Bioresource Technology 95 (2004) 269–279
Case study: design, operation, maintenance and water quality management of sustainable storm water ponds for roof runoff Miklas Scholz
*
Institute for Infrastructure and Environment, School of Engineering and Electronics, College of Science and Engineering, The University of Edinburgh, Faraday Building, The King’s Buildings, Edinburgh EH9 3JL, Scotland, UK Received in revised form 24 May 2003; accepted 15 July 2003 Available online 2 April 2004
Abstract The purpose of this case study was to optimise design, operation and maintenance guidelines, and to assess the water treatment potential of a storm water pond system after 15 months of operation. The system was based on a combined silt trap, attenuation pond and vegetated infiltration basin. This combination was used as the basis for construction of a roof water runoff system from a single domestic property. United Kingdom Building Research Establishment and Construction Industry Research and Information Association, and German Association for Water, Wastewater and Waste design guidelines were tested. These design guidelines failed because they did not consider local conditions. The infiltration function for the infiltration basin was logarithmic. Algal control techniques were successfully applied, and treatment of rainwater runoff from roofs was found to be largely unnecessary for recycling (e.g., watering plants). However, seasonal and diurnal variations of biochemical oxygen demand, dissolved oxygen and pH were recorded. 2004 Elsevier Ltd. All rights reserved. Keywords: Roof water runoff; Attenuation pond; Vegetated infiltration basin; Silt trap; British Building Research Establishment; Construction Industry Research and Information Association; German Association for Water, Wastewater and Waste; Water quality management; Sampling scheme; Algae
1. Introduction 1.1. Sustainable roof runoff water drainage Conventional storm water and urban drainage systems are designed to dispose of rainfall runoff water as quickly as possible. This results in end of pipe solutions that often involve the provision of large interceptor and relief sewers, huge storage tanks in downstream locations and centralised wastewater treatment facilities (Butler and Davies, 2000; Abbott and Comino-Mateos, 2001; Ellis et al., 2002). In contrast, combined attenuation pond and infiltration basin systems can be applied as cost-effective end of pipe drainage solutions for local source control; e.g., collection of roof drainage. It is often possible to divert all roof drainage for infiltration or storage and subsequent recycling. As runoff from roofs is a major con*
Tel.: +44-131-6-506780; fax: +44-131-6-506554. E-mail address:
[email protected] (M. Scholz).
0960-8524/$ - see front matter 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.biortech.2003.07.015
tributor to the quantity of surface water requiring disposal, this is a particularly beneficial approach where suitable ground conditions prevail (Butler and Davies, 2000). Furthermore, roof runoff water is usually considered to be cleaner than road runoff water. However, diffuse pollution can have a significant impact on the water quality of any surface water runoff (Ellis et al., 2002). 1.2. Case study: site description A domestic property in Sandy Lane (Bradford, West Yorkshire, England) was selected for this pilot plant case study. The study area is located approximately 1.8 West of Greenwich and 53.8 North of the Equator. The surface water (subject to disposal) came from the roofs of the house and a tandem (double) garage. Concrete roof tiles were used for the house, and the garage roof was covered with a layer of gravel. The roof materials had not been cleaned for at least five years prior to the study (current and previous house owners,
270
M. Scholz / Bioresource Technology 95 (2004) 269–279
Nomenclature coefficient (unknown function of various variables including rainfall intensity and infiltration rate) ATV-DVWK German Association for Water, Wastewater and Waste b maximum experimental depth (mm) within the infiltration basin during an individual storm BOD5 five-day @ 20 C biochemical oxygen demand (mg/l) BRE Building Research Establishment CIRIA Construction Industry Research and Information Association D infiltration basin design depth (mm) DO dissolved oxygen (mg/l) ETAAS Electrothermal Atomic Absorption Spectrometer GAC granular activated carbon ICP-OES Inductively Coupled Plasma Optical Emission Spectrometer IR empirical infiltration rate (m/s) NTU nephelometric turbidity unit a
R SS T TS Z1 Z2
mean product moment correlation coefficient suspended solids (mg/l) infiltration time (s) total solids (mg/l) factor (defined by the BRE method) growth factor (defined by the BRE method)
Considering that sustainable urban drainage technology is a novel research area, definitions of the terminology used in this paper are required. Reduction of the rate of flow through a system, which has the effect of reducing the peak flow and increasing the duration of a flow event is defined as attenuation. A pond is called more specific a wet pond if it is a permanently wet depression designed to retain storm water for several days, and to permit settling of SS. It follows that an attenuation pond combines both the meaning of attenuation and pond. In comparison, an infiltration basin is a dry basin designed to promote infiltration of surface water to the ground. If vegetated, a small infiltration basin is often referred to as a dry pond.
personal communication). It follows that dirty roofs could be a source of inorganic and organic pollution. It is likely that the roof material supports micro-habitats that include algae. However, a detailed discussion of these problems is beyond the scope of this paper. In the original pipeline layout dated 1972, rainwater drained into the public sewer. However, in April 2001 this layout was changed in order to feed a semi-natural attenuation pond structure (Fig. 1) with rainwater. The storage water was predominantly used for watering garden plants in summer, but there is a much greater potential for other usage; e.g., recycling of storm water to flush toilets (Butler and Davies, 2000). If the attenuation pond structure overflows, the water is transferred to a vegetated infiltration basin structure (Fig. 1). 1.3. Purpose
Front Garden Property Border
House Roof
House Garage
Roof
Roof Pipework
Roof Drain
Back Garden
Pond System
Wet Pond
The purpose of this paper is not only to investigate a case study of a sustainable urban drainage system designed according to best management practice (Martin et al., 2000) but also to address the following key objectives in order to assess the potential for scaled-up systems: • to identify technical constraints associated with the design, operation and maintenance, • to identify the infiltration characteristics of storm water into the vegetated infiltration basin,
Slope North
Infiltration Point Dry Pond
= 1 meter
Fig. 1. Drawing of the case study site showing roof areas, pipework and the combined attenuation wet and dry pond system. The rainwater flows (indicated by arrows) from the roof areas into the drainage pipe network which conveys the water first into the silt trap, than into the wet pond and finally into the dry pond. Detailed dimensions are given in Section 2.1.
M. Scholz / Bioresource Technology 95 (2004) 269–279
• to suggest water quality monitoring and management strategies including algal control techniques and • to assess the water treatment potential of storm water ponds for roof runoff water.
2. Methods 2.1. Design of the study site The pilot plant was designed considering sustainable urban drainage system guidelines of BRE (1991), CIRIA (Bettes, 1996; Martin et al., 2000) and ATV-DVWK (2002). The system design allows flooding to occur only once within 10 years, and this is defined as the return period. However, the German guideline recommends construction for a five-year design storm only (ATVDVWK, 2002). The system was based on a combined attenuation pond and infiltration basin design (Fig. 1). Rainwater runoff from the roofs of one house (plus a tandem garage) was drained directly into a silt trap, which fulfills the purpose of a small sedimentation tank. The surface areas of both house roofs were 29 m2 each. The angle between each roof and the ceiling of the house was 23. Therefore, the total theoretical horizontal area of the house roofs was 53 m2 . The roof area of the tandem garage was 33 m2 . It follows that the total theoretical horizontal area to be drained was 86 m2 . The distances between the nearest building (garage) and the attenuation pond and infiltration basin were 1.5 and 5.0 m, respectively (Fig. 1). The total length of the horizontal plastic pipework close to the ground (mean angle of 2) was 19.6 m. Guttering and downpipes were not included in this sum. The inside diameter of all pipes was 6.5 cm. The pipeline layout could have been optimised but the sustainable urban drainage system was retrofitted in this case study, and it was therefore the aim to recycle as much of the old pipework as possible. The traditional drains were sealed off temporarily although the house owners preferred a system that would allow rainwater diversion into the main sewer in cases of exceptional rainfall or failure of the infiltration basin. However, neither of these events occurred during the duration of the experiment. The maximum horizontal dimensions (length · width) of the silt trap, attenuation pond and infiltration basin were 0.7 m · 0.4 m, 3.2 m · 1.7 m and 3.7 m · 2.5 m, respectively (Fig. 1). The maximum depths of the lined attenuation pond and the vegetated infiltration basin were approximately 40 cm each. All water level measurements were taken daily at a reference level structure that was part of the attenuation pond outflow structure (Fig. 1). The total area of the experimental attenuation pond was 5.5 m2 when completely filled with water (Fig. 1).
271
The slope ratios of the infiltration basin were 1:1.6 (towards the attenuation pond in the West), 1:1.9 (towards North and South) and 1:2.4 (towards the lower garden in the East). The attenuation pond and infiltration basin were planted with common aquatic plants, which were collected from local natural habitats. The dominant macrophytes of the attenuation pond were Common Reed (Phragmites australis), Reedmace (Typha latifolia) and Yellow Iris (Iris pseudacorus). Common Waterstarwort (Callitriche stagnalis) and Frogbit (Hydrocharis morsus-ranae) were common floating aquatic plants. Canadian Waterweed (Elodea canadensis) was the dominant submerged plant. Common Reed and Reedmace (both deep-rooting) were also planted in the infiltration basin in order to enhance infiltration properties although this has been disputed by some findings reviewed by Brix (1997). This combination of plants was chosen because it is typical for similar watercourses around Bradford. At the top of the food chain (attenuation pond) were two Koi fish (introduced during pond construction in April 2001), at least 25 European Three-spined Sticklebacks (Gasterosteus aculeatus; introduced in September 2001) and the Common Grass Frog (Rana temporaria; approximately 20 adults and at least 2500 tadpoles during spring and summer).
2.2. Engineering methods A simple rain gauge was used, comprised of a measurement cylinder fed through a funnel (diameter of 9 cm). Two gauges were used to give a more representative estimate of rainfall. Experimental rainfall data were compared with official data (Bradford measurement station) supplied by the Meteorological Office (2002). The rain gauges were located between the attenuation pond and infiltration basin (Fig. 1). A cost-effective, non-destructive test on a small-scale was required in order to assess the storm water (roof runoff) infiltration properties into the infiltration basin. This information was important in order to estimate the infiltration rates required for the design calculations (Section 2.1). Therefore, a simple infiltration device specifically designed to promote the passage of gravityfed tap water into the natural ground was used. Infiltration rates were determined fortnightly or monthly by measuring the actual infiltration time of 200 ml tap water through the first 4 cm of a 24 cm long drainage pipe (diameter of 6.5 cm). The pipe was buried vertically in the ground at a depth of approximately 20 cm. The infiltration test was carried out at three infiltration point stations in parallel under varying natural and particularly meteorological conditions. The infiltration stations are indicated by small circles in Fig. 1.
272
M. Scholz / Bioresource Technology 95 (2004) 269–279
2.3. Water quality analysis Daily or weekly sampling schemes were applied. Daily sampling and subsequent analysis took place at approximately 06:00 and/or 18:00. All analytical procedures to determine water quality were performed according to the American standard methods (Clesceri et al., 1995) that also outline the corresponding water quality criteria. Water samples were tested for temperature (air and water), BOD5 , SS, TS, conductivity, turbidity, DO and pH. Hanna instruments HI 9033 conductivity, C 102 turbidity, HI 9142 DO and HI 8519N pH meters were used throughout the study. Oxidised aqueous nitrogen was determined as the sum of nitrate and nitrite. Nitrate was reduced to nitrite by cadmium and determined as an azo dye at 540 nm (using a Peristorp Analytical EnviroFlow 3000 flow injection analyser) following diazotisation with sulfanilamide and subsequent coupling with N-1-naphthylethylenediamine dihydrochloride. This technique measured also nitrite at the same time. Furthermore, aqueous ammonia reacted with hypochlorite and salicylate ions in solution in the presence of sodium nitrosopentacyanoferrate (nitroprusside), and aqueous phosphate reacted with acidic molybdate to form a phosphomolybdenum blue complex. The associated coloured complexes were measured spectrophotometrically at 655 and 882 nm, respectively, using a Bran and Luebbe Autoanalyser (Model AAIII). All analyses for nutrients were carried out in triplicate. Vegetative, sediment and soil samples were dried at 105 C overnight in a drying oven (UM500, Memmert) prior to being ashed at 400 C for 12 h in a muffle furnace (ELF 11/14, Carbolite). Ashed samples (0.2–0.6 g) were then digested under reflux in aqua regia for 2 h, cooled, filtered through Whatman no. 5412 filter papers and made up to 100 ml with de-ionised water ready for analysis. An ICP-OES (maker of the TJA IRIS instrument: ThermoElemental, USA), at 1350 W with coolant, auxiliary and nebuliser argon gas flows of 15, 0.5 and 0.7 ml/min, respectively, and a pump flow rate of 1 ml/min was used to screen for total concentrations of analyte elements in filtered (0.45 lm) water samples and digests. Multi-element calibration standards in the concentration range 0.1–10 mg/l were used and the emission intensity measured at appropriate wavelengths in nm. Further information on elements and corresponding wavelengths used is available on request. For all elements, analytical precision (relative standard deviation) was typically 5–10% for individual aliquots. Three replicates for each sample were analysed. An ETAAS (maker of Varian SpectrAA 400: Varian Inc., Australia) with auto-sampler and powered by a GTA-96 graphite tube atomiser was used to analyse
some of the water samples for their zinc content. A 20-ll injection volume was used for samples and standards in notched GTA partition tubes (coated). Nitrogen was applied as the carrier gas. The char temperature was 300 C with a ramp rate of 10 C/s and a hold time of 3 s. For atomisation, temperature was 1900 C with ramp and hold times of 1 and 2 s, respectively. Precision in terms of the relative standard deviation was typically less than 5% for triplicate injections. Three sample replicates were taken from pre-filtered (Whatman 0.45 lm cellulose nitrate membrane filters) subsamples. 2.4. Control of algal growth Visual algal cover estimations were undertaken by using a 0.2 · 0.2 m reference metal grid located on top of the attenuation pond during the experiment that could also be used as a permanent safety barrier. Microorganisms, including algae like Microspora spp. (dominates blanket weed), were counted using a Sedgewick-Rafter Cell S50 (counting chamber: 1 · 50 · 20 mm) and a Wang Biomedical Research Microscope 6000 (bright field and phase-contrast). Excess algae and decomposing litter were occasionally removed from the attenuation pond and infiltration basin. Wet algae and litter were weighed after the biomass had drained for two minutes by applying maximum pressure with both hands to small portions of the organic waste. Blanket weed control with barley straw bales (commercial product called Frogmat) was also practised. The straw bales were sub-divided into smaller bales before application. The new bales (approximately 0.2 kg straw each) were located at eight sites (approximately 1.6 kg straw in total) near the attenuation pond margins, and stayed there almost fully submerged between 31 August and 31 October 2001. This procedure was repeated between 22 February and 18 May 2002. Between 15 March and 12 August 2002, an experimental GAC filter cleaned the storm water before it reached the silt trap. Approximately 1.4 kg GAC (Grade 207 EA, US mesh: 12 · 40) was used to test the purification potential with respect to trace elements which might contribute to algal growth. 2.5. System capacity The silt trap and attenuation pond had a combined total effective volume of approximately 1.9 m3 during storm events. The rainwater drained into the silt trap with a maximum capacity of 0.1 m3 . Suspended solids (e.g., weathered building materials, decayed leaves, bird droppings and particles from atmospheric pollution) from the roofs settle down predominantly in the silt trap. Water from the silt trap overflowed into the
M. Scholz / Bioresource Technology 95 (2004) 269–279
attenuation pond (volume of approximately 1.7 m3 ), which also served the purpose of a storage pond. If the attenuation pond was full, water flowed into the infiltration basin, which accommodated a maximum volume of approximately 1.8 m3 during heavy storm events before it overflowed (Fig. 1).
3. Results and discussion 3.1. Standard design considerations Tables 1–3 summarise the calculation procedures for the BRE, CIRIA and ATV-DVWK design guidelines (BRE, 1991; Bettes, 1996; Martin et al., 2000; ATVDVWK, 2002). Findings are based on the design assumption that flooding for the case study site might
273
statistically occur only once within 10 years (defined as return period). The estimated mean IR applied for all methods was 104 m/s (Tables 1–4). The critical storm durations for the BRE, CIRIA and ATV-DVWK design calculations were 1.0, 0.5 and 1.0 h, respectively. The associated maximum infiltration basin height requirements were 28, 21 and 26 cm, respectively. Furthermore, the maximum infiltration storage volume requirements were 1.41, 1.04 and 1.27 m3 , respectively. Water level fluctuations within the attenuation pond and infiltration basin are indicated in Fig. 2. Maximum infiltration basin height requirements calculated according to BRE (1991), CIRIA (Bettes, 1996) and ATV-DVWK (2002) guidelines were not sufficient for the period of the experiment. Fig. 2 indicates when the system would have failed if the
Table 1 Calculation of the required infiltration basin storage volume (italic number) using the Building Research Establishment guidelines (BRE, 1991) Duration (min) Factor Z1 Design rainfall M5-D (five-year return period;mm) Growth factor Z2 Design rainfall M10-D (10-year return period; calculated from M5 rainfall;mm) Inflow (10-year return period;m3 ) Outflow drainage (m3 ) Required storage (10-year return); M10-D (m3 )
10
15
30
60
120
0.46 9.2 1.22 11.2
0.56 11.2 1.22 13.7
0.75 15.0 1.24 18.6
1.00 20.0 1.24 24.8
1.30 26.0 1.24 32.2
0.97 0.12 0.85
1.18 0.18 1.00
1.60 0.36 1.24
2.13 0.72 1.41
2.77 1.44 1.33
Assumptions: ratio ¼ 0.24 (ratio between the M5-60 storm and the M5-2D rainfall); impermeable area ¼ 86 m2 ; M5-60 ¼ 20 mm; infiltration rate ¼ 0.0001 m/s (Table 4); perimeter ¼ 10 m; depth ¼ 0.4 m; surface area to 50% effective depth (excluding base) ¼ 2 m2 .
Table 2 Calculation of the required infiltration basin storage height (italic number) using the Construction Industry Research and Information guidelines (Bettes, 1996) Duration (min) Intensity of storm (mm/h) Maximum height of water to be stored (m)
10
15
30
60
120
67.3 0.154
54.6 0.184
37.2 0.213
24.8 0.210
16.1 0.142
Assumptions: infiltration rate ¼ 0.0001 m/s (Table 4); safety factor ¼ 2 (area to be drained < 100 m2 ; minor inconvenience if system fails); area of base ¼ 4.95 m2 ; perimeter ¼ 10 m; impermeable area ¼ 86 m2 ; void ratio of the soakaway fill material ¼ 1 (represents air).
Table 3 Calculation of the required infiltration basin storage volume (italic number) using the German Association for Water, Wastewater and Waste guidelines (ATV-DVWK, 2002) Duration (min) Runoff for each design storm (l/(s · h)) Volume to be stored (m3 )
10
15
30
60
120
186.9 0.85
151.7 0.99
103.3 1.19
68.9 1.27
44.7 0.96
Assumptions: impermeable area ¼ 86 m2 ; surface factor ¼ 0.9 (considering roof angle and building material); infiltration area ¼ 4.95 m2 ; infiltration rate ¼ 0.0001 m/s (Table 4); safety factor ¼ 1.1 (minor inconvenience if system fails).
274
M. Scholz / Bioresource Technology 95 (2004) 269–279
Table 4 Summary statistics: water quality of the attenuation pond and infiltration rates (IR) for the infiltration basin (26 April 2001 to 12 August 2002) Variable
Unit
Time
Sample number
Mean
First summera mean
Autumnb mean
Winterc mean
Spring meand
Air temperature Water temperature BOD5 g Suspended solids Total solids Conductivity Turbidity Dissolved oxygen pH Algal cover IR (Station 1)j IR (Station 2)j IR (Station 3)j
C C mg/l mg/l mg/l lS NTUi mg/l – % 107 m/s 107 m/s 107 m/s
PMf PMf AMh AMh AMh PMf AMh PMf PMf PMf PMf PMf PMf
345 325 57 40 40 281 39 319 263 306 23 23 24
12.2 11.1 4.3 46.8 193.3 39.8 2.8 9.3 7.77 44 8 158 9946
16.9 15.4 4.2 132.5 238.7 75.1 4.3 11.5 8.36 61 3 4270 8152
8.9 8.6 3.1 51.3 293.8 37.5 2.7 4.8 7.27 50 23 95 4778
6.5 9.1 5.6 3.8 104.7 33.0 2.2 10.1 7.86 36 1 42 1286
13.2 11.9 6.7 54.8 151.5 37.7 2.0 12.4 8.69 39 2 36 6342
Second summer meane 17.1 15.3 2.5 3.4 91.7 34.2 1.5 7.3 6.84 38 5 580 31,841
a
Summer: 21/06–21/09/01. Autumn: 22/09–20/12/01. c Winter: 21/12/01–19/03/02. d Spring: 20/03–20/06/02. e Summer: 21/06–12/08/02. f PM ¼ afternoon. g BOD5 ¼ five-day @ 20 C biochemical oxygen demand. h AM ¼ morning. i NTU ¼ nephelometric turbidity unit. j IR ¼ infiltration rates were calculated for the soil layers located 20 cm below the bottom of the infiltration basin centre (Station 1), the middle of the slope of the infiltration basin (Station 2) and the lawn (Station 3), see also Fig. 1. b
350
300
BRE ATV-DVWK
400
250 CIRIA
Jul-02
Aug-02
Jul-02
Jun-02
May-02
Apr-02
May-02
Apr-02
Mar-02
Feb-02
Jan-02
Feb-02
Jan-02
Nov-01
Dec-01
Oct-01
0
Nov-01
300
Oct-01
50
Aug-01
320
Sep-01
100
Jul-01
340
Aug-01
150
Jul-01
360
Jun-01
200
May-01
380
Water level (mm) of the dry pond
420
May-01
Water level (mm) of the wet pond
440
Time (d) Wet pond (mm)
Dry pond (mm)
Fig. 2. Maximum daily water level fluctuations within the wet pond and dry pond between 13 May 2001 and 12 August 2002. Maximum infiltration basin height requirements calculated according to Building Research Establishment, BRE (1991), Construction Industry Research and Information Association, CIRA (Bettes, 1996; Martin et al., 2000), and German Association for Water, Wastewater and Waste, ATV-DVWK (2002), guidelines are indicated by horizontal lines.
recommended design depths for the infiltration basin had been applied. Eq. (1) indicates the mathematical relationship between the infiltration basin design depth D (mm) and the
infiltration time T (h). The corresponding R for the function is 0.92. D ¼ a þ lnðT Þ þ b
ð1Þ
M. Scholz / Bioresource Technology 95 (2004) 269–279
a ¼ 2:54 b0:51 ;
R ¼ 0:58 ðin generalÞ
a ¼ 0:59 b0:83 ;
R ¼ 0:89 ðfor b < 190 mmÞ;
a ¼ 0:015 b1:51 ;
depths related to different infiltration times and vice versa. 3.2. System design comparisons
R ¼ 0:77
ðfor b P 190 mm during spring; summer and autumnÞ a ¼ 0:001 b
1:94
;
R ¼ 0:75 ðfor b P 190 mm during winterÞ: Fig. 3 indicates the mathematical relationship between the terms a and b. Both a and b are based on logarithmic trendline functions for 36 recorded storms (between 3 and 8 depth measurements each) between May 2001 and August 2002. Term a is a coefficient whereas b represents the maximum depth (mm) within the infiltration basin during an individual storm. The corresponding R accounts for differences in infiltration behaviour for systems that are either shallower or deeper than approximately 185 mm. The mathematical relationships summarised in Eq. (1) and Fig. 3 give the design engineer the opportunity to estimate design
70 0.51
a = 2.54 x b R = 0.58
60 50
a
40 30 20 10 0 50
100
150
200
250
300
b 70
The actual design for the infiltration basin (40 cm depth and 1.7 m3 volume) was acceptable when compared to BRE (1991), CIRIA (Bettes, 1996) and ATVDVWK (2002) guidelines. Signs of system failure (e.g., flooding of the nearby lawn and structural damage) have not yet been observed. However, strict application of all test guidelines (without adding a higher safety factor than recommended) would have led to system failures during the first year of operation. This finding is in contrast to a previous investigation by Abbott and Comino-Mateos (2001), where British design guidelines did not fail. However, a perforated concrete ring soakaway to infiltrate stormwater runoff was used. It has been shown in Fig. 2 that all three official designs (Tables 1–3) have failed at least once during the first 15 months of the 10-year design period. However, the period of study can be described as a particularly wet time (2160 mm precipitation between summer 2001 and spring 2002). Moreover, the study site is situated on a semi-exposed location approximately 240 m above sea level. In comparison, the annual average rainfall for Bradford (located in a valley) is approximately 830 mm (Meteorological Office, 2002). Considering a maximum depth of 40 cm, the infiltration basin is associated with under-utilised storage capacity. However, observations have shown that a depth of 40 cm is justified considering that 20, 10 and 1 rainfall runoff events have led to the maximum design water depths of 21, 26 and 28 cm, respectively, being exceeded (Fig. 2). It follows that there would have been no spare capacity for strict applications of the standard design guidelines. 3.3. Water quality management
60
0.83
a = 0.59 x b R = 0.89
50
a
275
40 30 20 10 50
70
90
110
130
150
170
190
b
Fig. 3. Relationships between the coefficient a and the maximum depth b (mm) within the infiltration basin during an individual storm (top and bottom figures). The bottom figure indicates an optimised trendline fit for small values of b. R ¼ mean product moment correlation coefficient.
The water quality of the attenuation pond (Tables 4 and 5) was acceptable for disposal (e.g., sustainable drainage) and recycling (e.g., irrigation, toilet flushing and washing cars) according to findings presented by Butler and Davies (2000) and Ellis et al. (2002). Figs. 4 and 5 show results from the analysis of waters and digests by ICP-OES. Measured elemental concentrations were either low (boron, barium, calcium, magnesium, manganese and zinc), close to the detection limit (aluminium, copper and iron) or below the detection limit (most heavy elements were analysed but not listed; complete list available on request). Fig. 5 indicates that the liquid phase (rainwater) within the treatment chain can actually take up soluble contaminants from the sediment; the sediment acts as a source, thereby polluting the liquid phase.
276
Table 5 Summary statistics: 24 h sampling on 15 March 2002 Variable
a
C mg/l mg/l mg/l mg/l mg/l lS mg/l – mg/l mg/l mg/l mg/l mg/l mg/l
Total number 23 36 40 40 40 43 24 24 24 21 21 21 21 39 21
Mean
Standard deviation
00:00–05:30
06:00–11:30
12:00–17:30
18:00–23:30
00:00–05:30
06:00–11:30
12:00–17:30
18:00–23:30
3.4 1.3 0.54 0.06 0.07 99.8 36.0 10.0 7.54 10.37 1.13 0.006 0.007 0.009 1.57
4.3 1.1 0.38 0.03 0.04 97.8 42.0 9.8 7.57 10.89 1.14 0.006 0.010 0.006 1.57
7.4 0.7 0.39 0.04 0.06 180.7 43.1 11.3 7.89 11.08 1.14 0.005 0.012 0.006 1.66
4.8 2.0 0.47 0.07 0.13 139.2 41.3 11.6 8.01 10.96 1.14 0.010 0.010 0.006 1.72
0.6 0.8 0.37 0.06 0.06 18.8 1.2 0.7 0.39 1.19 0.08 0.002 0.003 0.004 0.24
2.1 0.6 0.18 0.04 0.04 14.6 1.5 1.2 0.51 0.96 0.07 0.002 0.005 0.004 0.09
0.7 0.5 0.17 0.05 0.06 59.1 1.7 0.8 0.24 1.23 0.04 0.002 0.009 0.003 0.28
1.4 1.1 0.19 0.07 0.09 31.3 2.0 1.3 0.58 0.07 0.06 0.009 0.001 0.002 0.31
BOD5 ¼ five-day @ 20 C biochemical oxygen demand. Concentrations include nitrite. c Filtered sample (0.45 lm pore size). d Analysis with the Inductively Coupled Plasma Optical Emission Spectrometer. e Analysis with the Electrothermal Atomic Absorption Spectrometer. b
M. Scholz / Bioresource Technology 95 (2004) 269–279
Water temperature BOD5 a Nitrate-Nb Ammonia-N Phosphate-P Total solids Conductivity Dissolved oxygen pH Calciumc;d Magnesiumc; d Manganesec; d Zincc;d Zincc;e Bariumc; d
Unit
M. Scholz / Bioresource Technology 95 (2004) 269–279
277
Fig. 4. Summary of the Inductively Coupled Plasma Optical Emission Spectrometer results for different aquatic plants (dry weights) located in the attenuation pond and infiltration basin. Concentration readings for boron, barium and zinc require division by a factor of 100, 10 and 10, respectively. WP ¼ wet pond, DP ¼ dry pond.
3.4. Twenty-four hour water quality monitoring Manual sampling over 24 h was conducted on 15 March 2002 (Table 5). Most samples were taken either in 30 min intervals or hourly. Despite the cold climate, diurnal water quality variations (Wu and Mitsch, 1998) were apparent for water temperature, BOD5 , nitrate-N (including nitrite-N), ammonia-N, conductivity, pH, manganese, zinc and barium. Of the elements detected by ICP-OES, zinc was the only one that had a high overall standard deviation (approximately 0.0060 mg/l, for triplicate analyses of each sample) as measurements were made close to the detection limit. However, analysing the same samples using ETAAS reduced the overall standard deviation to 0.0036 mg/l (three replicates for each sample) because this technique is more sensitive for the determination of zinc. There were no significant fluctuations of zinc during the day.
3.5. Aquatic plant management In order to prevent algal washout into the infiltration basin during storm events, it is necessary to control the algae with barley straw and mechanical removal (Taylor P., Director, Frogmat International Holdings, personal communication). Algal washout allows decaying algae to block fine pores, thereby decreasing the infiltration capacity of the basin (see also Table 4).
In addition to biological (grazing zooplankton and tadpoles) and physical (removal by hand) algal control, biochemical control was also implemented (with barley straw extracts). The presence of small barley straw bales is likely to reduce algal growth by releasing a cocktail of phytotoxic chemicals (Everall and Lees, 1997). The wet weights of litter removed during summer 2001, autumn 2001, winter 2001/02 and spring 2002 were 1.1, 4.6, 1.0 and 6.2 kg, respectively. The number of occasions on which plant harvesting was carried out during these seasons were 6, 10, 3 and 14, respectively. Litter production depends on the season (Wu and Mitsch, 1998; Scholz and Xu, 2002). Fresh biomass is usually produced in spring and summer, usually requires harvesting in late autumn. Furthermore, algae (predominantly Microspora spp.) require frequent removal in spring. In contrast to summer and winter, the maintenance was more efficient during autumn and spring because less labour time was required and more litter and algae were removed during these maintenance sessions. Fig. 4 indicates the elemental content of various aquatic plants. It can be seen that macrophytes located in the attenuation pond contain higher concentrations of calcium, magnesium and manganese in comparison to macrophytes located in the infiltration basin. Furthermore, Frogbit was associated with high elemental concentrations in general. This was also the case for barium, iron, magnesium, manganese and zinc.
278
M. Scholz / Bioresource Technology 95 (2004) 269–279
Fig. 5. Top figure: summary of Inductively Coupled Plasma Optical Emission Spectrometer (ICP-OES) results for the liquid phase of the pond system. Concentration readings for boron, barium, magnesium, manganese and zinc require division by a factor of 100, 10, 10, 100 and 100, respectively. Bottom figure: summary of ICP-OES results for the sediment phase of the pond system. Concentration readings for aluminium, boron, barium, copper, iron, magnesium, manganese and zinc require division by a factor of 10, 100, 10, 100, 10, 10, 100 and 100, respectively. GAC ¼ granular activated carbon.
All investigated aquatic plants were associated with detectable concentrations of magnesium, manganese and zinc. Measured concentrations of these elements for aquatic plants were relatively higher than for the liquid
phase (pond water) but comparable to associated sediment samples (Figs. 4 and 5). However, a direct comparison between different concentrations (mg/kg and mg/l) is difficult. Nevertheless, plant harvesting reduced
M. Scholz / Bioresource Technology 95 (2004) 269–279
279
trace nutrients and elements (in addition to BOD5 , SS, etc.) in the attenuation pond as indicated previously by Wu and Mitsch (1998).
and Research Centre sponsored the project. Forgmat International Holdings Ltd. contributed barley straw bales (Frogmat) for the control of algae.
4. Conclusions
References
The case study described the successful design, operation and maintenance of a novel storm water pond system during the first 15 months of operation. However, design guidelines would only have been acceptable if local environmental conditions (precipitation and infiltration patterns) had been fully considered or high safety factors (much higher than widely recommended) had been applied. Infiltration through the base of the infiltration basin was low (despite the presence of macrophytes) and should not have been considered during the design. The infiltration function for the infiltration basin was logarithmic and not linear as indicated by most international design guidelines. The water quality of the attenuation pond was acceptable for recycling. Rainwater runoff did not require extensive treatment. Seasonal and diurnal variation in water quality for BOD5 , phosphate (only evidence for diurnal variation provided), DO and pH were apparent. Excess algal and litter biomass weighted approximately 2.3 kg wet biomass/m2 of the attenuation pond surface area per annum. The biomass required mechanical removal in order to avoid eutrophication in the attenuation pond and low infiltration rates in the infiltration basin. Plant harvesting reduced BOD5 , nutrient and trace element levels within the pond system.
Abbott, C.L., Comino-Mateos, L., 2001. In situ performance monitoring of an infiltration drainage system and field testing of current design procedures. J. Ch. Instn. Wat. Envir. Mangt. 15 (3), 198– 202. ATV-DVWK-Arbeitsgruppe ES-41, 2002. Planung, Bau und Betrieb von Anlagen zur Versickerung von Niederschlagswasser (Design, construction and operation of rainwater drainage systems). ISBN 3-935669-83-6, ATV-DTWK (German Association for Water, Wastewater and Waste), Gesellschaft zur F€ orderung der Abwassertechnik e.V., Hennef, Germany, Regelwerk A-138. Bettes, R., 1996. Infiltration drainage –– manual of good practice. Construction Industry Research and Information Association (CIRIA), London, UK, CIRIA Report 156. Brix, H., 1997. Do macrophytes play a role in constructed treatment wetlands? Wat. Sci. Tech. 35 (5), 11–17. Building Research Establishment, 1991. Soakaway Design. Building Research Establishment (BRE) Bookshop, Watford, UK, BRE Digest 365 (replaces 151). Butler, D., Davies, J.W., 2000. Urban Drainage. E&FN Spon, London, UK. Clesceri, L.S., Greenberg, A.E., Eaton, A.D., 1995. Standard Methods for the Examination of Water and Wastewater, 19th ed. American Public Health Association (APHA)/American Water Works Association (AWWA)/Water Environment Federation (WEF), Washington, DC, USA. Ellis, J.B., D’Arcy, B.J., Chatfield, P.R., 2002. Sustainable urban drainage systems and catchment planning. J. Ch. Instn. Wat. Envir. Mangt. 16 (4), 286–291. Everall, N.C., Lees, D.R., 1997. The identification and significance of chemicals released from decomposing barley straw during reservoir algal control. Water Res. 31, 614–620. Martin, P., Turner, B., Waddington, K., Dell, J., Pratt, C., Campbell, N., Payne, J., Reed, B., 2000. Sustainable Urban Drainage Systems––Design Manual for England and Wales. Construction Industry Research and Information Association (CIRIA), London, UK, CIRIA C522. Meteorological Office, 2002. Bradford monthly long-term average rainfall. Online data between 1908 and 2001, URL address: http:// www.metoffice.gov.uk/climate/uk/stationdata/bradforddata.txt. Scholz, M., Xu, J., 2002. Performance comparison of experimental constructed wetlands with different filter media and macrophytes treating industrial wastewater contaminated with lead and copper. Biores. Technol. 83 (2), 71–79. Wu, X.Y., Mitsch, W.J., 1998. Spatial and temporal patterns of algae in newly constructed freshwater wetlands. Wetlands 18 (1), 9–20.
Acknowledgements The author acknowledges support from previous Final Year Project students (Mr. Robin O. Draycott, Mr. Alex Onyeador, Mr. Alexis O. De Burlet and Mr. Yann Legay). The technical assistance of Dr. Peter Anderson, Mr. Andy Gray and Mr. Anthony J.F. Daron has been appreciated. The University of Bradford, The University of Edinburgh, the Pennine Water Group and the Contaminated Land Assessment, Remediation
