Irrigation of treated wastewater in Braunschweig ...

This is a postprint of the paper printet in Chemosphere 66 (2007) 894–904. The publishers version is available at: http://dx.doi.org/10.1016/j.chemosphere.2006.06.035.

Irrigation of treated wastewater in Braunschweig, Germany: An option to remove pharmaceuticals and musk fragrances Thomas A. Ternes1, Matthias Bonerz1, Nadine Herrmann1, Bernhard Teiser2, Henrik Rasmus Andersen3 1:

Federal Institute of Hydrology (BFG), D-56068 Koblenz, Am Mainzer Tor 1, Germany Abwasserverband Braunschweig, Celler Heerstraße 337, D-38112 Braunschweig, Germany 3: The Danish Technical University. Bygningstorvet 115, DK-2800 Kgs. Lyngby, Denmark 2:

Corresponding author: Dr. Thomas Ternes, telephone: ++49 261 1306 5560, Fax: ++49 261 1306 5363, e-mail: [email protected]

Ap

0.55 m

pathway Well 1

lysimeter C

lysimeterA

Field

14

Gr Fl oun ow dw d i ate re cti r on

Well 5

shield wind

B2 Road

I Well 2

II

lysimeter D

edge lysimeter

lysimeter B Well 3

medium sand

III

Dephts of the wells  12-15 m, saturated zone: 1.5-2 m

Lipid regulator antiphlogistics

6.0

12 m

March

Contrast media

September

4.0

Betablockers

3.0 2.0

Antibiotics

Musks

1.0

Iomeprol

Diatrizoate

Iopamidol

Tonalide

Galaxolide

Propranolol

Metoprolol

Sotalol

Atenolol

Carbamazepine

Clofibric acid

Diclofenac

Ibuprofen

Roxithromycin

Sulfamethoxazole

0.0 Trimethoprim

Concentration in µg/L

May 5.0

unsaturated zone

saturated zone

pathway

ing h

Well 6

1.60 m

wind shielding hedge

well 4

fine sand

Abstract In this study the fate of pharmaceuticals and personal care products which are irrigated on arable land with treated municipal wastewater was investigated. In Braunschweig, Germany, wastewater has been irrigated continuously for more than 45 years. In the winter time only the effluent of the sewage treatment plant (STP) of Braunschweig is used for irrigation, while during summer digested sludge is mixed with the effluent. In the present case study six wells and four lysimeters located in one of the irrigated agricultural fields were monitored with regard to the occurrence of 52 pharmaceuticals and 2 personal care products (PPCPs; e.g. betablockers, antibiotics, antiphlogistics, carbamazepine, musk fragrances, iodinated contrast media (ICM) and estrogens). No differences in PPCP pollution of the groundwater were found due to irrigation of STP effluents with and without addition of digested sludge, because many polar compounds do not sorb to sludge and lipophilic compounds are not mobile in the soil-aquifer. Most of the selected PPCPs were never detected in any of the lysimeter or groundwater samples, although they were present in the treated wastewater irrigated onto the fields. In the groundwater and lysimeter samples primarily the ICM diatrizoate and iopamidol, the antiepileptic carbamazepine and the antibiotic sulfamethoxazole were detected up to several µg·L-1, while the acidic pharmaceuticals, musk fragrances, estrogens and betablockers were likely sorbed or transformed while passing the top soil layer. Potential estrogenic effects are likely to disappear after irrigation, since the most potent steroid estrogens were not measurable. Key words: soil-aquifer treatment, estrogens, antibiotics, antiphlogistics, contrast media, musk fragrances Introduction The limited quantity of unpolluted water available for future use as a resource for food production and drinking water supply is one of the major challenges faced around the world, including Europe (Postel, 2000). Indirect reuse of treated wastewater can increase the water supply in areas in which the water demand by the urbanized population has exceeded the available natural water sources, becoming a limiting factor for economic (agricultural and industrial) requirements (e.g. Harhoff and van Merve, 1996; Vazquez et al., 1996; Crook et al., 1998; Lauer and Roger, 1996; Cuthbert and Hajanosz, 1999; Levine and Asano, 2004;; ). The fate of pharmaceuticals and personal care products (PPCPs) is rarely investigated during currently applied indirect reuse facilities (Drewes and Shore, 2001). Approximately 3000 different pharmaceutical ingredients are used in the EU today, including painkillers, antibiotics, antidiabetics, beta-blockers, contraceptives, lipid regulators, antidepressants, antineoplastics, tranquilizers, impotence drugs and cytostatic agents. As these compounds are frequently transformed in the body, a combination of unchanged pharmaceuticals and metabolites is excreted by humans. These pharmaceuticals are discharged from private households and from hospitals and eventually reach municipal STPs. Musk fragrances are used in high quantities in cosmetic products. In 1996 about 5600 t a-1 of polycyclic musk fragrances (PMF) were used world-wide (Geyer et al., 2000); the production of galaxolide (HHCB) was estimated to be 1000 t a-1. In recent years nitro musk fragrances have been successively replaced in some countries (e.g. Germany) by PMF. In contrast to pharmaceuticals, personal care products, such as PMF, do not pass through the human body. They enter the wastewater via their regular use during showering or bathing. Since pharmaceuticals, estrogens and musk fragrances are not totally removed during wastewater treatment (e.g. Ternes, 1998; Heberer, 2002; Simonich et al., 2002; Miao and Metcalfe, 2003; Andersen et al., 2003; Clara et al., 2004; Joss et al., 2005; ), they are discharged in appreciable quantities into receiving waters through treated wastewater effluents. The question arises whether PPCPs present in treated wastewater are removed by soil-aquifer treatment. Gray and Sedlak (2005) determined a removal of 17-estradiol (E2) and 17-ethinylestradiol (EE2) of 36% and 43%, respectively, in an engineered treatment wetland with dense macrophytes located in Riverside County, California, spiked these estrogens at conentrations with 2.4 mg·L-1 and 2.25 mg·L-1, respectively. Kinney et al. (2006) found that measurable but low concentrations of pharmaceuticals can be detected in soil irrigated with reclaimed water from municipal sewage treatment plants (STPs). The overall objective of this paper is to discuss the suitability of soil aquifer treatment as a tool within the indirect reuse scheme of municipal wastewater to remove PPCPs. Therefore, a case study in

Braunschweig, Germany, was launched where secondary treated sewage has been irrigated onto agricultural fields for more than 45 years. This rare case of long time irrigation of agriculture land in a temperate region is an ideal case study. Sorption and degradation for any recalcitrant compound would have come to equilibrium over the years. Furthermore, any adaptation of soil bacteria communities to degrade the chemicals have had ample time to develop. The soil in Braunschweig has a low water and nutrient retention capacity, since it is sandy with less than optimal clay and organic carbon content. Therefore, the cultured plants would suffer from drought and nutrient deficiency without watering and fertilising. Irrigiation is also a cheap possibility to add water and nutrients. As an unique situation in Braunschweig, digested sludge is mixed into the irrigation water. Sludge contains both organic matter which improves the water retention capacity of soil and the bulk of nutrients originally present in the sewage, mainly nitrogen and phosphate. Experimental section Sampling of STP Braunschweig

Water samples were taken as 24 h flow-proportional composite samples (cooled at 4°C) from the outlet of the grit removal tank and the secondary clarifier, collected at midnight (see Fig. 1). The hydraulic retention time (HRT) was not directly considered when sampling the STP inlet and outlet, since both were taken in the same 24 h period. Liquid samples were obtained by filtration through glass fibre filters (< 1 µm). The mechanical treatment of the current Braunschweig STP (population equivalent: 385000) consists of a screen, an aerated grit-removal tank and a primary clarifier (Fig. 1). The flow rate is about 60000 m3 d1 . The primary sludge collected in the primary clarifier (volume: 2900 m3) and the secondary sludge are pre-thickened and then further treated in a two stage anaerobic digester (first process: thermophilic, 55°C, 4 d retention time; second process: mesophilic, 37°C, 16 d retention time). The primary effluent is directed via mixing basins (volume: 2900 m3) to the activated sludge system for biological phosphate removal, nitrification and denitrification (volume: 51000 m3). After settling in the secondary clarifier (volume: 20800 m3), the activated sludge is returned to the inlet of the activated sludge tanks. The secondary effluent is irrigated on agricultural fields. The activated sludge system is operated with a solid retention time of 12-14 d, which is typical for a nitrifying plant with simultaneous denitrification. The HRT of the STP (calculated from the flow and the volume of the treatment tanks) varies between 25-35 h depending on seasonal variation and precipitation.

sampling Grit Screen removal

sampling Primary clarifier Mix tank

Denitr.

Nitrif.

Second. clarifier Second. effluent

Primary sludge

Secondary excess sludge

Water

Biogas

Agricultural fields Centrifuge

Centrifuge Digester

Digested Sludge March to November Spray Irrigation onto agricultural fields

Fig.1. Scheme of the water and sludge flux in the STP Braunschweig

Irrigation of the effluent and digested sludge

In 2003, approximately 15 Mio. m³ of secondary effluent were irrigated on agricultural fields (about 3000 ha). In the vegetation period from March to October digested sludge is mixed with the effluent prior to irrigation. About 1 g digested sludge is added to 1 L effluent. In the summer time up to 2,400 m³ h-1 and in the winter time up to 1,600 m³ h-1 were irrigated. Using a distribution net of about 100 km, the irrigation machines are fed with a effluent/sludge mixture. The mixture is irrigated to about 40 L·m-2 at each application. About 500 L·m-2 (in the winter time only effluents) are applied in total per year. When applying fertilizers, the farmers consider the input of the nutrients onto the agricultural fields by the effluent. For instance, the phosphate requirements of the crops is totally met by irrigation and thus there is no need to add further phosphate by fertilizers. In 2003, when the current case study took place, 264.7 t phosphorus and 412 t nitrogen were irrigated onto the agricultural fields. The nitrogen intake can be allocated to 142.4 t ammonia-N, 168.9 t org.-N and 100.7 t NO3-N. Due to hygienic aspects, it is not allowed in Braunschweig to produce fruits or vegetables such as strawberries or salad which are directly consumed without further refinement. Therefore, the main crops produced in the irrigation fields are cereals, potatoes and sugar beets. To avoid a loss of nitrogen after harvesting the agricultural goods, intermediate plants accumulating nitrogen such as mustard (Sinapis sp.) and rape (Brassica napus) are grown between main crop cultivation periods. The irrigation area (section I and II) monitored in this study contains 3.5 ha. It is a part of the whole irrigation area (about 3000 ha) and contains six wells and four lysimeters (see Fig. 2). Section I and III have been irrigated for more than 45 years, while section II has only been irrigated for 8 years. The irrigation area consists of a podzolic brown soil (cambisol, sandy soil) with an Ap horizon up to 0.55 m depths (Fig. 3). The parameters measured in the Ap horizon of section I and II are listed in Table 1. Below the Ap horizon mainly fine sand and medium sand can be found which are extremely water permeable. The water table varies 1-2 m below land surface. The saturated zone varies 1-2 m to 10-12 m. The HRT to reach the saturated zone is predicted to be less than one day.

pathway Well 1

lysimeter C

lysimeterA

Field

Well 5

Well 2

II

lysimeter D Well 6

hedg e lysimeter

lysimeter B Well 3

III

pathway

Gr Flo oun w dw d ir a t e ec r t io n

14

ding shiel wind

B2 Road

I

wind shielding hedge

well 4

Dephts of the wells  12-15 m, saturated zone: 1.5-2 m

Figure 2: Scheme of the irrigation field in Braunschweig with the included sampling sites investigated in the case study. I: section I; II: section II, III: section III.

Ap

0.55 m

fine sand 1.60 m

medium sand

unsaturated zone

saturated zone

12 m Fig. 3: Soil-aquifer column of lysimeter A used for irrigation.

Table 1: Hydrological parameters of section I and II

horizon

section I Section I section I section II section II section II 0-5 cm 20-25 cm 25-55 cm 0-5 cm 20-25 cm 25-55 cm pH 6.9 7.1 6.9 7.0 7.2 6.5 Clay in % 3.6 3.4 4.2 2.2 2.6 1.8 Silt in % 6.5 8.1 12.2 5.6 9.1 7.2 Fine sand in % 29.3 22.5 21.3 30.0 24.1 21.5 Medium sand in % 57.8 60.0 56.6 58.6 57.4 62.9 Coarse sand in % 3.4 4.9 4.8 3.4 4.1 4.8 Corg in % 0.61 0.55 0.70 0.60 0.19 0.14 CEC mmol kg-1 56 28 29 49 31 21 Water holding capacity in % 20 17.3 16.3 15.4 20.6 13.2 Kf in 10-4 m s-1 -1.2 1.0 -2.6 3.1 Porosity in % 36.7 38.8 37.8 43.7 43.7 41.4

Sampling of lysimeters and wells

In the current study water samples were collected from lysimeters (inner diameter: 80 mm; depths: 1215 m) of stainless steel located in the selected agricultural field (Fig. 2) at three different depths together with groundwater probes. In total 52 PPCPs (e.g. betablockers, antibiotics, lipid regulators, antiphlogistics, carbamazepine, musk fragrances, ICM and estrogens) were covered by the analytical methods. The lysimeters provide specific data where the irrigation quantity is known and where primarily a vertical water movement occurs. The wells covered a much broader irrigation area primarily influenced by a horizontal groundwater flow. For sampling, PTFE (Teflon) pipes were introduced into the monitoring wells 3 m below the groundwater table. Subsequently, the groundwater was pumped out without lowering the groundwater surface until temperature and conductivity were on a constant level (temperature: alteration below  0.1 K within 5 min; conductivity: alteration below 1% of the actual value within 5 min). In 2002, four sampling campaigns (5. Feb., 14. May, 18. Oct., 10. Sep.) were run from spring to the fall within one continuous vegetation period to investigate the infiltration of PPCPs into the groundwater and to elucidate differences in pollution due to irrigation of effluent with and without stabilized digested sludge. On 5.2.2002 only the STP effluent was irrigated, while on the 14.5.2002, 10.09.2002 and 18.10.2002 additionally digested sludge was mixed. The standard parameters of the irrigated effluent and the digested sludge are given in Table 2. The following samples were always analysed for PPCPs: raw sewage entering the STP, STP effluent used for irrigation, groundwater from six wells within the area of a high irrigation intensity, two lysimeters A and B at different depths (0.40, 0.80, 1.2 m) and due to less water volumes the two lysimeters C and D only at 1.2 m within the irrigation fields. Additionally, the digested sludge was analyzed for estrogens. For the other PPCPs, the analytical methods for sludge were not available in 2002 when the case study took place. All water samples were extracted as soon as possible, at least within three days after sampling. Table 2: Paramters of the effluent and digested sludge irrigated 05.02.2002 Effluent Flow rate m³ d-1 64,100 Irrigation rate m³ d-1 33,300 Hydraulic residence time h 28.5 in STP Suspended solids (SS) mg·L-1 10 pH 7.6 BOD5, filtrated mg·L-1 2.4 COD, non filtrated mg·L-1 33 Norg mg·L-1 2.4 ammonium-N mg·L-1 0.43 nitrate-N mg·L-1 14 nitrite-N mg·L-1 0.19 Phosphate mg·L-1 2.57 Cr mg·L-1 5 Zn mg·L-1 32 Cd mg·L-1 0.1 Pb mg·L-1 2 Ni mg·L-1 10 Cu mg·L-1 5 Hg mg·L-1 0.2 MgO mg·L-1 13.6 K2O mg·L-1 31.4 CaO mg·L-1 93.5

14.05.2002

10.09.2002

18.10.2002

69,800 43,800 26.1

54,100 46,300 33.7

53,200 45,600 34.3

330 7.3

156 7.3 9 246 12 8.7 7.7 0.73 19.1 11 241 0.6 16 10 59 0.3 18.7 31.2 124

336 7.5 10 342 20 12 7.21 2.04 29.8 11 340 1.3 15 12 107 0.34 18.9 36.2 114

326 18 11 8.31 0.15 30.9 6 353 0.9 19 10 109 0.4 20 29.8 131

Analytical methods

Several methods have been used for the determination of 52 PPCPs in the lower ng·L-1 range. Analytical methods used are described in the following literature: estrogens (Ternes et al., 1999; Andersen et al., 2003), antibiotics (Hirsch et al., 1999; Göbel et al., 2004), iodinated contrast media (Ternes and Hirsch, 2000), acidic pharmaceuticals (Löffler and Ternes, 2003 as well as neutral pharmaceuticals and

betablockers (Ternes et al., 1998) and musk fragrances (Ternes et al., 2004). The analysis of estrogens in sewage sludge can be found in Ternes et al. (2002). Calibration, limit of quantification (LOQ) and blank samples

Calibration over the whole procedure was performed with groundwater spiked with the analytes. A 710 point calibration curve was used, ranging from 10 ng·L-1 to 2 µg·L-1 (contrast media: 5 µg·L-1). The LOQ allowing for the quantification of analytes in native samples was calculated according to the German DIN method 32645 with the standard deviation of a linear regression curve for a detection range from 0.005 to 5 µg·L-1. The calculated LOQs were always higher than 10 times the signal/noise ratio of the extracts. Blank samples of water which are only spiked with the surrogate standard were included in each series of analysis. The analyte concentrations in native samples were calculated in relation to the surrogate standards spiked at the beginning to compensate losses caused by the presence of matrix ingredients. LOQs for estrogens can be achieved down to 0.5 ng·L-1 analyzing 1 L of groundwater and raw or treated wastewater. LOQs can be achieved for all other PPCPs down to 0.025 µg·L-1 for ground water and down to 0.10 µg·L-1 for treated wastewater analyzing a volume 500 mL and 200 mL, respectively; for raw wastewater only 100 mL were enriched and hence the LOQ were 0.20 µg·L-1. The LOQ for determing estrogens in freeze-dried sludge was 1.5 ng g-1. Results and discussion Occurrence of PPCP in the irrigated effluent and estimated removal in the municipal STP Braunschweig

PPCPs are permanently present in the raw sewage and the effluents of the municipal STP Braunschweig. Although the HRT of influent and effluent sampling was not directly considered the following conclusions can be drawn: Estimated removal of PPCPs (see Tables 3, 4)

Carbamazepine, diatrizoate and iopamidol were not significantly removed when passing through the municipal STP of Braunschweig. Obviously, these compounds are neither degraded, nor sorbed onto sewage sludge and might therefore appropriate to be used as wastewater indicators. These observations are consistent with those of several other studies (Heberer, 2002; Ternes et al., 1998; Ternes and Hirsch, 2000; Miao and Metcalfe, 2003; Kreuzinger et al., 2004). A major removal of higher than 90% was found for caffeine and ibuprofen. An appreciable removal was observed for many other PPCPs such as atenolol, metoprolol, propranolol, bezafibrate, trimethoprim, AHTN (tonalide) as well as for the ICM iohexol, iomeprol and iopromide. Although no differentiation between sorption and biodegradation was made, based on the physical-chemical properties and the literature data the elimination of caffeine, ibuprofen and the ICM should be mainly caused by biodegradation (see also Joss et al., 2006). The sludge retention time (SRT) plays an important role for the biodegradation of trace pollutants (Kreuzinger et al., 2004; Ternes et al., 2004). For instance, in the Braunschweig STP with a SRT of about 12-14 d a removal of three ICMs was observed, whereas in a previous case study with a SRT of about 4-6 d for the Wiesbaden STP no ICM removal occurred (Ternes and Hirsch, 2002). The estrogens (EE2, E1 and E2) were significantly removed from the water phase in the Braunschweig STP (see Table 4). However, since the effluent was mixed with digested sludge

Table 3. Concentrations (µg L-1) pharmaceuticals and musk fragrances in the influent and effluent of Braunschweig STP and wells in the irrigation area (LOQ: 0.025 µg·L-1). Mean result from the four sampling rounds is indicated with 95% confidence interval. Influent effluent Removal well 1 Well 2 well 3 well 4 well 5 well 6 Antibiotics Trimethoprim Sulfamethoxazole Clarithromycin Erythromycin Roxithromycin Antiepileptic Carbamazepine Antiphlogistics Ibuprofen Diclofenac Lipid regulators Clofibric acid Fenofibric acid Bezafibrate Betablockers Atenolol Sotalol Celiprolol Propranolol Metoprolol Musk fragrances AHTN (Tonalide) HHCB (Galaxolide) Contrast media Diatrizoate Iopromide Iomeprol Iohexol Iopamidol Caffeine

1.10.26 0.820.23 0.460.10 0.830.27 0.810.42

0.340.08 0.620.09 0.210.04 0.620.44 0.540.07

69 24 54 25 33

< LOQ 0.110.07 < LOQ < LOQ < LOQ

< LOQ < LOQ < LOQ < LOQ < LOQ


< LOQ 0.0460.020 < LOQ < LOQ < LOQ

< LOQ 0.0410.018 < LOQ < LOQ < LOQ

< LOQ 0.100.02 < LOQ < LOQ < LOQ

2.01.3

2.10.7

0

0.570.14

(0.030; 0.035)

< LOQ

0.330.08

0.0410.018

0.230.07

3.41.7 2.01.5

0.130.06 1.30.9

96 33

< LOQ < LOQ

< LOQ < LOQ

< LOQ < LOQ

< LOQ < LOQ

< LOQ < LOQ

< LOQ < LOQ

0.250.09 --

52

4.94.1

0.120.04 0.130.07 0.130.07

97

< LOQ < LOQ < LOQ

< LOQ < LOQ < LOQ

< LOQ < LOQ < LOQ

< LOQ < LOQ < LOQ

< LOQ < LOQ < LOQ

< LOQ < LOQ < LOQ

2.32.0 2.51.3 0.440.34 0.510.35 4.93.4

0.360.02 1.30.2 0.280.02 0.180.02 1.70.07

84 48 36 65 65







0.330.20 1.30.7

0.100.05 0.730.24

70 44

< LOQ < LOQ

< LOQ < LOQ

< LOQ < LOQ

< LOQ < LOQ

< LOQ < LOQ

< LOQ < LOQ

3.31.8 186 104 9.02.0 2.31.6 8236

3.31.7 3.01.7 1.10.9 1.00.3 1.90.5 0.220.06

0 83 89 89 17 99.7

5.71.8 < LOQ < LOQ < LOQ (0.011; 0.012) < LOQ

5.12.5 < LOQ < LOQ < LOQ < LOQ < LOQ

9.66.3 < LOQ < LOQ < LOQ (0.07; 0.11) < LOQ


4.11.6 < LOQ < LOQ < LOQ (0.066) < LOQ


(about 1 g TSS to 1 L water) prior to irrigation, reasonable concentrations of estrogens were irrigated. For instance, in total a sum of 37 ng·L-1 for E1 and E2 were measured in the irrigated slurry. EE2 is biodegraded more slowly in STPs (Joss et al., 2004; Ternes et al., 1999) and has a sorption constant for sludge similar to the natural estrogens (Clara, 2004b; Ternes et al., 2004; Andersen et al., 2005). Therefore, it might be possible that EE2 is added to the irrigation water with the sludge as well, but this could not be directly quantified, as the concentration of EE2 in sludge is lower than the limit of quantification of 1.5 ng·g-1. Table 4. Concentrations of estrogens in the influent and effluent (LOQ: 0.50 ng·L-1) and digested sludge (LOQ: 1.5 ng·g-1) of the Braunschweig STP. 95% confidence interval is indicated. Natural estrogen Influent Effluent Removal Sludge

E1+E2 EE2

ng·L-1

ng·L-1

3620 3.31.5

1.40.6 1.10.4

% 96 67

ng g-1 36