Integrated Constructed Wetlands for Pollution Control

Integrated Constructed Wetlands for Pollution Control

Chapter 32 Integrated Constructed Wetlands for Pollution Control 32.1 INTRODUCTION The first research work related to the free-water surface integrat...

257KB Sizes 2 Downloads 95 Views

Chapter 32

Integrated Constructed Wetlands for Pollution Control 32.1 INTRODUCTION The first research work related to the free-water surface integrated constructed wetlands (ICW) database has been published elsewhere (Carroll et al., 2005; Dunne et al., 2005a,b), providing clear evidence that phosphorus removal correlates positively with an increase in wetland area. Current mass pollutant removal by surface flow constructed wetland treatment of agricultural dirty water usually varies between 30% and virtually 100% of the phosphorus input, depending on its type (Braskerud, 2002; Newman et al., 2000; Reddy et al., 2001). However, there are also some other references indicating insufficient phosphorus removal for constructed treatment wetlands. For example, constructed wetlands were not efficient in removing total phosphorus in Nordic countries, especially during periods of high loading rates (Smith et al., 2006). Good initial removal efficiency of total phosphorus, which, however, became less effectively removed after five years of operation, was reported by Tanner et al. (1999). The variation in wetland sizes and configurations makes it difficult to compare phosphorus removal performances. The litter from decaying macrophytes provides surface area for attachment of biofilms and is therefore important for microbial processes such as the transformation of nutrients in wetlands. For most aquatic systems, the bulk microbial conversions are undertaken by microorganisms immobilized in sediments, microbial mats, and biofilms attached to solid surfaces. Therefore, sediment and litter samples should be collected to gain an insight into the microbial transformations taking place in removing nutrients from ICW. The aim of Chapter 32 is to assess the ICW performance, putting special emphasis on the specific water treatment potential of selected ICW examples. The objectives are: l l l

l

to analyze and assess ICW performance; to assess the corresponding phosphorus and nitrogen removal capacity; to identify the presence or absence of ammonia-oxidizing and denitrifying groups of organisms; and to recommend ICW design details for subsequent guidelines.

Wetlands for Water Pollution Control. http://dx.doi.org/10.1016/B978-0-444-63607-2.00032-0 Copyright © 2016 Elsevier B.V. All rights reserved.

361

362

Wetlands for Water Pollution Control

32.2 CASE STUDY SITES, MATERIALS, AND METHODOLOGIES 32.2.1 Sites within the Case Study Catchment The entire Annestown-Dunhill case study catchment (near Waterford, Ireland) is divided into subcatchments serving each farmyard curtilage and comprises 19 working farms. In 1999 and 2000, 12 ICW (e.g., Figure 32.2.1.1) were designed, and construction was carried out to intercept and treat farmyard dirty water. Additionally, a sewage treatment ICW (site 7) was constructed to intercept the discharge from a septic tank (Table 32.2.1.1). The ICW comprised several wetland cells operated in series, and their total sizes ranged between approximately 3000 and 23,000 m2. Components of farmyard dirty water discharged to the wetlands were variable. The runoff typically consisted of yard and dairy washings and rainfall on open yard and farmyard roofed areas, along with silage and manure effluents. All ICW were in operation for at least seven years. ICW sites 1, 2, 5, 6, 7, 8, and 13 were subject to surface runoff and/or groundwater infiltration (Table 32.2.1.1). This led to reductions of toxic ammonia concentrations due to dilution, which supports healthy plant growth.

32.2.2 Sampling and Analytical Methods Grab samples for each wetland cell inlet and outlet were taken on an approximately fortnightly basis. Water analysis was conducted at the Waterford County Council water laboratory using standard methods (APHA, 1998). To better understand nutrient removal processes in different parts and components of selected ICW examples, sediment and litter samples were

FIGURE 32.2.1.1 Irish integrated constructed wetland example comprising four wetland cells and one polishing pond. Photo taken by Dr Rory Harrington.

TABLE 32.2.1.1 Site Characteristics of Farms and Corresponding Integrated Constructed Wetland Systems in the Anne Valley (near Waterford, Ireland), after Scholz et al. (2010) ICW No.

Farm Enterprise

Farmyard Area (m2)

Effective ICW Area (m2)

ICW-to-YardArea Ratio

1

Dairy

4500

3906

0.9

2

Dairy

14,750

22,966

1.6

3

Dairy; beef

5400

10,288

1.9

c

4

Dairy

9200

10,327

1.1

5

Dairy; tillage

4000

3940

1.0

6

Dairy

9800

12,691

1.3

7

Sewage

n/a

3075

n/a

8

Beef

2300

3940

2.0

9

Mixed

4800

7964

1.7

10

Mixed

2100

4375

2.1

11

Dairy

5000

7676

1.5

d

12

Mixed

13,600

10,748

0.8

13

Sheep; tillage

5000

5610

1.1

ICW No.

Aspect Ratio of ICW Cellsa

Dairy Washings (Cow Number)

1

2.0

2

b

Roof Water

Spring Water Input

Yes (60)

Yes

Yes

3.1

Yes (60)

Yes

Yes

3

9.0

Yes (50)

Yes

No

4

6.6

Yes (100)

Yes

No

5

4.3

Yes (35)

Yes

Yes

6

4.7

Yes (80)

No

Yes

7

4.0

n/a

No

Yes

8

2.5

No

Yes

Yes

9

3.0

Yes (55)

Yes

No

10

3.0

Yes (50)

Yes

No

11

5.6

Yes (77)

Yes

No

12

3.6

Yes (85)

Yes

No

13

2.3

No

Yes

Yes

n/a ¼ not applicable. a Nonvegetated first and last cells were excluded. Total cell numbers for ICW systems 1e13 were 9, 4, 6, 6, 4, 6, 4, 5, 5, 5, 4, 7, and 6, respectively. b A “Yes” entry indicates discharges to the wetland and a “No” entry indicates that there were none. c 500 m2 added in 2003. d 2000 m2 and 1100 m2 added in 2004 and 2005, respectively.

364

Wetlands for Water Pollution Control

collected in April and May 2007 from three different wetlands (ICW 3, 9, and 11), frozen, and sent off to Linko¨ping University (Sweden) for subsequent molecular microbiological analysis. Samples from ICW 3 and 9 were taken on April 24, 2007, while samples for ICW 11 were taken on May 8, 2007. Mostly triplicate field sediment and litter samples were collected 1 m to the left, 1 m in front, and 1 m to the right of each sampling point located near the inlet of each ICW cell and the outlet of each ICW system. Sediment samples were collected with a sediment sampler (diameter 4 cm). Green plant material was removed from litter samples before collection. Samples were stored at 20  C before analysis.

32.2.3 Deoxyribonucleic Acid Extraction Sediment and litter samples were subjected to deoxyribonucleic acid (DNA) extraction using a FastDNAÒ SPIN kit for Soil (Bio 101, Inc., La Jolla, CA, USA). Samples (0.25 g) were suspended in a sodium phosphate buffer supplied with the FastDNAÒ SPIN kit as stipulated by the manufacturer and homogenized for 180 s with a hand-held blender (DIAX 900 Homogeniser Tool G6, Heidolph, Kelheim, Germany). DNA was extracted from soil samples by bead beating, a procedure in which soil aggregates are disrupted and bacterial cells are lysed mechanically. Bead beating was extended to 3  30 s to achieve good homogenization of the samples. The subsequent centrifugation was prolonged to 2  5 min and the centrifugation after washing with SEWS-M, a salt and ethanol wash solution (Qbiogene, Inc., USA), was extended to 5 min. The extracted DNA was stored at 20  C.

32.2.4 Polymerase Chain Reaction The ammonia-oxidizing bacterial community was investigated using groupspecific polymerase chain reaction (PCR) primers targeting the 16S ribosomal ribonucleic acid (rRNA) gene, while the denitrifying bacterial community was assessed using the functional gene nitrous oxide reductase (nosZ), which is the gene for the terminal enzyme in denitrification (Hallin et al., 1999; Sundberg et al., 2007). A primer is a nucleic acid strand that serves as a starting point for DNA replication and is required because most DNA polymerases (i.e., enzymes that catalyze the replication of DNA) cannot synthesize a new DNA strand from scratch; rRNA is one of the three major types of ribonucleic acid (RNA) and is part of ribosomes and composed of RNA of different sizes, such as 5S, 16S, and 23S in prokaryotes. The extracted DNA from all samples was diluted 10-fold to avoid inhibition of the PCR by humic substances. This was determined by testing for different dilution ratios. PCR amplification was undertaken using forward and reverse primers for ammonia-oxidizing bacteria. The PCR was performed on a

Integrated Constructed Wetlands for Pollution Control Chapter j 32

365

PTC-100TM thermal cycler (MJ Research Inc., San Francisco, CA, USA) in a 50 ml mixture (Sundberg et al., 2007b). For the nitrous oxide reductase, the forward and reverse primers targeting the nosZ gene were used in the subsequent PCR. The PCR was performed on a PTC-100TM thermal cycler in a 50 ml mixture including 1.33 U of Taq polymerase and 5 ml of the supplied buffer (including 1.5 mM MgCl2; Roche Diagnostic GmbH, Mannheim, Germany), each nucleotide at a concentration of 200 mM, the primers at 0.125 mM each, 600 ng/ml bovine serum albumin, and 2 ml of the DNA template (Sundberg et al., 2007a).

32.2.5 Agarose Gel Electrophoresis The PCR products of DNA extraction and PCR reactions were examined by agarose gel electrophoresis. The agarose was melted by heating the mixture (agarose and buffer), and subsequently pouring it in the agarose gel casting tray. The solidified gel was covered with an electrophoresis buffer before running electrophoresis. The electrophoresis buffer was the same as the one used to prepare the agarose. The PCR products and dye supplied with the DNA extraction kit (2 ml of dye and 4 ml of PCR products) were placed into the loading wells formed by the gel comb. The first well of each row was loaded with 2 ml of Gene Ruler (1 kb DNA ladder; 1000 base pairs for ammonia-oxidizing bacteria and nosZ) and 4 ml of distilled water. The electrophoresis was run for 40 min at 120 V (Owl Scientific, Inc., Woburn, MA, USA). The gel was then placed in ethidium bromide solution (immersed for 15 min) located in the fume cupboard and washed subsequently with tap water. The ethidium bromide-stained gel was then visualized by ultraviolet illumination.

32.3 RESULTS AND DISCUSSION 32.3.1 Wetland Water Quality Water quality data for all wetland influents and effluents for the monitoring period between August 2001 and August 2007 were analyzed. Molybdate reactive phosphate (MRP) concentration reductions were generally greater than 90% (Table 32.3.1.1), with the most notable exception of ICW site 7, which is likely to be due to the high hydraulic loading rates and its similarity with a sewage treatment works. Owing to the high loading rates and treatment of domestic wastewater from a relatively high population of approximately 200, it exhibited only a 30% MRP reduction (i.e., an indication for an undersized plant). No obvious signs of phosphate saturation along with wetland maturation were observed. There were also no significant differences between cold (winter) and warm (summer) months (seasons). Phosphorus is likely to accumulate owing to soil sorption and other processes within the sediment and detritus of

366

ICW Site Number Flow

Statistic

1

2

3

4

5

6

7

Inflow

Max.

918.00

28.30

31.70

60.00

99.30

40.00

16.40

Min.

0.20

0.20

0.05

1.45

0.01

0.20

0.65

Mean

75.69

15.46

15.22

22.32

14.33

10.76

7.51

SD

198.91

6.734

9.730

11.515

20.613

9.202

3.211

N

25

28

72

82

25

27

33

Max.

1.30

0.73

11.47

4.70

3.01

0.96

10.35

Min.

0.01

0.01

0.02

0.02

0.01

0.01

0.16

Mean

0.22

0.27

3.43

1.66

0.24

0.13

5.25

SD

0.384

0.231

2.334

0.940

0.639

0.218

2.977

N

28

28

76

64

28

28

64

99.71

98.22

77.46

92.56

98.34

98.83

30.08

Outflow

% Phosphate reduction

Wetlands for Water Pollution Control

TABLE 32.3.1.1 Summary Data for Molybdate Reactive Phosphate (mg/l), after Scholz et al. (2010)

ICW Site Number Statistic

8

9

10

11

12

13

Inflow

Max.

5.00

124.71

42.80

50.96

250.00

4.49

Min.

0.05

0.01

0.14

0.35

0.91

0.02

Mean

1.46

8.57

7.16

11.55

43.53

0.94

SD

1.250

13.754

9.357

10.05

58.061

1.413

N

26

98

69

131

37

21

Max.

0.29

1.59

0.36

4.41

5.22

0.32

Min.

0.00

0.03

0.01

0.00

0.01

0.00

Mean

0.04

0.47

0.06

0.94

0.74

0.06

SD

0.063

0.307

0.094

0.628

0.916

0.089

N

27

74

24

131

65

24

97.21

94.52

99.16

91.86

98.30

93.34

Outflow

% Phosphate reduction

min. ¼ minimum; max. ¼ maximum; SD ¼ standard deviation; N ¼ number of entries.

Integrated Constructed Wetlands for Pollution Control Chapter j 32

Flow

367

368

Wetlands for Water Pollution Control

the wetland cells (i.e., sinks for pollutants), as discussed by Carroll et al. (2005) and Dunne et al. (2005a). Plants were not harvested to remove phosphorus. Concerning the inflow mean water quality values, chemical oxygen demand (COD) was 1813  3763 mg/l, standard five-day biochemical oxygen demand (BOD) was 539  1041 mg/l, suspended solids (SS) was 679  2856 mg/l, ammonia-nitrogen was 61  124 mg/l, MRP was 15  21 mg/l, and there were 321,600  384,800 Escherichia coli colony-forming units (CFU) per 100 ml. This indicates very high water quality variability concerning the farmyard dirty water discharged to the ICW systems. The mean effluent water quality was 78  73 mg/l COD, 18  21 mg/l BOD, 24  46 mg/l SS, 1.1  2.6 mg/l ammonia-nitrogen, 1.3  1.6 mg/l MRP, and 228  306 Escherichia coli CFU per 100 ml during the monitoring period. This suggests that most ICW systems had a high capacity to remove pollutants. If the two farm sites (ICW sites 8 and 13) without dairy washings had been excluded, the BOD concentration of water entering the other ICW systems would have been approximately 1800 mg/l. This value is consistent with soiled water such as dairy washings diluted with cleaner water, including yard and roof runoff (European Community, 2006).

32.3.2 Flows Hydraulic fluxes from extraneous sources were minimal at ICW 11. Therefore, flow meters were installed at the inlets to cells 1 and 2 and the outlet at this example site. Mean daily inflows and outflows from the ICW system during the monitoring period (April 2003eJuly 2007) were 5.97  7.35 m3/day and 1.59  4.47 m3/day, respectively. Although there was usually inflow to the ICW system between August and December, there was no outflow from it during most of these periods. It follows that there was only a short discharge period to receiving surface waters. The patterns of flow continuously recorded at ICW 11 were similar to those estimated for other ICW systems, where instantaneous flows were measured manually. Decreases in flow through the wetlands were recorded during the summer, and none of the 12 ICW systems were discharging to receiving watercourses. During dry periods, the loss of water from ICW due to evapotranspiration and slow infiltration created an increased storage capacity (freeboard) within most ICW cells. This provided a buffering capacity for wet weather periods such that waters are stored within the ICW system prior to discharge during storm events. Furthermore, groundwater contamination due to the infiltration of partially treated runoff was not observed in nearby wells.

32.3.3 Design Considerations The variations in ICW design and the observed variations in effluent quality allowed examination of the influence of some key design factors on water

Integrated Constructed Wetlands for Pollution Control Chapter j 32

369

treatment performance. Mean effluent MRP concentrations (Table 32.3.1.1) were considered to be an appropriate indicator of the treatment performance. The ICW were built with a mean size of 1.3 times the farmyard area and with a minimum of four cells. Most (10 out of 12) of the ICW built to this design produced a good-quality effluent with MRP concentrations below the nominal target concentration of 1 mg/l (Table 32.3.1.1). Thus the sizing of at least 1.3 times interception area and a minimum of four (approximately equally sized) vegetated cells are taken to be the primary design criteria. ICW systems 3 and 4, which did not meet the desired effluent quality, had long, narrow ponds (i.e., high individual aspect ratio). The mean aspect ratio is defined as the mean length of the wetland system divided by the mean width. In this study, individual aspect ratios ranged from 2 to 9. A linear regression analysis (MRP ¼ 0.45  aspect ratio) between the aspect ratio and the effluent MRP concentrations had a coefficient of determination of 0.55 (P < 0.05). An aspect ratio of less than 4 resulted in good MRP removal. These are considered to be the minimum criteria for good runoff treatment performance (i.e., effective nutrient removal). Ideally, for system robustness, farm ICW should be built to a larger relative size (twice the interception area) and with more than four cells. The observed impact of the ICW total cell number, cell sizing, and cell aspect ratio on effluent quality reflects the fact that the effluent should stay as long as possible in the ICW (and also have a flow velocity as low as possible) to allow the ICW ecosystems to develop, thrive, and be able to remove pollutants (Keffala and Ghrabi, 2005; Scholz, 2006). This is achieved by round wetland cells (which are, however, not economical if land costs are high) and low gradients, thus allowing the pollutant plume to spread as slowly as possible throughout each ICW cell.

32.3.4 Nitrogen Transformation Processes In comparison to ammonia oxidizers, denitrifiers were more abundant in most of the collected litter and sediment samples (Table 32.3.4.1). Since the nitrate concentrations within the ICW systems were low, it is likely that oxygen and nitrate have served as electron acceptors in the lower layer of the wetland cells, and this might have promoted the growth of denitrifying bacteria. Each ICW system contained denitrifying bacteria, but they were present in varying quantities. For example, ICW 11 had lower denitrifying bacteria numbers than both ICW 3 and ICW 9. Samples analyzed for ICW 3 and ICW 9 did not indicate the presence of ammonia-oxidizing bacteria. Most denitrifiers are heterotrophs. The supply of organic carbon by macrophytes raised the overall heterotrophic activity, leading to the consumption of oxygen. Thus, oxygen availability in the sediment was reduced, and subsequently denitrification was supported (Bastviken et al., 2005).

370

Wetlands for Water Pollution Control

TABLE 32.3.4.1 Bacterial Community Abundance in Integrated Constructed Wetland Litter and Sediment Samples Compared to Corresponding Mean Ammonia-Nitrogen and NitrateeNitrogen Concentrations within the Top Water Phase, after Scholz et al. (2010) ICW No.

AmmoniaOxidizing Bacteriaa

Denitrifying Bacteriaa

AmmoniaNitrogen (mg/l)

Nitratee Nitrogen (mg/l)

3

0

73

13.95

0.27

9

0

80

5.20

1.69

11

27

53

9.73

1.73

Relative presence (%) of bacterial community (0 ¼ absent; 100 ¼ present) for multiple samples.

a

Ammonia-oxidizing bacteria were present at ammonia-nitrogen concentrations between approximately 5 and 20 mg/l. Denitrifying bacteria were observed at nitrateenitrogen concentrations between 0.1 and 4.5 mg/l. In comparison to ammonia-oxidizing bacteria, more denitrifiers were present in most ICW systems, while ammonia-oxidizing bacteria were found in samples collected from ICW 11 only.

32.3.5 Comparison of Ammonia Oxidizers and Denitrifiers Interesting differences between three selected ICW example sites are discussed in this section (see also Table 32.3.4.1). For example, concerning ammoniaoxidizing bacteria, samples analyzed from ICW 3 and ICW 9 did not indicate the presence of ammonia-oxidizing bacteria. Concerning denitrifying bacteria, ICW 3 had lower denitrifying bacteria numbers than ICW 9. Furthermore, ICW 9 had a higher aquatic plant cover density than ICW 3. It follows that the decaying macrophytes within ICW 9 contributed to organic matter, which became a source of carbon and therefore energy for denitrifying bacteria. Concerning ICW 9 versus ICW 11, samples obtained from ICW 9 did not indicate the presence of ammonia-oxidizing bacteria, while samples taken from ICW 11 indicated the presence of these bacteria. There was a reduced availability of organic matter at the bottom of ICW 11, which led to decreased numbers of heterotrophic bacteria and consequently created conditions in which ammonia-oxidizing bacteria proliferated. Regarding denitrifying bacteria, ICW 9 had higher denitrifying bacteria numbers than ICW 11. Furthermore, ICW 9 had a higher plant density than ICW 11. The high number of denitrifying bacteria was linked to high concentrations of nitrate. The decaying plants also contributed to organic matter that became a source of energy for denitrifying bacteria.

Integrated Constructed Wetlands for Pollution Control Chapter j 32

371

32.4 CONCLUSIONS AND RECOMMENDATIONS Findings suggest that integrated constructed wetlands were capable of treating farmyard dirty water and that they provided a sustainable management option to effectively reduce nutrient and contaminant loss from farmyards to watercourses. Significant concentration reductions in suspended organic matter, nutrients, and fecal bacteria between ICW influents and effluents were observed. Surface discharges from the ICW sites had seasonal patterns. None of the farm ICW had surface discharges during most summer months. In terms of ICW design for the southeast of Ireland, the effective ICW area should be sized from the farmyard area by applying a multiplying factor of at least 1.3 to the latter. The optimal theoretical ICW cell aspect ratio (to achieve an effluent MRP concentration below 1 mg/l) was approximately 1:1; i.e., the ICW wetland cells become relatively round. The small aspect ratio is specific for ICW comprising planted cells, which are characterized by very low flow velocities. The recommended minimum number of cells within each ICW was four. However, constructing a greater number of cells with greater surface areas was advised, as both the treatment performance and the aesthetic appeal of the ICW system increases with increasing total wetland area. The proposed design results in relatively large treatment systems, which are unsuitable if land prices are high (e.g., inner-city areas). However, the absence of both a liner as well as standard engineering control and operational structures makes these systems particularly affordable for farmers. The number of denitrifying bacteria detected in different ICW systems was higher than the number of ammonia-oxidizing bacteria. The high vegetation density impeded the water flow at the surface. Anoxic conditions at the wetland bottom (coupled with a high organic carbon matter content available from the decaying macrophytes) created conditions that were favorable for denitrifying bacteria. It follows that sediment and litter within the ICW supported denitrification. Further research is required on decreasing the costs for molecular microbiological analysis to make the proposed techniques more widely available. Too little is also known about the ecology and engineering application of many new microbes that have been discovered since 2005. Engineers are usually only interested in organisms that pose a risk to human health or that can be used to treat waters.