column experiments consisted of a series of lab-scale filters designed in two different sizes: (i) two small-size filters of about 0.04 m3, and (ii) two big-size filters of about 0.08 m3. The small-size filters were filled one with EAF-slag size 5-16 mm (EAF-small), and one with BOF-slag size 6-12 mm (BOF-small). The big-size filters were filled one with EAF-slag size 20-40 mm (EAF-big), and one with BOF-slag size 20-50 mm (BOF-big). The filters were fed with a synthetic P solution (about 10 mg P/L) and operated with a continuous horizontal subsurface flow (HSSF) and HRTv of 1 d under room temperature condition (around 20 °C). Field experiments consisted of two pilot-scale filters of 6 m3, one filled with EAF-slag size 20-40 mm (EAF-filter), and one with BOF-slag size 20-40 mm (BOF-filter). The filters were fed with a fraction (2-4 %) of the effluent from a CW which treats the municipal wastewater of a French village (800 people equivalents), and operated with a HSSF and HRTv of 1-2 d. Inlet and outlet pH, Ca, total P (TP), phosphate (PO4-P), and total alkalinity (TA) of all the filters were monitored over a period of almost two years of operation. RESULTS AND DISCUSSION The experimental results indicated that P removal predominantly occurred via two reactive phases: (i) first, CaO-slag dissolution with increase in pH of the solutions; (ii) then, Ca-P ABSTRACTS - WETPOL 2013 - October 13-17, 2013 - Nantes - FRANCE 239

precipitation followed by adsorption/crystallisation of the precipitates on the slag surface. Column experiments showed that P removal improved with increasing CaO content and with decreasing the size of slag, most probably because the smaller was the size, the greater was the specific surface available for CaO dissolution and adsorption of Ca-P precipitates. Also, field experiments showed that P removal increased with increasing temperature and HRTv, probably because this affected the rate of CaO dissolution and Ca-P precipitation (Barca et al., 2013). Overall, column experiments showed higher outlet pH and TP removal efficiencies than field experiments (Fig. 1), likely because the higher pH favoured Ca-P precipitation. Most likely the higher TA of the CW effluent treated in the field experiments (mean value 282 mg CaCO3/L) compared to that of the synthetic solution (mean value 140 mg CaCO3/L) may have buffered the increase in pH due to CaO dissolution, thus limiting Ca-P precipitation. Also, the presence of organic colloids, humic, fulvic and tannic acids in the CW effluent may have affected the processes of Ca-P nucleation and precipitation. inlet TP

outlet TP

14

outlet pH

12

TP (mg P/L)

16

10

12

8 8

pH (-)

20

6

4

4

0

2 EAF-small BOF-small

EAF-big

Column experiments

BOF-big

EAF-filter

BOF-filter

Field experiments

Fig. 2. Mean inlet and outlet TP, and mean outlet pH values over 100 weeks of experiments (with the exception of column BOF-big (65 weeks)). Bars indicate the range min-max values.

CONCLUSIONS This study showed that EAF and BOF-slag are efficient filter materials to remove P from wastewater via CaO-slag dissolution followed by Ca-P precipitation mechanisms. P removal improved with increasing HRTv and temperature, and with decreasing the slag size, most probably because CaO-slag dissolution was favoured. However, TA of wastewater may buffer the increase in pH due to CaO dissolution, thus limiting Ca-P precipitation. ACKNOWLEDGEMENTS The research leading to these results has received funding from the European Union's Research Fund for Coal and Steel (RFCS) research programme under grant agreement n° RFSP–CT–2009–00028 (SLASORB). REFERENCES Barca C., Gérente C., Meyer D., Chazarenc F., Andrès Y. (2012) Phosphate removal from synthetic and real wastewater using steel slags produced in Europe. Wat. Res. 46:2376-84. Barca C., Troesch S., Meyer D., Drissen P., Andrès Y., Chazarenc F. (2013) Steel slag filters to upgrade phosphorus removal in constructed wetlands: two years of field experiments. Environ. Sci. Technol. 47:549-56 Vohla C., Kõiv M., Bavor J., Chazarenc F., Mander U. (2011) Filter materials for phosphorus removal from wastewater in treatment wetlands - A review. Ecol. Eng. 37:70-89.

ABSTRACTS - WETPOL 2013 - October 13-17, 2013 - Nantes - FRANCE 240

Pilot-scale experimental study of phosphorus retention in deposit layer of partially saturated vertical flow constructed wetland (O.135)

B. Kima,b, M. Gautiera, P. Mollec, P. Michelb, R. Gourdona a

University of Lyon, INSA of Lyon, LGCIE, 20 av. A. Einstein, Villeurbanne cedex, F-69621, FRANCE ([email protected])

b

SCIRPE, Society of design and production engineering for wastewater purification, 5 Allée Alban Vistel, Sainte Foy Lès Lyon, F-69110, FRANCE

c

Irstea, Wastewater Treatment Team, 5 rue de la Doua – CS70077, Villeurbanne cedex, F-69626, FRANCE

INTRODUCTION The Vertical-Flow Constructed Wetland (VFCW) is well known to have a good efficiency for organic load removal and nitrification. However, its phosphorus (P) retention capacity is still questioned (Brix et al., 2001). Indeed, the release of phosphates into sensitive aquatic ecosystems may promote eutrophication. The French company SCIRPE has developed a compact VFCW process called Azoé® (patent EP1857419A1 – Michel, 2007). The process line consists of: (1) a biological aerobic trickling filter as a pre-treatment stage of the organic load, (2) ferric chloride (FeCl3) addition for P treatment and (3) two stages of partially saturated vertical-flow constructed wetlands (psVFCW). In this process, dissolved phosphates are precipitated by injection of FeCl3 at the output of the trickling filter. The precipitates thus formed, mixed with suspended solids, progressively accumulate as sludge on the surface of psVFCWs. The deposit layer is removed by dredging every 10 to 15 years. Before final evacuation, the sludge layer may undergo various bio-physico-chemical changes depending on the environmental conditions (pH, Eh, etc.). The objective of the present study was to better understand the P retention/release dynamics within the deposit layer of psVFCW by pilot-scale experiments. Moreover extreme conditions of water saturation levels could be easily simulated. METHODS Description of pilot-scale design and operation Pilot-scale units simulating the different operational units of a real scale plant (excluding the second stage filter) were designed to treat a hydraulic load of 1 m3·day-1. Total volume of TF pilot was 0.29 m3. It was filled with plastic matrix media for development of bio-film. The VFCW pilot had a surface of 2 m2 and a depth of 0.7 m. It is a core sample of the real scale plant in operation since 2004 (Vercia) and equipped with redox probes placed respectively at 10 cm, 30 cm and 50 cm from the bottom of the filter. A layer of 10 cm of sludge sampled from the Vercia plant was then spread on the surface of the filter. The pilot was then operated by alternating feeding stages of 1/3 of week and resting stages of 2/3 of week in order to simulate the operational conditions of real scale systems. In order to study the effects of redox conditions on P retention, three different watersaturation levels were investigated: (1) unsaturated phase (free water drainage, i.e. contact with air possible from the top and the bottom), (2) partially saturated phase (contact with air only possible from the top) and (3) total saturation phase (i.e. even the sludge layer was flooded).

ABSTRACTS - WETPOL 2013 - October 13-17, 2013 - Nantes - FRANCE 241

MONITORING AND ANALYSES Samples of solutions were taken at the beginning and the end of each feeding period from four different points between each treatment steps. Each sample was analysed for the following parameters according to French standard methods (AFNOR): particle-size distribution, suspended solids, dissolved chemical oxygen demand, ammonium and nitrate. Moreover, total P and dissolved reactive P (DRP) were determined. Total elemental concentrations (Fe, Ca, Al, Mn and Si) were determined by ICP-AES. RESULTS AND DISCUSSION The oral presentation will focus on presenting the mechanisms implied in the stability of Fe-P precipitates (precipitation conditions, precipitates evolutions inside of the system, etc.). The first experimental results revealed the influence of the level of water saturation on the redox potential changes within the psVFCW, and the release of dissolved reactive P. The key observations at the different water-saturation levels were: 1) Unsaturated phase: the filter was under aerobic condition with redox potentials ranging between 400 and 670 mV (with respect to standard hydrogen electrode). The maximum concentration of dissolved reactive phosphorus (DRP) was 0.7 mg of P/L. 2) Partially saturated phase (same as the Vercia plant): the redox measured at the bottom of the filter within the saturated zone was around -250 mV indicating anaerobic conditions. In the unsaturated upper part of the filter the redox remained between 400 and 670 mV as previously, thereby confirming aerobic conditions. An excellent treatment performance was observed during this experimental period for organic matters (SS, dissolved COD). Denitrification was also quite significant. On the other hand, the release of P increased (DRP concentrations between 0.7 and 2.3 mg of P/L). 3) Total saturation phase: redox values in the entire depth of the filter were around -260 mV, indicating that the whole system was under anaerobic conditions. The treatment performances were degraded for all parameters. The concentration of DRP increased drastically (5.2 – 7.3 mg of P/L). This phase represent the extreme condition which could probably arrive at real treatment plant in case of a long period of heavy rain. CONCLUSIONS This study clearly highlighted the influence of the water-saturation level within the psVFCW on the release of dissolved reactive phosphorus and on the global treatment performance. ACKNOWLEDGEMENTS The authors thank Irstea Lyon and SCIRPE for the pilots that they kindly provided. Financial support for this work was provided by SCIRPE and CIFRE. REFERENCES Brix, H., Arias, C. A. and del Bubba, M. (2001) Media selection for sustainable phosphorus removal in subsurface flow constructed wetlands. Water Science and Technology. 44(1112):47-54. Michel, P. (2007) Process and installation of effluent treatment. EP 1 857 419 A1.

ABSTRACTS - WETPOL 2013 - October 13-17, 2013 - Nantes - FRANCE 242

Phosphorus and Iron Dynamics in Constructed Wetlands Domestic Treating Sewage in Small Works (O.141) Gabriela Dotroa,b, Raul Prieto Forta, Jan Baraka, Bruce Jeffersona, Peter Valeb a

Cranfield University, College Road, Cranfield, Bedfordshire, MK43 0AL, United Kingdom

b

Severn Trent Water, Severn Trent Centre, 2 St John’s Street, Coventry, West Midlands, CV1 2LZ, United Kingdom ([email protected], [email protected])

INTRODUCTION With the advent of tighter environmental regulations, phosphorus removal has become an increasingly growing area of research, as even small sewage works ( 6 mg/L, whilst anaerobic reactors remained undisturbed during the study. Soluble orthophosphates and total iron release rates were calculated at linear changes in the P and Fe mass in the supernatant as a function of time divided by the area of the sediment in the reactor. RESULTS AND DISCUSSION Results indicate that wetlands retain particulate phosphorus and iron effectively, removing 94% and 83% of the influent particulate fraction respectively, enabling the works to meet ABSTRACTS - WETPOL 2013 - October 13-17, 2013 - Nantes - FRANCE 243

their current discharge consents. Dissolved and colloidal fractions were found to pass through the wetland (Figure 1). Phosphorus release in the dissolved phase was observed in three wetland systems and further studied in laboratory reactors.

Fig. 1. Phosphorus and iron fractions in influent and effluent of example wetland full-scale site.

Quantities of phosphorus and iron release were significantly different in the aerobic and anaerobic reactors, with steady state final concentrations of 2.9 mgFe/L and 53 mgFe/L, respectively, achieved after 48 hours. Phosphorus release in the aerobic reactor was higher than in the anaerobic reactor, reaching P concentrations in the top water of 9.5 mg/L and 0.3 mg/L, respectively. The anaerobic phosphorus release rates were comparable to that reported by Lai and Lam (2008) in eutrophic marshes under anaerobic conditions and to that observed by Pant and Reddy (2003) in a constructed wetland. However, phosphorus release under aerobic conditions was unexpected and illustrates the complex dynamics of these pollutants. CONCLUSIONS Tertiary wetlands can retain particulate iron and phosphorus, storing it as sludge. Unfortunately, trapping of these pollutants is partially reversible and under anaerobic conditions, such as those observed in mature surface sludge layers, can result in phosphorus release in dissolved form. In laboratory reactors, both phosphorus and iron were released under anaerobic conditions but only P was released in full scale systems. In addition, iron release rates were orders of magnitude greater in the laboratory reactors when compared against full scale systems. This suggests P and Fe dynamics are significantly influenced by multiple factors which are variable within full scale wetland systems and thus, behave differently from a steady state reactor. Further research is underway to better understand pollutant dynamics and management strategies to enable tertiary wetlands to permanently retain phosphorus and iron in chemically-dosed sewage works. REFERENCES Brix, H, and Arias, C. (2005) The use of vertical flow constructed wetlands for on-site treatment of domestic wastewater: New Danish guidelines. Ecol. Eng. 25(5): 491-500 Lai, D.Y.F., Lam, K.C. (2009). Phosphorus sorption by sediments in a subtropical constructed wetland receiving stormwater runoff. Ecol. Eng.35: 735–743. Malecki, L.M., White, J.R., Reddy, K.R. (2004) Nitrogen and phosphorus flux rates from sediment in the Lower St. Johns River Estuary. J. Environ. Qual. 33: 1545–1555. Prieto Fort, R. (2013) Assessment of tertiary reed beds in chemically-dosed wastewater treatment plants for phosphorus removal. MSc thesis. Cranfield University p.96. Pant, H.K., Reddy, K.R. (2003) Potential internal loading of phosphorus in a wetland constructed in agricultural land. Water Res. 37: 965–972.

ABSTRACTS - WETPOL 2013 - October 13-17, 2013 - Nantes - FRANCE 244

Experimental steel slag filters to upgrade phosphorus removal from small wastewater treatment plants (O.160) D. Meyer c,b, H. Rustige a, F. Chazarenc b a

AKUT Umweltschutz Ingenieure Burkard und Partner, Wattstr. 10, D-13355 Berlin, Germany ([email protected]) b L’UNAM Université, Ecole des Mines de Nantes, CNRS, GEPEA UMR 6144, 4, rue Alfred Kastler. B.P. 20722 F-44307 Nantes, France ([email protected]) c IRSTEA centre de Lyon (formerly Cemagref), 5 rue de la Doua - CS70077, 69626 Villeurbanne, FRANCE ([email protected]) INTRODUCTION Phosphorus (P) removal requirements for treated wastewater discharge increased due to recent EU-legislations. Small treatment plants - such as constructed wetlands (CWs) or small activated sludge systems - are usually not able to ensure higher effluent qualities. One appropriate P-removal technology for treatment plants under unregulated operation appears to be the use of additional reactive filters. Previous international studies demonstrated potentials of steel slag, a by-product of the steel industry (Vohla et al., 2011). The European research project SLASORB integrates different scales and aims of slag investigations. Slag samples of type BOF (basic oxygen furnace) and EAF (electric arc furnace) were tested in batch and column tests to select proper materials (Barca et al., 2012, 2013b; Meyer et al., 2012). Afterwards full scale pilot filters in France (Barca et al.; 2013a) and Germany (this abstract) started operation. Further investigations targeted valorisation of the retained P as an agricultural fertilizer and finally economical aspects. METHODS The experimental site is located at Kappe, a small village north of Berlin. A sequencing batch reactor (SBR) serving 190 p.e. is situated next to a natural small stream, which is classified as sensitive according to European frame water directive. To prevent the prototype from frost a heated enclosure for the slag reactors was built. Two different kinds of slag and three different kinds of flow were tested in 6 reactors: BOF slag (8-32 mm) and EAF slag (5-10 mm). The slag volume of each cylindrical reactor is 0.31 m³ (height = 0.7 m). Three different kinds of flow were tested: fluctuating water level (fill-anddrain), ponded downstream and ponded upstream conditions. All reactors are equipped with pumps and valves for experimental operational control. Before slag filter outlets were discharged they had to pass an acidification step (CO2 bubble reactors). Technical CO2 and atmospherical CO2 was tested. Daily loads were regulated as needed for experiments. Because 8 hours of hydraulic retention time (HRT) seemed to be efficient after pre-testing, this value was used for a short period of time in comparison to the standard of 24 hours. Inflow and outflow water samples were taken on a weekly basis and analysed on standard wastewater parameter (COD, NH4-N), total phosphorus, phosphate, pH, conductivity and temperature. In addition to water quality control, tracer tests were conducted after 14 and 32 weeks of operation: fluorescein was added directly into the inlet of each reactor. The research period started accordingly to system operation launch in May 2011 and lasted for 15 month. RESULTS AND DISCUSSION The hydraulic behaviour of the filters was strongly influenced by several accidental sludge releases from the unregulated SBR. Especially the ponded upstream reactors suffered from sludge accumulation and correlated shortcut flows – it had to be taken out of operation. In ABSTRACTS - WETPOL 2013 - October 13-17, 2013 - Nantes - FRANCE 245

contrast, the opposite flow direction showed the highest hydraulic efficiencies which seemed little influenced by clogging. During the tracer test it could visually be observed, that the tracer gets mixed with the water in the ponding zone and passes the slag afterwards. The filland-drain containers also showed high hydraulic efficiencies, but not as much as assumed – most likely due to the sludge deposit. P-removal performances were strongly depending on the HRT. Within 8 hours sufficient retention could not be achieved in practice, 24 hours seemed to be a reasonable minimum. Efficiencies were also varying due to the used slag, the hydraulic conditions and also to highly fluctuating inflow concentrations. In summary, higher hydraulic efficiencies and longer reaction times lead to higher P-removal rates. The EAF slag showed higher mass retentions than the BOF slag in direct comparison of equal reactor operations. The maximum retained value of 0.4 g/kg slag (EAF) under break through concentration (P < 2 mg/l) seems to be relatively low. In general, removal performances were lower than in the corresponding column experiments and lower than in the pilot scale filters in France – most probably due to differing water quality compositions Acidification of the outflow by the use of atmospheric CO2 pressure (air) caused technical problems by foam production. Technical CO2 showed to be very effective and easy to handle. CONCLUSIONS Highest efficiencies could be reached in permanently ponded reactors with flow from top to bottom. This is easy to construct and easy to operate and is recommended for up-scaling. Hydraulic retention times of not less than 24 hours up to 48 hours is proposed. In practice the accidental occurrence of sludge from technical treatment systems has to be considered. For use with subsurface flow constructed wetlands this is not critical. The smallest used grain size 5-10 mm did not show an increased clogging behaviour. It is not clear whether the double specific surface of EAF slag compared to BOF slag may be the reason for better phosphorus performance. In order to achieve a maximum P storage within the slag cascading reactors may be used in practice. At least to filter beds may be used in a row with changing sequence. For small treatment plants the use of big bags is proposed for easy reloading with fresh material. ACKNOWLEDGEMENTS The research leading to these results has received funding from the European Union's Research Fund for Coal and Steel (RFCS) research programme under grant agreement n° RFSP–CT–2009–00028 (SLASORB). REFERENCES Barca C., Gérente C., Meyer D., Chazarenc F., Andrès Y. (2012) Phosphate removal from synthetic and real wastewater using steel slags produced in Europe. Wat. Res. 46:2376-84. Barca C., Troesch S., Meyer D., Drissen P., Andrès Y., Chazarenc F. (2013a) Steel slag filters to upgrade phosphorus removal in constructed wetlands: two years of field experiments. Environ. Sci. Technol. 47:549-56 Barca, C., Meyer, D., Troesch, S., Gérente, C., Andrès, Y., Chazarenc, F. (2013b) Steel slag filters to upgrade phosphorus removal in constructed wetlands - main factors controlling removal efficiency. this issue Meyer, D., Barca, C., Drissen, P., Rex, M., Andrès, Y., Chazarenc, F. (2012). Slag filters for upgraded phosphorus removal in constructed wetlands: two years of field experiments. Proceedings 13th Int. Conf. on Wetland Systems for Wat. Pol. Cont., Perth, Australia Vohla C., Kõiv M., Bavor J., Chazarenc F., Mander U. (2011) Filter materials for phosphorus removal from wastewater in treatment wetlands - A review. Ecol. Eng. 37:70-89

ABSTRACTS - WETPOL 2013 - October 13-17, 2013 - Nantes - FRANCE 246

S and COD cycle in wetlands

ABSTRACTS - WETPOL 2013 - October 13-17, 2013 - Nantes - FRANCE 247

Evaluation of sulfur turnover in different types of pilot-scale constructed wetlands treating domestic wastewater (O.24) Rania A. B. Saada, Peter Kuschka, Jaime Nivalab, Heinz Köserc a

Department of Environmental Biotechnology and bCentre for Environmental Biotechnology (UBZ), Helmholtz Centre for Environmental Research – UFZ, Permoserstraße 15, Leipzig, D-04318, GERMANY ([email protected] - [email protected] - [email protected])

c

Department of Environmental Engineering, Martin-Luther-University Halle-Wittenberg, Geusaerstraße 135, Merseburg, D-06217, GERMANY ([email protected])

INTRODUCTION Only limited research was dedicated to the sulfur transformation processes in constructed wetlands (CWs), despite their importance. It is evident that the sulfur transformation processes interact with the other contaminant removal processes and alter the overall performance of the CWs (e.g Wießner et al., 2010). From real-scale systems, sulfur deposits as white patches in the outflow were documented for the CWs Morina (Czech Republic) whose occurrence correlated with low NH4+ removal; which indicates the impact of sulfur transformations on CWs treatment efficiency (Vymazal, 2005; Sturman et al., 2008). Task of our research was to characterize the changes in inorganic sulfur pools associated with the treatment of domestic sewage in different types of CWs. Such a comparison of different CW technologies has not been conducted in the past. METHODS The research facility is situated in Langenreichenbach (LRB) in Saxony, Germany. It contains 15 individual pilot-scale subsurface flow CWs, representing 8 different design variants, between vertical, horizontal and reciprocating flow systems. All the systems receive pre-treated (septic tank) wastewater from the adjacent villages (Nivala et al., in press). 14 systems (7 pairs of unplanted and planted with Phragmites australis horizontal and vertical flow CWs) were sampled in the period April 2012 – March 2013 at inflow and outflow points. The main influent wastewater characteristics are summarized in Table 1. Table 1. Mean influent wastewater characteristics

Average Standard deviation Count (n)

BOD5 (mg/L)

TSS (mg/L)

240 74 65

140 59 66

Total Nitrogen (mg/L) 72 16 66

Redox potential (mV) -148 78 66

SO42--S (mg/L) 72.6 13.6 14

S2(mg/L) 8.6 3.6 19

S0 (mg/L) 5.2 2.3 9

RESULTS AND DISCUSSION The results are shown for 5 pairs (VG & VGp: vertical flow systems, unsaturated, gravel media; VA &VAp: vertical flow, saturated, gravel, active aeration; HA & HAp: horizontal flow, saturated, gravel, active aeration; H25, H25p, H50 & H50p: horizontal flow, saturated, gravel, depth of 25 & 50 cm; p refers to the planted bed of each pair). All vertical beds and the horizontal aerated beds (‘HA’ pair) hosted high redox levels throughout the duration of sampling, and the sulfate concentrations and loads did not change significantly by passing through these systems. In addition, the influent sulfide and elemental sulfur (Table 1) were oxidized in these systems and were not detected in the outflow. On the contrary, the unaerated horizontal flow systems (H25 and H50 pairs) accommodated low redox states (data not shown). In these systems, the sulfate concentrations and loads ABSTRACTS - WETPOL 2013 - October 13-17, 2013 - Nantes - FRANCE 248

decreased from inflow to outflow, and concentrations of the reduced species sulfide and elemental sulfur were higher in the outflow than inflow; indicating -though no investigation of microbial communities was conducted- the occurrence of microbial dissimilatory sulfate reduction and the subsequent re-oxidation of sulfide. 350 300

Sulfate (mg/L)

250

Inflow

200 150 100 50 0

Fig. 1. Concentration of sulfate in the inflow (horizontal dotted line) and the outflow of 5 different pairs of horizontal and vertical flow CWs. Striking differences of sulfur transformations depending on CW technology, HF CWs showed the highest SO42- turnover

CONCLUSIONS Clear shift from oxidized to reduced sulfur species in the horizontal unaerated systems, whilst an opposite trend prevailed in vertical and aerated systems. The main factor found to be dictating the inorganic sulfur transformations in CWs is the redox potential of the system. ACKNOWLEDGEMENTS Utmost gratitude is due to the German Federal Ministry of Education and Research ‘BMBF’, International Bureau, IPSWaT program; and the Helmholtz Interdisciplinary Graduate School ‘HIGRADE’ for the sponsorship of this PhD research. REFERENCES Nivala, J, Headley, T, Wallace, S, Bernhard, K, Brix, H, van Afferden, M, Müller, R (in press): Comparative analysis of constructed wetlands: The design and construction of the ecotechnology research facility in Langenreichenbach, Germany. Ecol. Eng. 2013 Sturman, P.J., Stein, O.R., Vymazal, J, Kröpfelova, L (2008): Sulfur cycling in Constructed Wetlands. In: J. Vymazal (ed.) Wastewater Treatments, Plant Dynamics and Management in Constructed and Natural Wetlands. © Springer Science + Business Media B. V. 2008 pp. 329-344 Vymazal, J and Kröpfelova, L (2005): Sulphur deposits at the outflow from horizontal subsurface-flow constructed wetlands – what does it indicate? In: Proceedings of the International Symposium on Wetland Pollutant Dynamics and Control, pp. 200-201. Ghent, Belgium: University of Gent Wießner, A, Rahman, K. Z., Kuschk, P, Kästner, M, Jechorek, M (2010): Dynamics of sulphur compounds in horizontal sub-surface flow laboratory-scale constructed wetlands treating artificial sewage. Water Research 44: 6175-6185

ABSTRACTS - WETPOL 2013 - October 13-17, 2013 - Nantes - FRANCE 249

Relation between Oxygen and E. coli in a Constructed Wetland and a Pond (O.154) S. Rühmland, M. Barjenbruch, R. Richter Technische Universität Berlin, Department of Urban Water Management, FG Siedlungswasserwirtschaft, Sekr. TIB 1-B16, Gustav-Meyer-Allee 25, 13355 Berlin, GERMANY www.siwawi.tu-berlin.de ([email protected]) INTRODUCTION Constructed wetlands (CW) and stabilisation ponds have a certain capacity to disinfect wastewater. If the elimination mechanisms (hostile environment, sedimentation, filtration, UV-radiation, predation) were identified, design and operation of constructed wetlands could be optimised for pathogen removal. Previous studies have found a relation between disinfection performance and oxygen content of the water (Schreijer et al., 1997). The objective of the current study is to get a deeper understanding of the favourable oxygen conditions for disinfection. METHODS The treatment plants are located on former sanitary landfills on the outskirts of Berlin, Germany. They are fed with the same effluent from a large conventional wastewater treatment plant with nutrient removal. Nitrification process is completed at the influent of the constructed wetland and the pond. Cases examined were: • SSF: sandy subsurface flow wetland, continuous vertical-flow, planted with Phragmites australis, water level constantly above filter bed, 1,450 m²; mean hydraulic loading 90 mm/d. • pond: stabilisation pond, 1,550 m²; mean hydraulic loading 110 mm/d. Both treatment plants have a good denitrification performance reducing NO3-N from 8.7 mg/L at the influent to 1.9 (SSF) and to 4.3 mg/L (pond) at the effluents. Grab samples were taken three times a day on two successive days between 2007 and 2012. Each sampling day resulted in one data point. According to the EU Bathing Water Quality Directive, E. coli was determined with the miniaturised MPN-method (ISO 9308-3, 1999). RESULTS AND DISCUSSION The influent contained a median concentration of 3.9 log E. coli/100 mL. The number of E. coli was reduced to 2.1 for unplanted pond and to 2.0 for sandy SF. Statistical comparison with temperature, season and flow rate did not show any correlations (Rühmland and Barjenbruch, accepted article). Moreover, the redox potential and phosphorus concentrations in the outlet showed only weak correlations with E.coli. The results of the correlations with oxygen differ with the type of treatment plant: 1) The sandy SF has generally very low oxygen concentration at the effluent (mean = 0.6 mg/L). Low oxygen concentration and low number of E. coli occur simultaneously (s. figure 1). Oxygen concentrations above 0.7 mg/L have limited E. coli elimination. 2) The pond has more aerobic conditions (mean oxygen concentration = 6 mg/L). Best elimination rates were measured at medium oxygen concentrations (see figure 2). An optimal oxygen range for disinfection was identified. As long as oxygen was in the range between 0.4 and 5 mg/L (equivalent to log 10-0.4 to log 100.7), E. coli was lower than 32 MPN / 100mL at the effluent. This outcome is contrary to the earlier

ABSTRACTS - WETPOL 2013 - October 13-17, 2013 - Nantes - FRANCE 250

observations that high oxygen concentrations above 5 mg/L enhance the removal of pathogens (Schreijer et al., 1997). Other parameters such as nitrate concentration also shows a positive correlation with the number of E. coli, especially in summer (SSF: log E. coli = 0.4cnitrate+1.6, R²=0.92). However, this positive relation does not apply for the pond. 3.5

log (E. coli [MPN/100 mL])

3.0

summer winter

y = 1.4x + 2.7 R2 = 0.73

2.5 2.0 1.5 1.0 0.5 0.0 -1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

log (oxygen [mg/L])

Fig. 1 Relation between oxygen concentration and number of E. coli in the outlet of the sandy subsurface flow wetland 3.5

log (E. coli [MPN/100 mL])

3.0

y = 0.8x2 - 0.3x + 1.4 R2 = 0.56

2.5 2.0 1.5 1.0 0.5 0.0 -1.5

summer winter

-1.0

-0.5

0.0

0.5

1.0

1.5

log (oxygen [mg/L])

Fig. 2 Relation between oxygen concentration and number of E. coli in the outlet of the pond

It is assumed that both treatment plants work best at their normal operating conditions. Divergences of the optimal disinfection performance are indicated by extreme oxygen conditions. There is no lower measureable limit for oxygen in the SSF. Different oxygen conditions lead to the same disinfection performance. That allows two alternative conclusions: a) The elimination mechanisms in the pond are facilitated by medium oxygen concentrations. Therefore, these removal mechanisms of pathogens are different than those in the SSF where lower oxygen concentrations are best. b) There is no causal correlation between the abundance of oxygen and the hygienic quality of the water. The oxygen concentration at the outlet only indicates the state of the current operation conditions. REFERENCES Rühmland, S.and Barjenbruch (accepted) Disinfection capacity of seven constructed wetlands and ponds. Water Science and Technology. Schreijer, M., Kampf, R., Toet, S., Verhoeven, J. (1997) The use of constructed wetlands to upgrade treated sewage effluents before discharge to natural surface water in Texel island, the Netherlands – pilot study. Water Science and Technology 35(5):231-237. ABSTRACTS - WETPOL 2013 - October 13-17, 2013 - Nantes - FRANCE 251

A large-scale constructed wetland for managing 95,000 m3/day of production water from the Nimr oilfield, Oman (O.167) Tom Headleya, Stephane Prigenta, Roman Breuerb, James Johnstona, Badar AlSharjic a b

BAUER Nimr LLC, P.O. Box 1186, Postal code 114, Al Mina, SULTANATE OF OMAN ([email protected], [email protected], [email protected])

BAUER Resources GmbH, In der Scherau 1, 86529, Schrobenhausen, GERMANY ([email protected]) c Petroleum Development Oman LLC, P.O. Box 81, Postal code 100, Muscat, SULTANATE OF

OMAN ([email protected]) INTRODUCTION Oilfield production water, or “produced water”, refers to the large quantities of water that are brought to the surface during oil extraction operations and which remain after the bulk of the crude oil has been separated. Produced water is typically saline (brackish to hypersaline), depending on the hydrogeological conditions in the oil reserve, and contaminated with residual hydrocarbons that remain after the oil-water separation process. In most oil fields more water is extracted from the ground than oil. The water to oil ratio can range from 1:1 up to 20:1 depending on the quality and age of the oil reserve. In land-based oil fields, produced water is typically disposed of via re-injection into the oil reserve, or into deep or shallow aquifers. Owing to the vast quantities of produced water involved, the depth of injection and the quantity of residual oil remaining in the water, below-ground disposal of produced water is generally a costly, energy intensive and environmentally damaging aspect of the oil production process. Hence, there is a growing interest from oil companies to develop more sustainable and cost-effective methods for managing produced water. Seeking to become an innovative industry leader, Petroleum Development Oman (PDO), the national oil company of Oman, made a strategic decision to explore more environmentally sensitive methods for managing produced water from its Nimr oil field located in the southern desert region of Oman. In 2000, PDO established a constructed wetland pilot system to investigate the potential of wetlands to remove and degrade the hydrocarbons in produced water. Following the success of this initial pilot study, PDO released a tender in 2008 for the construction and operation of a large-scale produced water management facility based on wetland technology. The environmental remediation company, BAUER, was successful in winning that tender and set about designing and constructing the Nimr Water Treatment Plant (NWTP) in 2009 – 2010 under a build, own, operate and transfer model with a 20 year operations contract. Commissioning of the NWTP started in November 2010, when the first produced water started flowing through an initial set of surface flow wetlands. The system has subsequently been expanded over time until in September 2012 it reached a capacity of 95,000 m3/day, which it continues to treat to this day. The NWTP currently consists of 350 ha of surface flow wetlands which treat the produced water to remove the hydrocarbons remaining after oil-water separation. Some of the brackish treated water is reused for different purposes, while the remainder is evaporated in 350 ha of evaporation ponds. METHODS The Nimr oilfield is located approximately 700 km south of the capital, Muscat, in the Sultanate of Oman. The NWTP is located in the desert with high daily and annual temperature variations and low rainfall. The inflowing produced water (average of 95,000 ABSTRACTS - WETPOL 2013 - October 13-17, 2013 - Nantes - FRANCE 252

m3/day), with an average total oil and grease content of 250 ppm, passes through an oil-water separator consisting of hydrocyclones where approximately 250 barrels of oil are recovered per day. The water then flows into a 3 km long buffer pond from where it is distributed into the 350 ha surface flow wetlands consisting of nine parallel streams of four cells in series (Figure 1). The functions of the wetlands are to remove the remaining hydrocarbons and reduce water volume via evapotranspiration. Part of the effluent from the wetlands is reused for drilling water, irrigation or further polishing, while the remainder passes through a long series of ponds (350 ha) where the water is evaporated leaving a final salt residue. The NWTP operates as a zero-discharge system with no outflow of effluent into the environment. The entire treatment process operates via gravity without any need for electricity.

Figure 1. Layout of the NWTP in Oman.

Water quality is monitored through the system for salinity, Total Petroleum Hydrocarbons and physio-chemical parameters on a fortnightly basis using portable electrodes and a TPH analyser in an on-site laboratory. A more detailed set of parameters are analysed for on a quarterly basis. Weather data is monitored using a set of automatic weather stations and class-A evaporation pans. The volume of produced water entering the NWTP and flowing out of the wetlands is monitored using flow meters in order to compile a water balance for the system. RESULTS The average TPH concentration throughout the wetland system is shown in Figure 2. The Surface Flow wetlands have proven highly effective at removing hydrocarbons.

Figure 2. Average TPH concentration through the wetland system (4 cells in series) for 2012 and 2013 (n = 53).

Additional results relating to the water balance, salinity, nutrients, vegetation, operational experiences and habitat values (over 100 bird species identified to date) will also be presented.

ABSTRACTS - WETPOL 2013 - October 13-17, 2013 - Nantes - FRANCE 253