On-site infiltration of a copper roof runoff: Role of clinoptilolite as an artificial barrier material

ARTICLE IN PRESS WAT E R R E S E A R C H

41 (2007) 3251 – 3258

Available at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/watres

On-site infiltration of a copper roof runoff: Role of clinoptilolite as an artificial barrier material Konstantinos Athanasiadis, Brigitte Helmreich, Harald Horn Institute of Water Quality Control, Technical University of Munich, Am Coulombwall, 85748 Garching, Germany

art i cle info

ab st rac t

Article history:

On-site infiltration may be considered as a promising way of managing rainwater runoffs in

Received 11 December 2006

urban areas, provided the hydrological and ecological conditions allow infiltration, and

Received in revised form

provided there is adequate treatment of the contaminants to avoid a risk of soil and

20 March 2007

groundwater pollution. The aim of this study was to evaluate the feasibility of the

Accepted 11 May 2007

application of a new technical infiltration system equipped with clinoptilolite as an

Available online 23 May 2007

artificial barrier material for the treatment of the copper roof runoff of the Academy of Fine

Keywords:

Arts in Munich, Germany. During the 2-yr sampling period, 30 rain events were examined.

Infiltration

The cover material of the roof and the drainage system was responsible for the high copper

Clinoptilolite

concentrations in the roof runoff. The rain height and the rain intensity were of great

Copper

significance regarding the establishment of the copper runoff rate. The technical

Roof runoff

infiltration system applied was able to reduce the copper from the roof runoff by a factor

Rain intensity

up to 96%. The mean measured copper concentration in percolation water was lower than

Rain height

the critical value of 50 mg/l set by the German Federal Soil Protection Act and Ordinance, indicating no risk for soil and groundwater contamination. & 2007 Elsevier Ltd. All rights reserved.

1.

Introduction

Copper is, on the one hand, a natural component in most ecosystems and, on the other, a metal that has always found many applications in old and modern societies, e.g. in jewellrey and sculptural art, in plumbing, as a building material, in electronics, and in many other industrial artifacts. Nowadays, copper is extensively used as a roofing material, for facades and guttering. According to a report released in the year 2001 by the Metal Trade Association in Du¨sseldorf, a total amount of 888,000 t of copper, representing a surface of 153,500,000 m2, was assumed to be used in claddings of buildings in Germany since 1950 (Hullmann, 2003). A surface area of approximately 64.1 million m2 is exposed to atmospheric conditions and is subjected to corrosion processes. During a precipitation event, a part of the corrosion copper products formed will be retained on the roof surface (patina)

and a part will be released in the roof runoff. The copper runoff rate from copper roofs varies between 1.1 and 1.7 g/m2 yr, as reported by several research groups in related literature (Odnevall Wallinder et al., 2001; Karlen et al., 2002; Priggemeyer, 2003; UBA, 2005; Faller, 2003). According to a study of the German Federal Environmental Agency (UBA) in 2005, almost 85.2 t of copper is released in roof runoff waters every year in Germany U.B.A. (2005). Ekstrand et al. (2001), using the digital air photo processing for the mapping of copper roof distribution, estimated a yearly copper load of approximately 1200 kg for the city of Stockholm. Traditionally, copper roof runoffs are sent to sewers through which the rainwater is either directly transported to the receiving waters or, in case of combined sewer systems, sent to wastewater treatment facilities. Economically, this drainage system is far from being optimal because of the high costs for providing sewer capacity, which is rarely exploited. From an ecological point of view, the traditional urban

Corresponding author. GHD, 201 Charlotte Street, Brisbane, Old. 4000, Australia. Tel.: +61 7 3316 3277; fax: +61 7 3316 3333.

E-mail addresses: [email protected], , [email protected] (K. Athanasiadis). 0043-1354/$ - see front matter & 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.watres.2007.05.019

ARTICLE IN PRESS 3252

WAT E R R E S E A R C H

41 (2007) 3251– 3258

drainage concept is to be critically assessed as well, since it results in the following negative effects: (a) increase of runoff volume and runoff peak, (b) lowering of the groundwater table underneath the sealed urban area, (c) pollution of the receiving waters by direct discharge of copper-polluted storm waters or by sewer overflow, and (d) loading of activated sludge from wastewater treatment facilities with copper stemming from roof and road runoffs. Davis et al. (2001) reported that the roof contribution on the copper loadings of urban commercial storm water runoffs was up to 75%. The German Federal Environmental Agency estimated that about 69% of the annual released copper (85.2 t) in roof runoff is distributed in receiving waters and the remaining 31% in soil. Seventy-three percent of the copper released in soil originates in the utilization of activated sludge from wastewater treatment facilities for agricultural activities (UBA, 2005). On-site infiltration may be considered as a promising way of managing roof rainwater situations in urban areas, provided the hydrological and geological conditions allow infiltration, and provided the pollutants contained in the roof runoff water are effectively removed before the water enters the soil and the groundwater. Otherwise, the pollutants may accumulate in the soil leading eventually to highly contaminated sites, and they may contribute to deterioration of groundwater quality (Zimmermann et al., 2005; Ammann et al., 2003; Bucheli et al., 1998). As an alternative, it is proposed to pass the roof runoff water through an artificial barrier before it enters the soil and the groundwater (Steiner et al., 2006; Scholz, 2004). Clinoptilolite is a mineral zeolite that belongs to the Heulandite group. Its chemical formula varies depending on its composition and origin, but a typical representation is Na6((AlO2)(SiO2)30)  24H2O. Its characteristic tabular morphology shows an open rectangular structure of a three-dimensional framework with a negatively charged lattice. Exchangeable cations such as Na+, Ca2+, Mg2+, and K+ commonly balance this negative charge. Clinoptilolite, one of the most frequently studied natural zeolites, has a high selectivity for certain heavy metals such as Cu2+, Zn2+, Pb2+, Cd2+, and Ni2+ (Athanasiadis and Helmreich, 2005; Inglezakis et al., 2003; Rodriguez-Iznaga et al., 2002; Berber-Mendoza et al., 2006). It is used in many applications such as chemical sieve, gas absorber, feed additive, food additive, as an odor control agent, as a water filter for municipal and residential drinking water, and as a filter material in aquariums. It is well suited for these applications due to its large amount of pore space, high resistance to extreme temperatures, and chemically neutral basic structure. The main objective of this study is to evaluate the feasibility of the application of a new technical infiltration system equipped with clinoptilolite as an artificial barrier material for the treatment of the copper roof runoff of the Academy of Fine Arts in Munich, Germany. This paper also intends to determine the influence of weather parameters, such as rain height and rain intensity, on the performance of the infiltration system. Finally, the risk of soil and groundwater contamination, as well as the cost of installation and maintenance of the infiltration facility, is investigated.

2.

Experimental section

2.1. Description of the field site and the applied technical infiltration system The Academy of Fine Arts is located in the center of Munich (1,260,597 inhabitants), Germany. The building is surrounded by four major streets: Leopold street, Akademie Street, Tu¨rken Street, and Georgen Street. Leopold Street is one of the main roads in Munich, with a daily vehicular flow of 40,000 cars. The traffic load of the other streets is less than 2000 vehicles per day. The 4-yr-old copper roof covers a total area of 4800 m2. The monitored infiltration system, which dewaters the runoff of a 412 m2 roof surface at the south-west site, is composed of a filtration pit followed by an infiltration trench (Fig. 1). The roof runoff is channeled into the filtration pit (6), developed by HydroCon GmbH, tangentially. The filtration unit (8) (Athanasiadis et al., 2004) equipped with clinoptilolite as a barrier material is vertically seated over the hydrodynamic sediment separator (7). Since the barrier material could fail hydraulically under extreme rain weather conditions, an emergency overflow (4) is arranged, so that the runoff can reach the infiltration trench (9) without passing through the barrier material. The applied infiltration trench, developed also by HydroCon, is a special construction, which contributes to an additional reduction of copper concentration. The ditch consists of a partial seeping pipe of impermeable concrete in the lower part and of porous concrete with a high concentration of CaCO3, coated with iron hydroxide, in the remaining part. The clinoptilolite used (750 kg) was supplied by Silver & Baryte Ores Mining Co. S.A., Greece. The particle size of the sieved material was in the range of 0.7–2 mm. Clinoptilolite was chemically conditioned with 1 M NaCl solution at room temperature, before application (Athanasiadis and Helmreich, 2005).

2.2.

Sampling and analysis

Samples are taken from four different positions: (a) directly from the rain by means of a sampling device installed at the top of the roof; (b) from the inflow pipeline to the filter pit; (c) directly after the filtration unit; and (d) underneath the infiltration trench. pH values were measured in the laboratory with a glass electrode (WTW Sentix 60). The electrical conductivity (EC) readings were performed by an EC electrode (WTW TETRACON 325) connected to a WTW LF 340 EC meter. Total and dissolved copper was measured in acidified samples (1% Suprapur HNO3) by means of flame atomic absorption spectrometry. The dissolved fraction was obtained by filtration (0.45 mm membrane filter) prior to acidification.

3.

Results and discussion

3.1.

Quality of roof runoff

Thirty representative rain events were sampled and analyzed from the copper roof of the Academy of Fine Arts in Munich,

ARTICLE IN PRESS 3253

4 1 (200 7) 325 1 – 325 8

550 mm

WAT E R R E S E A R C H

600 mm

500 mm

2750 mm 800 mm

1000 mm 400 mm

1000 mm

900 mm

900 mm

200 mm 150 mm 130 mm

940 mm 420 mm

Fig. 1 – Technical infiltration system with monitoring devices: 1. control tank; 2. control unit; 3. flow meter; 4. overflow; 5. pressure transducer; 6. filtration pit; 7. hydrodynamic sediment separator; 8. filter unit; 9. infiltration trench; 10, 11, 12. samplers; 13. signal cables; 14. sampling pipe lines.

during the period March 2004–2006. The measured roof runoff volume, obtained through the flow meter recordings in the same sampling period, was almost 441 m3. The pH of rainwater varied between 6.6 and 8.0, with a mean value of 7.1 (pHmedian ¼ 7.1). Environmental atmospheric data, such as SO2, NO2, and O3 obtained during the same sampling period, were supplied by the Bavarian State Office for Environment (2005). The mean SO2, NO2, and O3 concentrations were 3.5, 42, and 40 mg/m3, respectively. Odnevall Wallinder et al. (2001) reported a mean pH value of 4.7 in rainwater collected from an urban site in the southern part of the city of Stockholm, with measured mean SO2, NO2, and O3 concentrations of 4.0, 23, and 63 mg/m3, respectively. Negrel and Roy (1998) measured pH levels ranging from 4.3 to 6.2 in rainwater samples collected in the Massif Central, in France. Reimann et al. (1997) reported pH levels ranging from 4.0 to 5.0 in rainwater collected in Finland, Norway, and Russia. The relatively alkaline pH values of the rainwater samples measured in this study are not due to the lack of acidity in precipitation but rather due to the neutralization of acidity with major cations, such as Ca, Mg, and K, stemming from the construction site next to the Academy of Fine Arts. Similar results have been reported by Mouli et al. (2005) for an urban site in Tirupati in southern India. The lowest pH value measured in the roof runoff water was 6.2 and the highest was 8.0. The mean pH value obtained was 6.8 (pHmedian ¼ 6.8). The pH of the runoff water after its passage through the clinoptilolite barrier material varied between 6.4 and 9.3, with a mean value of 7.2 (pHmedian ¼ 7.2). The increase of pH can be explained by the ion exchange of H+ ions with the exchangeable ions located on clinoptilolite and by the protonation of

neutral and negative surface hydroxyl groups of clinoptilolite (Doula et al., 2002). As was expected, different amounts of copper were washed off the roof surface during all rain events. In the rainwater, the copper concentration, stemming mainly from brake pad materials (Davis et al., 2001), varied between 10.0 and 230 mg/l; the mean value was 53.0 mg/l. After contact with the roof material, the concentration of copper in the roof runoff increased significantly and varied between 200 and 21,700 mg/l. The mean copper concentration in roof runoff was 1623 mg/l (Cumedian ¼ 1500 mg/l); thus, the cover material of the roof and the drainage system were responsible for the high copper concentrations in roof runoff. The phase distribution of the copper concentration in roof runoff was dominated by the dissolved phase, around 96% during winter and almost 85% during the remaining seasons of the year. The concentration and the phase distribution of copper in roof runoff are of prime importance with regard to its bioavailability after its release from the roof surface. Karlen et al. (2002) demonstrated that nearly all copper in runoff water, sampled directly after its release from the roof, was in dissolved phase, and thus bio-available and toxic toward the green algae Raphidocelis subcapitata. However, the toxicity of copper roof runoff water should not be simply extrapolated to effects on the environment since dilution processes and organic and inorganic complexations may alter its phase distribution and hence its toxicity on its way to the receiving water bodies. According to Mason et al. (1999), immediate percolation of roof runoff at the point of discharge to an infiltration pit, designed according to the normal standards for water infiltration sites, resulted in a minor retention of the metals

ARTICLE IN PRESS WAT E R R E S E A R C H

41 (2007) 3251– 3258

Cu, Zn, Pb, Cr, and Cd by the top organic-rich humus layer. Metals, in particular Cu, Cd, and Cr, were transported in soil water downward through the unsaturated zone and could eventually reach the groundwater (Mason et al., 1999). To identify the influence of rain intensity on copper runoff rate, the washed copper mass of the roof surface during a precipitation event was plotted against the mean rain intensity of the precipitation event (Fig. 2). The data demonstrate a linear correlation between the rain intensity and the amount of released copper. Rain events with high rain intensities generate higher copper runoff rates. Fig. 3 evaluates the influence of roof runoff volume, and thus rain height, on released copper mass. It can be seen, as by the rain intensity, that there is a linear correlation between rain height and released copper mass, and thus copper runoff rate. The slope of the regression analysis curve indicates, independently from the weather parameters, that there is a mean copper concentration in the roof runoff of 1610 mg/l. This result is in agreement with the measured mean copper concentration reported above, as well as with mean concentrations of copper roof runoffs reported in related literature (Persson and Kucera, 2001; Zobrist et al., 2000). The copper runoff rate obtained from the 2-yr exposure period was 0.83 g/m2 yr. A 4-yr exposure program performed by Leuenberger-Minger et al. (2002), in Switzerland, reported runoff rates of 0.4–1.7 g/m2 yr for copper. The lowest value was found for the least polluted site Davos and the highest values for the relatively polluted sites Cadenazzo and Ha¨rkingen. Odnevall Wallinder and Leygraf (2001) reported, for a 2-yr field exposure research, copper runoff rates ranging between 1.1 and 1.7 g/m2 yr for an urban site in Stockholm, and runoff rates ranging between 0.6 and 1.0 g/m2 yr for a rural site. All these variations are primarily attributed to differences in weather parameters, such as rain intensity and rain height, and roof characteristics, such as orientation, inclination, and age. Aspect, slope and height of the roof have a significant impact on the generated roof runoff volume

4 Roof runoff y=0.32143x r: 0.93

Copper mass (g)

3

2

1

0 0

2

4 6 Rain intensity (l/s ha)

8

10

Fig. 2 – Linear correlation between washed copper mass of the roof surface during a precipitation event and mean rain intensity of the rain event.

4 Roof runoff y=0.00161x r: 0.95

3 Copper mass (g)

3254

2

1

0 0

500

1000 1500 2000 Roof runoff volume (l)

2500

Fig. 3 – Linear correlation between released copper mass in the roof runoff during a rain event and roof runoff volume of the rain event.

(Maksimovic, 1996; Hollis and Ovenden, 1988; Davies and Hollis, 1981). Roofs facing the prevailing wind receive higher rainfall; thus they generate higher roof runoff volumes. South-facing roofs experience more evaporation than those facing east, north, and west (Ragab et al., 2003). In our study, the sampled roof surface has a southern aspect, with the prevailing wind coming from the west and therefore generated the low copper runoff rate of 0.83 g/m2 yr.

3.2.

Performance of the technical infiltration system

The copper concentration in the runoff after the filtration unit equipped with clinoptilolite as a barrier material varied between 12.0 and 980 mg/l, with a mean value of 102.8 mg/l (Cumedian ¼ 51.5 mg/l). The copper concentration in percolation water varied between 12.1 and 180 mg/l, with a mean value of 40.9 mg/l (Cumedian ¼ 29.0 mg/l). It can be seen that the application of the special infiltration trench seems to be necessary to achieve copper concentrations in percolation water lower than the critical value of 50 mg/l given by the German Federal Soil Protection Act and Ordinance (1999). The performance of the technical infiltration system regarding copper elimination, during the sampling period March 2004–2006 is demonstrated in Fig. 4. The copper loading of the technical infiltration system, of 710 g, was estimated by using the best linear fit of Fig. 3 (y ¼ 0.00161x) and the roof runoff volume recorded (441 m3) during the sampling period. The copper mass that was not removed by the hydrodynamic sediment separator and the clinoptilolite filtration unit during a sampled rain event was plotted against the runoff volume of the sampled rain event (Fig. 5). The regression estimate (best linear fit) was obtained, as in Figs. 2 and 3, by minimizing the sum of the squares of the residuals. The copper mass that was not removed by the hydrodynamic sediment separator and the clinoptilolite filtration unit, during the sampling period March 2004–2006, was 56.7 g and it was estimated by using the predicted line of

ARTICLE IN PRESS WAT E R R E S E A R C H

3255

4 1 (200 7) 325 1 – 325 8

0.30

710

Percolation water

Copper loading

55 597,4

Copper mass (g)

0.25

y=58.9 10-6x r: 0.76

0.20 0.15 0.10

30, 26

0.05 0.00 0

Fig. 4 – Performance of the technical infiltration system, regarding copper elimination, in the sampling period March 2004–2006.

1000 1500 2000 Roof runoff volume (l)

2500

Fig. 6 – Linear correlation between copper mass reaching percolation water during a rain event and runoff volume of the rain event.

Table 1 – Influence of the cement surface on the copper phase distribution

0.3 Runoff after the filtration unit y=128.5 10-6x r: 0.81 Copper mass (g)

500

Parameter

0.2

pH Total Cu2+ (mg/l) Dissolved Cu2+ (mg/l) Total Ca2+ (mg/l)

0.1

Copper solution 1 mg/l

2 mg/l

5 mg/l

10.0 980 5.7 4600

10.3 1980 7.8 4500

10.2 5100 2.2 5100

0.0 0

200

400 600 Roof runoff volume (l)

800

1000

Fig. 5 – Linear correlation between copper mass that was not removed by the filtration pit during a rain event and runoff volume of the rain event.

Fig. 5 (y ¼ 128.5  106x) and the recorded runoff volume of 441 m3 obtained during this sampling period. Fig. 6 illustrates the linear correlation between the copper mass reaching the percolation water during a rain event and the runoff volume of the rain event. It was found that almost 26.0 g of copper reached the percolation water during the sampling period March 2004–2006. To determine the contribution of clinoptilolite on the retention capability of the technical infiltration system regarding copper elimination, a representative sample of the clinoptilolite barrier material was taken at the end of the sampling period, and was regenerated in a discontinuous modus in the laboratory with 1 M NaCl solution (Athanasiadis, 2005). It was found that only 55.9 g of copper was sorbed on clinoptilolite (0.075 g Cu2+/kg clinoptilolite) during this sampling period. The performance of the technical infiltration system, including the infiltration trench, was 96.3% in respect to copper elimination. Ninety-two percent of the copper loading was retained by the hydrodynamic sediment separator and the clinoptilolite filtration unit. Only 8.6% of this retained

copper was removed by the clinoptilolite barrier. The remaining copper mass was supposed to be eliminated by sedimentation as CuCO3 or as Cu(OH)2. As was mentioned above, in Section 3.1, the copper concentration in roof runoff was dominated by the dissolved phase. This means that the cement surface of the hydrodynamic sediment separator (0.96 m2) plays a significant role in the transformation of the copper concentration from the dissolved phase to the particulate phase. To highlight this role, the following experiment was conducted. Three different copper solutions of 1, 2, and 5 mg/l of pH 7.0 were allowed to react for 15 min with the surface area of three equal cement tanks (1 l volume and 53.1  103 m2 contact surface area). After a contact time of 15 min, samples were collected and analyzed for pH, copper, and calcium concentration. The results obtained are presented in Table 1. It is evident that even a short retention time of 15 min of the roof runoff in the hydrodynamic sediment separator is enough to increase the pH of the runoff by up to three units. This increase of pH results from the dissolution of calcium hydrates, such as Ca(OH)2 and calcium silicate hydrate at the concrete surface, as well as from the carbonation process (Okochi et al., 2000; Yang et al., 2003; Chang and Chen, 2006). At a pH level of 10.0, almost all the copper concentration of the runoff is dominated by the particulate phase. This phenomenon can influence the performance of copper retention capability of the clinoptilolite barrier significantly.

ARTICLE IN PRESS 3256

WAT E R R E S E A R C H

500

Table 2 – Installation and maintenance costs of the technical infiltration system

8

300

6

200

4

100

2

0

Rain intensity (l/s ha)

10 Cu2+ Rain intensity

400 Cu2+ (µg/l)

41 (2007) 3251– 3258

Installation costs Filtration pit (h) 4839

Infiltration ditch (h)

Barrier material (h)

Sedimental removal (h)

1000

1000

100.0

0 0

4

8

12 16 20 Rain event

24

28

32

Fig. 7 – Influence of rain intensity on the copper concentration of runoff after the filtration unit.

Clinoptilolite, as a natural ion exchanger, is able to remove copper from water solutions only if the copper concentration of the solution is dominated by the dissolved phase. Particulate copper can be transported through the clinoptilolite barrier. Fig. 7 demonstrates the influence of rain intensity on the copper concentration of runoff after the clinoptilolite filtration unit for sampled rain events. This irregular behavior between rain intensity and copper concentration indicates that copper particles can be transported through the clinoptilolite filter body during strong rain events and can be eventually washed out by strong as well as by weak rain events. Additionally, the dissolved calcium ions generated by the dissolution of calcium hydrates and by the carbonation process at the concrete surface are able to compete successfully with the copper ions of roof runoff for ion exchange sites on the clinoptilolite barrier surface. According to Okochi et al. (2000), the intensity of dissolution of calcium hydrates and of the carbonation process decreases gradually during exposure time and after 3 yr of exposure the pH of the runoff rainwater reduced from 10.0 to 7.5. This indicates that copper precipitates in the hydrodynamic sediment separator could re-dissolve in the runoff water, resulting in unpredictable situations. To avoid such undesirable situations, it is proposed to encase the surface of the hydrodynamic sediment separator with a synthetic plastic material of good dimensional stability.

the risk of a biofilm development on the surface of the clinoptilolite barrier material. A biofilm could reduce the hydraulic conductivity of the barrier material by clogging its pores, resulting probably in the colmation of the filtration unit. As a consequence, the maintenance time of the technical infiltration system becomes unpredictable, resulting in higher maintenance costs as well. The clinoptilolite barrier material applied has an effective ion exchange capacity of 7.0 g/kg regarding copper (Athanasiadis, 2005). Considering a mean copper runoff rate of 1.3 g/m2 yr and that the technical infiltration system is designed to drain the runoff of a roof surface of 500 m2, it could be estimated that the clinoptilolite barrier material has a minimum maintenance life of 8 years. This refers only to the case of the copper concentration entering the clinoptilolite barrier, which is dominated by the dissolved phase.

4.

 The cover material of the roof and the drainage system





Installation and maintenance costs

Table 2 demonstrates the costs of installation and maintenance of the applied technical infiltration system. These prices refer to a system designed to dewater the runoff of a 500 m2 roof surface. The phenomenon described above, regarding the transformation of copper phase distribution in the hydrodynamic centrifugal sediment separator, degrades the performance capability of the clinoptilolite barrier material. Therefore, the application of the special infiltration trench becomes necessary in order to achieve copper concentrations in percolation water lower than the value of 50 mg/l set by the German Federal Soil Protection Act and Ordinance, leading of course to higher investment costs. The same phenomenon increases

Conclusions

The results obtained can be summarized as follows:

 3.3.

Maintenance costs



were responsible for the high copper concentrations in the roof runoff. The amount and the runoff rate of precipitation were the most important parameters for the magnitude of the copper runoff rate. The phase distribution of the copper concentration in the roof runoff was dominated by the dissolved phase. The technical infiltration system equipped with clinoptilolite as a barrier material was able to reduce the copper from the roof runoff by a factor up to 96%. The application of this dewatering system contains no risk of soil and groundwater contamination, regarding copper.

According to these results, the application of clinoptilolite as an artificial barrier material for on-site infiltration of runoff from copper roofs appears to be an advanced way of managing roof runoff situations in urban areas.

Acknowledgment This work was financed by the Bavarian State Office of Environment (33-4402.3).

ARTICLE IN PRESS WAT E R R E S E A R C H

Appendix A.

4 1 (200 7) 325 1 – 325 8

Supplementary materials

Supplementary data associated with this article can be found in the online version at doi:10.1016/j.watres.2007.05.019.

R E F E R E N C E S

Ammann, A., Hoehn, E., Koch, S., 2003. Ground water pollution by roof runoff infiltration evidenced with multi-tracer experiments. Water Res. 37 (5), 1143–1153. Athanasiadis, K., 2005. On site infiltration of roof runoff by using clinoptilolite as an artificial barrier material. Doctor’s Thesis, Institute of Water Quality Control, Technical University of Munich, Garching. Athanasiadis, K., Helmreich, B., 2005. Influence of chemical conditioning on the ion exchange capacity and on kinetic of zinc uptake by clinoptilolite. Water Res. 39 (8), 1527–1532. Athanasiadis, K., Helmreich, B., Christ, O., Winfried, R., 2004. Decentralised treatment of roof and street runoff prior infiltration (Vorrichtung und Schacht zur dezentralen Behandlung vom Wasser sowie Verfahren hierzu). European Patent Office, Pattentblatt 2005/15. Bavarian State Office for Environment, 2005. Air quality report 2004 (Lufthygienischer Jahresbericht 2004). Augsburg 2005. Berber-Mendoza, M.S., Leyva-Ramos, R., Alonso-Davila, P., Mendoza-Barron, J., Diaz-Flores, P.E., 2006. Effect of pH and temperature on the ion exchange isotherm of Cd2+ and Pb2+ on clinoptilolite. J. Chem. Technol. Biotechnol. 81 (6), 966–973. Bucheli, T.D., Muller, S.R., Heberle, S., Schwarzenbach, R.P., 1998. Occurence and behavior of pesticides in rainwater, roof runoff, and artificial stormwater infiltration. Environ. Sci. Technol. 32 (22), 3457–3464. Chang, C.-F., Chen, J.-W., 2006. The experimental investigation of concrete carbonation depth. Cement Concr. Res. 36 (1), 1760–1767. Davies, H., Hollis, T., 1981. Measurements of rainfall runoff volume relationships and water balance from roof and roads. In: Yen, BC. (Ed.), Second International Conference on Urban Storm Drainage. Department of Civil Engineering, University of Illinois, Urbana, IL, pp. 434–443. Davis, A.P., Shokouhian, M., Ni, S.B., 2001. Loading estimates of lead copper, cadmium, and zinc in urban runoff from specific sources. Chemosphere 44 (5), 997–1009. Doula, M., Ioannou, A., Dimirkou, A., 2002. Copper adsorption and Si, Al, Ca, Mg and Na release from clinoptilolite. J. Colloid Interf. Sci. 245 (2), 237–250. ¨ stlund, P., Hansen, C., 2001. Digital air photo Ekstrand, S., O processing for mapping of copper roof distribution and estimation of related copper pollution. Water Air Soil Pollut. 1 (1), 267–278. Faller, M., 2003. Runoff behaviour of different materials used for roofing and facades—5 years exposure study in Switzerland. In: Fitz, S., Kucera, V. (Eds.), Release of Metals Due to Corrosion of Materials—Proceedings of a UN ECE Workshop, Swedish Corrosion Institute, Munich, Germany. German Federal Soil Protection Act and Ordinance (BBodSchV), 1999. Bundesbodenschutz- und Altalastenverordnung. BGB1 I 1999, 1554. Hollis, G.E., Ovenden, J.C., 1988. The quantity of storm water runoff from 10 stretches of road, a car park and 8 roofs in Hertfordshire, England during 1983. Hydrol. Process. 2 (3), 227–243. Hullmann, H., 2003. Runoff rates, quantities and concentrations of copper and zinc. In: Fitz, S., Kucera, V. (Eds.), Release of

3257

Metals Due to Corrosion of Materials—Proceedings of a UN ECE Workshop, Swedish Corrosion Institute, Munich, Germany. Inglezakis, V.J., Loizidou, M.D., Grigoropoulou, H.P., 2003. Ion exchange of Pb2+, Cu2+, Fe3+, and Cr3+ on natural clinoptilolite: selectivity determination and influence of acidity on metal uptake. J. Colloid Interf. Sci. 261 (1), 49–54. Karlen, C., Wallinder, I.O., Hejerlick, D., Leygraf, C., 2002. Runoff rates, chemical speciation and bioavailability of copper released from naturally patinated copper. Environ. Pollut. 120 (3), 691–700. Leuenberger-Minger, A.U., Faller, M., Richner, P., 2002. Runoff of copper and zinc caused by atmospheric corrosion. Mater. Corros. (Werkstoffe und Korrosion) 53 (3), 157–164. Maksimovic, C., 1996. Rain and floods in our cities, gauging the problem. Technical Report in Hydrology and Water Resources, No. 53, Geneva, p. 168. Mason, Y., Ammann, A.A., Ulrich, A., Sigg, L., 1999. Behaviour of heavy metals, nutrients and major components during roof runoff infiltration. Environ. Sci. Technol. 33 (10), 1588–1597. Mouli, P.C., Mohan, S.V., Reddy, S.J., 2005. Rainwater chemistry at a regional representative urban site: influence of terrestrial sources on ionic composition. Atmos. Environ. 39 (6), 999–1008. Negrel, P., Roy, S., 1998. Chemistry of rainwater in the Massif Central (France): a strontium isotope and major element study. Appl. Geochem. 13 (8), 941–952. Odnevall Wallinder, I., Leygraf, C., 2001. Seasonal variations in corrosion rate and runoff rate of copper roofs in an urban and rural atmospheric environment. Corros. Sci. 43 (12), 2379–2396. Odnevall Wallinder, I., Verbiest, P., He, W., Leygraf, C., 2001. A comparison between corrosion rates and runoff rates from new and aged copper and zinc as roofing material. Water Air Soil Pollut. 1 (1), 67–82. Okochi, H., Kameda, H., Hasegawa, S., Saito, N., Kubota, K., Igawa, M., 2000. Deterioration of concrete structures by acid deposition— an assessment of the role of rainwater on deterioration by laboratory and field exposure experiments using mortar specimens. Atmos. Environ. 34 (1), 2937–2945. Persson, D., Kucera, V., 2001. Release of metal from buildings, constructions and products during atmospheric exposure in Stockholm. Water Air Soil Pollut. 1 (1), 133–150. Priggemeyer, S., Priggemeyer, S., 2003. Investigation of the metal runoff from copper surfaces exposed to atmosphere. In: Fitz, S., Kucera, V. (Eds.), Release of Metals Due to Corrosion of Materials—Proceedings of a UN ECE Workshop, Swedish Corrosion Institute, Munich, Germany. Ragab, R., Bromley, J., Rosier, P., Cooper, J.D., Gash, J.H., 2003. Experimental study of water fluxes in a residential area: 1. Rainfall, roof runoff and evaporation: the effect of slope and aspect. Hydrol. Process. 17 (12), 2409–2422. Reimann, C., DeCaritat, P., Halleraker, J.H., Volden, T., Ayras, M., Niskavaara, H., Chekushin, V.A., Pavlov, V.A., 1997. Rainwater composition in eight arctic catchments in northern Europe (Finland, Norway, and Russia). Atmos. Environ. 31 (2), 159–170. Rodriguez-Iznaga, I., Gomez, A., Rodriguez-Fuentes, G., BenitezAguilar, A., Serano-Ballan, J., 2002. Natural clinoptilolite as an exchanger of Ni2+ and NH4+ ions under hydrothermal conditions and high ammonia concentration. Micropor. Mesopor. Mater. 53 (1–3), 71–80. Scholz, M., 2004. Case study: design, operation, maintenance and water quality management of sustainable storm water ponds for roof runoff. Bioresourc. Technol. 95 (3), 269–279. Steiner, M., Pronk, W., Boller, M.A., 2006. Modelling of copper sorption onto GFH and design of full scale GFH adsorbers. Environ. Sci. Technol. 40 (5), 1629–1635. UBA, 2005. Distribution of copper, zinc and lead in receiving waters and in soil. Research Report: 202 242 20/02.

ARTICLE IN PRESS 3258

WAT E R R E S E A R C H

41 (2007) 3251– 3258

Yang, T., Keller, B., Magyari, E., Hametner, K., Gu¨nther, D., 2003. Direct observation of the carbonation process on the surface of calcium hydroxide crystals in hardened cement paste using an atomic force microscope. J. Mater. Sci. 38 (1), 1909–1916. Zimmermann, J., Dierkes, C., Gobel, P., Klinger, C., Stubbe, H., Coldewey, W.G., 2005. Metal concentrations in soil and seepage

water due to infiltration of roof runoff by long term numerical modelling. Water Sci. Technol. 51 (2), 11–19. Zobrist, J., Muller, S.R., Ammann, A., Bucheli, T.D., Mottier, V., Ochs, M., Scho¨nenberger, R., Eugster, J., Boller, M., 2000. Quality of roof runoff for groundwater infiltration. Water Res. 34 (5), 1455–1462.