Nutrient removal from polluted stream water by artificial ... - Springer Link

3 downloads 0 Views 438KB Size Report
May 9, 2009 - phic stream water polluted by non-point sources, an artificial aquatic food ... tank of 3 m3 capacity, and one zooplankton growth chamber of 1.5 ...
Hydrobiologia (2009) 630:149–159 DOI 10.1007/s10750-009-9788-7

PRIMARY RESEARCH PAPER

Nutrient removal from polluted stream water by artificial aquatic food web system Dawoon Jung Æ Ahnna Cho Æ Young-Gun Zo Æ Seung-Ik Choi Æ Tae-Seok Ahn

Received: 26 November 2008 / Revised: 3 April 2009 / Accepted: 14 April 2009 / Published online: 9 May 2009 Ó Springer Science+Business Media B.V. 2009

Abstract For the removal of nutrients from eutrophic stream water polluted by non-point sources, an artificial aquatic food web (AAFW) system comprising processes of phytoplankton growth and Daphnia magna grazing was developed. The AAFW system was a continuous-flow system constructed with one storage basin of 3 m3 capacity, one phytoplankton tank of 3 m3 capacity, and one zooplankton growth chamber of 1.5 m3 capacity. The system was optimized by setting hydraulic retention time of phytoplankton tank as 3 days and D. magna density as 740–1000 individual l-1. When the system was operated on eutrophic stream water that was delivering 471 g of total nitrogen (TN) and 29 g of total phosphorus (TP) loadings for 45 days, 250 g (53%) of TN and 16 g (54%) of TP were removed from the water during its passage through the phytoplankton tank. In addition, 64 g (14%) of TN and 4 g (13%) of TP were removed from the water by harvesting zooplankton biomass in the zooplankton growth chamber, resulting in significant overall removal

Handling editor: J. Padisak D. Jung  A. Cho  Y.-G. Zo  T.-S. Ahn (&) Department of Environmental Science, Kangwon National University, Chuncheon 200-701, Republic of Korea e-mail: [email protected] S.-I. Choi Institute of Environmental Research of Kangwon National University, Chuncheon 200-701, Republic of Korea

rates of TN (69%), nitrate (78%), TP (73%), and dissolved inorganic phosphorus (94%). While the removal efficiency of the AAFW system is comparable to those of other ecotechnologies such as constructed wetlands, its operation is less limited by the availability of space or seasonal shift of temperature. Therefore, it was concluded that AAFW system is a highly efficient, flexible system for reducing nutrient levels in tributary streams and hence nutrient loading to large aquatic systems receiving the stream water. Keywords Non-point source  Artificial food web system  Nutrient removal  Phytoplankton  Daphnia magna  Ecological engineering  Biomanipulation

Introduction Eutrophication is one of the most pervasive water quality problems in most urban and farming communities. At present, increasing levels of nitrogen and phosphorus loading, originating from intensive agriculture, poultry, and urbanization, are major factors forcing eutrophication of lakes and rivers (Leeben et al., 2008). For the prevention of eutrophication, the removal of phosphorus from waste water is essential, and there are several advanced treatment systems available for this purpose: Bardenpho, AO, A2O,

123

150

UCT, and Phostrip (Environmental Protection Agency, 1987). However, applications of these treatment systems are limited mainly to point sources such as sewage, and they are space costing and economically expensive for construction and operation (Xiaolian et al., 2006). Nowadays, non-point source pollution has been becoming the major reason for eutrophication, as point sources are relatively well treated in many countries (Jiang et al., 2007). However, it is very difficult to manage a non-point source, because its pollutants are coming from many diffuse sources at relatively low concentrations. In order to achieve compliance with water quality standards for river and lake, removing nutrients from polluted tributary streams heading into large river or lake is necessary. For this purpose, ecotechnologies, such as constructed wetland (Braskerud, 2003), artificial floating islands (Somodi & Botta-Duka´t, 2004), floating mats (Azza et al., 2006), and biomanipulation (Olin et al., 2006), were developed. Among them, the constructed wetland system is one of the typical ecotechnologies for the removal of nutrients, and its usefulness has been proven by several studies (Greenway & Simpson, 1996; Shutes, 2001; Coveney et al., 2002). However, a constructed wetland needs large area and is difficult to operate in rainy or winter season. In order to overcome these drawbacks, we designed an artificial aquatic food web (AAFW) system for the removal of nutrients from polluted stream water, a typical non-point source of river and lake pollution. The design is a revision of the AAFW system developed by Kim et al. (2003), which is proven to remove nutrients effectively from sewages containing high concentrations of nutrients. In this study, the AAFW system was applied to stream water containing relatively low concentrations of nutrients. The continuous-flow AAFW system developed in this study is based on a simple aquatic food chain, each trophic process of which is controlled to occur in separate water bodies. At the first process, nutrients are transformed into biomass of phytoplankton, which is subsequently transferred into zooplankton biomass by grazing activities of Daphnia magna, a cladoceran. Finally, the nutrients are removed from the system by disposing phytoplanktonic organisms and D. magna biomass, and sludge. In this study, we optimized hydraulic retention time (HRT) for phytoplankton and population density for D. magna, by

123

Hydrobiologia (2009) 630:149–159

investigating the growth parameters of phytoplanktonic organisms and D. magna under conditions simulating the operation of the AAFW system. In order to evaluate its efficacy, the system was operated for 45 days, and steady and efficient removal of nitrogen and phosphorus from stream water was confirmed.

Materials and methods Characteristics of target stream The target stream used for the construction of the AAFW system and nutrient removal experiments was located in Sachang-ri, Hwacheon-gun, Korea (388030 39.5000 N, 1278310 21.2800 E). The stream is known to receive poultry wastes and septic tank pollutions. Total nitrogen (TN), nitrate (NO3-), total phosphorus (TP), and dissolved inorganic phosphorus (DIP) concentrations in the stream water during the period of experiments (April 24–June 8, 2004) were 10.5 ± 0.1 (mean ± SD), 8.7 ± 0.2, 0.7 ± 0.1, and 0.6 ± 0.1 mg l-1, respectively. The ratio of nitrate to TN ranged from 65 to 94%, and the ratio of DIP to TP from 65 to 98% (Table 1).

Table 1 Mean and range of water quality variables observed in the target stream during operation of the artificial aquatic food web system Itema

Minimum

Maximum

Mean

Temp (°C)

12.5

26.7

19.3

pH

7.2

8.5

7.8

DO (mg l-1)

2.9

5.6

3.9

BOD5 (mg l-1) TN (mg l-1)

6.3 8.4

13.2 12.8

10.1 10.5

NO3- (mg l-1)

7.0

11.5

8.7

NO3-/TN (%) -1

65

94

83

TP (mg l )

0.49

0.85

0.65

DIP (mg l-1)

0.40

0.79

0.56

TP/DIP (%)

65

98

86

TN/TP

11.3

21.0

16.4

a

DO the concentration of dissolved oxygen, BOD5 the amount of dissolved oxygen consumed in 5 days by biological processes breaking down organic matter, TN total nitrogen, NO2 3 nitrate, TP total phosphorus, DIP dissolved inorganic phosphorus

Hydrobiologia (2009) 630:149–159

151

Facilities for AAFW system The artificial food web system consisted of one storage basin, one phytoplankton tank, and one zooplankton growth chamber, arranged sequentially in downward steps (Fig. 1). The system was designed by adopting the basic aspects of a laboratory-scale AAFW system that was developed by Kim et al. (2003). The storage basin was a cylindrical container with a holding capacity of ca. 3 m3 (Ø = 1.48 m, d = 1.7 m). Stream water was supplied by a drain pump at the velocity of 100 l per min. The major function of the storage basin was to dampen fluctuations in quality and quantity of the input water. The phytoplankton tank was a cylindrical tank of ca. 3 m3 (Ø = 2.6 m, d = 0.75 m) storage capacity. For the collection of sludge, the floor of the tank had a 10° slope toward the center of the tank bottom. On every 5 days, ca. 10 l of sludge at the bottom of the tank was drained out. For the optimum growth of phytoplankton, water in the phytoplankton tank was circulated at the velocity of 100 l min-1 by a water pump during the day time. The zooplankton chamber was a rectangular chamber of about 1.5 m3 capacity (w = 1.2 m 9 2.4 m, d = 0.6 m), which had a 10° vertical slope from the inlet to outlet for the collection of sludge. Water temperature of the zooplankton chamber was kept as 18°C ± 5°C (mean ± SD) with an aquarium heater. A shade was placed on the top side of the chamber, to avoid irradiation by direct sunlight. About 5 l of sludge at

the bottom of the chamber was drained out on every 5 days. The effluent was filtered through the sand and gravel layers to prevent the loss of individuals of D. magna population. Kinetics of phytoplankton growth In order to prepare phytoplankton inoculum, 5 l of domestic sewage was mixed with the same volume of lake water from Lake Ui-am, Chuncheon, Korea in a 10-l transparent glass bottle, and incubated under sunlight at 19.5 ± 2.5°C for 1 week with intermittent 15-min stirring at 2-h intervals. Ten liters of phytoplankton inoculum was mixed into 2 m3 target stream water in the phytoplankton tank of the AAFW system. The growth rate and the carrying capacity of phytoplankton in the stream water were assessed by monitoring the changes in the chlorophyll a concentration in the tank water, without input or output of water, for 16 days. The optimum HRT of the phytoplankton tank for the continuous-flow AAFW system was determined based on the growth curve of phytoplankton which was represented as a change in the chlorophyll a concentration during this batch culture. Kinetics of zooplankton growth We employed D. magna, a cladoceran, as the herbivorous zooplankton to feed on assemblages of phytoplankton in the effluent from the phytoplankton

S

P

[Storage basin] V : 3 ton

[Phytoplankton tank] V : 3 ton

S

Solar energy

P

Pump

H

Heater

H

Eff.

[ D. magna chamber] V : 1.5 ton

Fig. 1 Schematic diagram of artificial aquatic food web system

123

152

Hydrobiologia (2009) 630:149–159

tank. In order to determine the optimum density of D. magna in the zooplankton chamber of the continuous-flow AAFW system, growth curve of D. magna in a laboratory batch culture was obtained. D. magna was added into 10 l (w = 18 cm 9 30 cm, d = 20 cm) of water taken from the 5-day-old phytoplankton batch culture that was established by the procedures described above. The initial density of D. magna was adjusted to 225 ± 70 individual l-1. The culture was incubated in the dark at 20°C ± 0.2°C for 10 days and supplied with 83 ± 6 mg m-3 of phytoplankton biomass daily.

sewage and lake water in laboratory for 1 week, was 124 mg m-3. Since the 10-l inoculum was mixed into 2 m3 of target stream water, chlorophyll a concentration in the phytoplankton tank increased, showing a typical growth curve of a batch culture (Fig. 2). The maximum level of chlorophyll a was observed on the third day, as 89 mg m-3. Since then, it remained stable with the mean of 83 ± 6 mg m-3 until the 16th day. The increment of chlorophyll a concentration during the growth phase (days 0–3) was 19– 24 mg m-3 day-1. Based on these results, the HRT in the phytoplankton tank was set to 3 days.

Continuous-flow AAFW system

Kinetics of zooplankton growth

The continuous-flow AAFW system was operated with the target stream water for 45 days after the establishment of the phytoplankton community and acclimation of the D. magna population for 15 days. During the period of operation, water temperature was 20.5°C ± 5.4°C in the phytoplankton tank, and 17.8°C ± 3.7°C in the zooplankton chamber. The HRT of the phytoplankton tank was set to 3.0 days, based on the phytoplankton growth curve. Density of D. magna in the zooplankton chamber was maintained from 740 to 1,000 individual l-1 by harvesting D. magna standing crop daily. For harvesting of D. magna, the entire area of the zooplankton chamber was partitioned into ten parts, and D. magna biomass in 35–40% of the total area was removed by a net with nominal cut-off size of 5.5 lm. In order to determine the efficacy of the AAFW system in removing the nutrients from stream water, water samples were taken from the storage basin, the phytoplankton tank, the zooplankton chamber, and the effluent at 14:00 of a day at 3-day intervals, and analyzed for TN, nitrate, TP, DIP, and chlorophyll a concentrations by the standard methods (APHA, 2001). Phytoplankton species were identified to species level by the method of Mizuno (1975). Measurements were taken in triplicate for all water quality variables.

The growth curve of D. magna grown on phytoplankton preys, raised in the batch culture described above, is shown in Fig. 3. The number of D. magna was 225 individual l-1 initially and increased to 1,005 individual l-1 on the fourth day. Then, the growth of D. magna population entered into a stationary phase maintaining a mean abundance of 1,125–1,245 individual l-1. During the growth phase (days 0–4), the mean growth rate of the D. magna population was 38% day-1. Based on these results, the density of D. magna in the continuous-flow system was set to be controlled

Chlorophyll a(mg m -3 )

90 K 80 70 60 50 40 30 20 0

3

5

10

15

Time (days)

Results Kinetics of phytoplankton growth The chlorophyll a concentration of the phytoplankton inoculum, prepared by incubating the mixture of

123

Fig. 2 Variation of chlorophyll a concentration in the phytoplankton batch culture. Non-linear regression on daily triplicate measurements (open circles) with a logistic growth model yielded the regression line (solid line). The carrying capacity (K; gray dotted line) determined by the regression model was 84 ± 1 mg mg-3, and intrinsic growth rate (r) was 1.4 ± 0.2 day-1. The gray solid line indicates the time point for 95% of carrying capacity

Hydrobiologia (2009) 630:149–159

153 14

1400

K

1000

83% K

10

800

TN (mg l -1)

-1 D. magna (ind. l )

12

1200

62% K

8 6

600

4

400

2 0

200

14

0

2

4

6

8

10

12

Fig. 3 Variation in the abundance of Daphnia magna in the laboratory culture. Non-linear regression on daily triplicate measurements (open circles) with a logistic growth model yielded the regression line (solid line). The carrying capacity (K; gray dotted line) determined by the regression model was 1200 ± 27 ind. l-1, and intrinsic growth rate (r) of D. magna population was 1.02 ± 0.14 day-1. The gray solid lines indicate the upper and the lower limit of D. magna abundance maintained during the operation of the continuous-flow system -1

at 740–1,000 individual l by harvesting about 35– 40% of D. magna in the zooplankton chamber. The zooplankton population was expected to sustain its growth phase by this harvesting scheme.

NO3- (mg l -1)

Time (days) 10 8 6 4 2 0 0 3

6

9 12 15 18 21 24 27 30 33 36 39 42 45

Time (days)

Fig. 4 Variations of total nitrogen (TN) and nitrate (NO3-) during operation of the continuous-flow AAFW system. Black circle storage tank, inverted triangle phytoplankton tank, and open circle effluent from system

Continuous-flow AAFW system The concentrations of TN and NO3- in the storage basin averaged 10.5 and 8.7 mg l-1, respectively, whereas those in the effluent of the continuous-flow system were 3.3 and 1.9 mg l-1, respectively (Fig. 4). The concentrations of TP and DIP in the storage basin were 0.65 and 0.56 mg l-1, respectively, and decreased to 0.17 and 0.03 mg l-1 in the effluent, respectively (Fig. 5). Based on these mean values, the mean removal efficiencies of TN, NO3-, TP, and DIP were calculated to be 69, 78, 73, and 94%, respectively. The spatial separation of biological processes in the artificial food chain was obvious from vivid contrast in the pH values of waters in different containers. While the mean pH of the target stream was 7.7 ± 0.3, pH values in the phytoplankton tank and the zooplankton chamber were 10.8 ± 0.2 and 8.4 ± 0.3, respectively. The marked decrease in the pH value from the phytoplankton tank to the zooplankton chamber

indicated negligible level of primary production in the zooplankton chamber. The concentration of chlorophyll a was 84.4 ± 12 mg m-3 in the phytoplankton tank and 8.1 ± 4 mg m-3 in the effluent of the zooplankton chamber (Fig. 6). About 90% of the phytoplankton biomass, when represented as chlorophyll a, was grazed by herbivores in the zooplankton chamber. During the 45 days of operation, D. magna sustained the density of 894 ± 137 individual l-1 in the zooplankton chamber. One of the basic features of the AAFW system was the spatial separation of phytoplankton growth and zooplankton grazing. This design was adopted to provide resilience of the aquatic food web. The food chain from phytoplankton to D. magna may collapse due to overgrazing or starvation of D. magna, caused by difference between phytoplankton growth rate and zooplankton grazing rate. Also, the spatial separation was expected to prevent the emergence of phytoplankton species resistant to D. magna grazing, which

123

TP (mg l -1)

154

Hydrobiologia (2009) 630:149–159 1.0

Table 2 Abundance (individual l-1) of phytoplankton species in the phytoplankton tank

0.8

Species name

Day 1

Day 30

0.6

Cyanobacteria Oscillatorean sp.

0.4

1.5 ± 0.1 9 103 2.8 ± 0.2 9 106

Bacillariophyceae 0.2

0.0 1.0

DIP (mg l -1)

Abundance (mean ± SD)a

Asterionella gracillima

B100

ND

Nitzschia frustulum

B100

ND

Chlorophyceae Chodatella sp.

B100

ND

0.8

Crucigenia tetrapedia B100

ND

Scenedesmus acuminatus

B100

0.6

ND

Scenedesmus arcuatus B100 0.4

Scenedesmus longispina

0.2

ND

1.8 ± 0.1 9 107 1.8 ± 0.1 9 105

a

ND not detected in 10-ml sample, B100 detection of one individual in 10-ml sample

0.0 0

3 6

9 12 15 18 21 24 27 30 33 36 39 42 45

Time (days)

1200 1100 1000 900 800 700 600 500 400 300 200 100 0

200

Chlorophyll a (mg m -3)

180 160 140 120 100 80 60 40 20 0

D.magna (ind. l -1)

Fig. 5 Variations of total phosphorus (TP) and dissolved inorganic phosphorus (DIP) during operation of the continuousflow AAFW system. Black circle storage tank, inverted triangle phytoplankton tank, and open circle effluent from system

0 3 6 9 12 15 18 21 24 27 30 33 36 39 42 45

Time (days)

Fig. 6 Variations in chlorophyll a concentration and abundance of Daphnia magna during operation of the continuousflow AAFW system. Black circle chlorophyll a concentration in the phytoplankton tank, open circle chlorophyll a concentration in the effluent, and inverted triangle number of D. magna in the zooplankton chamber

might occur by natural selection if the two processes occur in the same space. The high removal rate of chlorophyll a in the zooplankton chamber indicated

123

the absence or negligible presence of grazing-resistant phytoplankton species. In order to identify the phytoplankton species assuring high clearance rate by D. magna grazing, we analyzed the species composition of the phytoplankton community in the phytoplankton tank. A succession was observed when the compositions of the initial phytoplankton community and that on the 30th day were compared (Table 2). On the first day, the predominant species was Scenedesmus longispina (Chlorophyceae). On the day 30, a filamentous oscillarorean Cyanobacterium species was predominating over the Scenedesmus species. However, the cyanobacterium species did not form long filaments; typical dimensions were: width, 5 ± 0.01 lm; length, 51 ± 27 lm, respectively, which were similar to the dimensions of S. longispina (diameter = 5 ± 0.01 lm; length = 40 ± 11 lm). Since the succession did not noticeably affect the abundance of D. magna and the removal efficiency of nutrients during and after the succession, these results indicated that both Scenedesmus and the cyanobacterium species were well grazed by D. magna. The material budget of the continuous-flow system was calculated and presented in Fig. 7. For 45 days, the total amounts of TN and TP flowing into the system were 471.3 and 29.3 g, respectively, and the amounts flowing out were 144.6 g (31%) and 8.0 g

Hydrobiologia (2009) 630:149–159

Target water TN: 471.3 g (100%) TP: 29.3 g (100%)

155 Phytoplankton tank Standing crop TN: 4.2 g (0.9%) TP: 0.4 g (1.4%)

TN: 216.8 g TP: 13.1 g

Ammonia stripping,

sedimentation, algae uptake TN : 250.3 g (53.1%) TP : 15.8 g (53.9%)

Zooplankton chamber Standing crop TN: 8.6 g (1.8%) TP: 1.4 g (4.8%)

Effluent TN: 144.6 g (30.7%) TP: 8.0 g (27.3%)

Elimination of D. magna and sludge TN : 63.6 g (13.5%) TP : 3.7 g (12.6%)

Fig. 7 Material budget of nitrogen and phosphorus in the continuous-flow AAFW system. Numbers in parentheses = proportion (%) to the amounts of nutrients in target stream water

(27%), respectively. Biologically, 250.3 g (53.1%) of TN and 15.8 g (53.9%) of TP were removed by phytoplankton and 63.6 g (13.5%) of TN and 3.7 g (12.6%) of TP by zooplankton.

Discussion Efficiency of the food web in the removal of nutrient

water. The aquatic food chain system employed in this study could treat the low concentration pollution, with the removal efficiency of 69% and 73% for TN and TP, respectively. In the previous study by Kim et al. (2003), concentrations of TN and TP (26.9 and 2.6 mg l-1) were higher than those for this study (10.6 and 0.65 mg l-1). Therefore, aquatic food web system can be applied to both severely or slightly polluted water. Carrying capacity for Daphnia magna

There were several studies that used artificial food webs for the removal of nutrients from sewage or treated sewage effluent. Tam & Wong (1989) removed nutrients from sewage by a batch system using Chlorella pyrenoidosa and Scenedesmus, and about 80% of both TN and TP were removed. Kawasaki et al. (1982) used artificial food chains comprising Scenedesmus, Daphnia, and silver carp on laboratory scale, and the removal rates of TN and TP from domestic sewage were 30–78% and 55–98%, respectively. Kim et al. (2003) removed 68% of TN and 56% of TP from sewage by artificial food web system with phytoplankton and D. magna. These studies focused on sewage which can also be treated by advanced treatments such as A2O process. Using A2O treatment, the ranges of the removal rate of nutrients in sewage were from 67 to 77% for TN and from 47 to 94% for TP (Xiaolian et al., 2006). Comparing AAFW methods with advanced treatments, one finds little difference in efficiency. However, advanced treatment systems are complicated and expensive to operate, therefore, not suitable for large quantity of slightly polluted water such as stream

In the zooplankton chamber, D. magna ingested phytoplankton effectively, even though both were grown with external resources. D. magna is an easyhandled, conveniently managed, and effective grazer of phytoplankton (Kim et al., 2003). The role of D. magna in the AAFW system was to collect the scattered phytoplankton and making fecal pellets which contain high density of phytoplankton remnants. One individual D. magna was ingesting 0.056lg chlorophyll a per day (see Fig. 6) and making fecal pellets which sank to the bottom of the chamber as sludge. On draining this sludge, which was consisted of fecal pellets and debris of D. magna, the process of removing the nutrients from the target stream was completed. D. magna contained 0.2% TN and 12.2% TP of dry weight (Kim et al., 2003). Therefore, the standing crop in the zooplankton chamber was calculated to be 8.6 g for TN and 1.4 g for TP (Fig. 7). The maximum number of D. magna observed in this study was 1,245 individual l-1, when 83 ± 6 mg m-3 of chlorophyll a was supplied. In a previous study with

123

156

high TN and TP in input water (Kim et al., 2003), the total number of D. magna reached to 1,460 individual l-1, while the chlorophyll a concentration of supplied water was 874 mg m-3, which was ten times higher than that of this study. Therefore, the maximum density of D. magna was similar in the two studies, with 1,200–1,400 individual l-1 range, regardless of prey abundance, i.e., 10-fold difference in chlorophyll a concentration. Abundance of D. magna may be limited by shortage of food, space, or oxygen, or by accumulation of growth-inhibitory wastes (Zeiss, 1963; Mourelatos & Lacroix, 1990; Persson et al., 2007). Regarding the question on the major factor determining the carrying capacity of the zooplankton chamber for D. magna, shortage of food can be eliminated from the list of the potential limiting factors because the 10-fold difference in chlorophyll a concentrations does not seem to cause significant difference in the density of D. magna in the zooplankton chamber. We believe that the major factor for regulation of D. magna abundance might be shortage of space, based on the fact that regular harvesting of zooplankton biomass supports continuous growth of D. magna population and that about 10% of chlorophyll a was still detected in the effluent. Comparison with constructed wetland technology The removal rates of TN and TP from stream water were 69 and 73%, respectively, in this study. These removal efficiencies were similar to those of previous studies using artificial food web systems (Kawasaki et al., 1982; Kim et al., 2003) or other industrialized techniques (Xiaolian et al., 2006) even though the nutrient concentrations in the target stream water were significantly lower than those in the sewage water. Constructed wetland technology has been often used for the purpose of reducing nutrient loading from non-point sources. The removal rates of TN and TP in constructed wetlands for the same stream water were 33% and 25% (Hur et al., 2007), respectively, only in warm period. Therefore, the removal rate of the AAFW system was better than that of the constructed wetland. In addition, this AAFW system has another advantage over constructed wetland method. It is more efficient than constructed wetland in the use of space by having a compact organization of facilities.

123

Hydrobiologia (2009) 630:149–159

Nutrient removal process In the AAFW system, most nutrient removal occurred in the phytoplankton tank (77% and 76% of totally removed TN and TP, respectively). In general, TN can be removed by ammonia stripping, denitrification, and precipitation or algal uptake (Kawasaki et al., 1982). Considering that the stream water contained dissolved oxygen ([3.9 mg l-1) and that water in the phytoplankton tank was circulated, anaerobic processes such as denitrification might not occur. Therefore, most of the TN appears to be removed by sedimentation and algae uptake. Unknown amount of ammonia stripping is also likely to have occurred via a spontaneous process since the water in the tank was highly alkaline and circulated while open to the atmosphere. In this case, the source of ammonia could be deamination of organic nitrogen compounds by decomposers such as heterotrophic bacteria. The high removal rate of TP can be explained as algal uptake. In the target stream water, most phosphorus existed as DIP, and its level diminished while the water was passing the phytoplankton tank. Since DIP is an essential compound for most aquatic organisms including bacteria, algae, and zooplankton, they tend to uptake surplus amount of DIP, via the process called luxury uptake (Powell et al., 2008). Therefore, it is likely that massive growth of phytoplankton has utilized most of the DIP diminishing in the phytoplankton tank. Also, the formation of Ca–Mg-PO4 in the alkaline condition (Kawasaki et al., 1982) can account for a part of the reduction of DIP in the phytoplankton tank. Succession of phytoplankton A significant change in the composition of the phytoplankton community was observed during the operation of the AAFW system. The most predominant species was Scenedesmus longispina on the first day, but it changed to oscillatorean cyanobacterium sp. by the 30th day. However, the grazing rate of D. magna appeared unaffected by the shift of phytoplankton species (Fig. 6). This observation implies that the cyanobacterium species growing in the phytoplankton tank was well grazed by D. magna. Several studies have shown inhibition in the feeding activity of zooplankton by filamentous cyanobacteria (Emerson & Peters, 1978; Dawidowicz

Hydrobiologia (2009) 630:149–159

et al., 1988; Hawkins & Lampert, 1989). The inhibitory effect could be due to two aspects of the phytoplankton: morphology and toxicity. Long filaments of Cyanobacteria, often forming dense mats floating on the surface of eutrophic waters, are not easily harvestable to most cladoceran zooplankton (Work & Havens, 2003). Size of the cyanobacterial cells observed in this study was similar to that of the Scenedesmus species. Some species of oscillatorean Cyanobacteria show cytotoxicity to zooplankton and herbivorous fishes. However, cyanobacterium cells observed in the AAFW of this study do not appear to be toxic to D. magna since the abundance of D. magna on or after the 30th day did not decrease (Fig. 6). Phytoplankton succession in the phytoplankton tank should be managed to avoid predominance by grazing-resistant phytoplankton species. In a phytoplankton community under top-down control by zooplankton and/or fish, populations with grazingresistant traits, such as toxicity and bulky morphology, are selected to predominate in the community. In the AAFW system, grazing of algae by zooplankton was spatially separated from algal growth. This design relieves pressure of selection for resistance to grazing, by minimizing the top-down control in the phytoplankton tank. Grazing pressure in the phytoplankton tank can be further relieved by employing mechanisms removing grazers from the inflow and/or tank water. Filtering out particles larger than or equal to D. magna can be one example. Size-based fractionation and elimination of particles in the phytoplankton tank also have the effect of filtering out large filamentous or matforming phytoplankton and reduce their chance to become predominant. Installing a filter system in connection with the circulation pump and the sludge discard system can be cost effective. Role of heterotrophic organisms in the nutrient removal In the AAFW system, the role of phytoplankton is crucial in transforming the dissolved form of nutrients into the particulate forms that can be easily removed from water by sedimentation or harvesting. Considering that kinetics of D. magna grazing also depends on phytoplankton density, biosynthesis by phytoplankton is the strongest determinant of nutrient removal efficiency.

157

On the other hand, activities of heterotrophic organisms may affect the nutrient removal rate indirectly, by modifying the phytoplankton growth. Heterotrophic bacteria are typical decomposers mineralizing organic materials into dissolved nutrients. Therefore, bacteria in the phytoplankton tank may contribute to the overall efficiency of nutrient removal by converting organic forms of nutrients into dissolved inorganic forms that can be readily used by the phytoplankton for biosynthesis. Daphnia magna ingests algal biomass in the zooplankton chamber, and a part of it is mineralized via respiration and ammonia excretion. Nutrients mineralized in the dark zooplankton chamber are not utilized by phytoplankton. Therefore, D. magna respiration reduces the nutrient removal efficiency of the system. However, sedimentation of fecal pellets and chitin-rich exuvia produced during molting of D. magna contributes to nutrient removal. Variability in growth kinetics of planktons under fluctuating conditions Understanding the characteristics in growth kinetics of planktons and their mode of response to fluctuating environmental condition is crucial for practical application of the system in the field. In order to cope with variable conditions of streams which fluctuate daily and seasonally, growth kinetics of phytoplankton and D. magna needs to be robust to those fluctuations or, at least, components of the AAFW system need to be adaptable to dampen the fluctuations. According to the results in this study, growths of the phytoplankton and D. magna follow logistic growth models (Figs. 2, 3). A logistic growth could be described with three parameters: initial abundance (N0), intrinsic growth rate (r), and carrying capacity of system (K). Therefore, stability of the AAFW system can be interpreted in terms of these three parameters for phytoplankton or D. magna. The efficiency of nutrient removal in the phytoplankton tank directly depends on the rate of algal growth during a given HRT, and it is a function of N0, r, and K. Algal biomass in the inflow water may positively contribute to N0, but its effect will be negligible because N0 in the tank will be higher than algal abundance in inflow at most times (Fig. 6). The r value of phytoplankton community will mostly reflect the intrinsic growth rate of the most predominant

123

158

species, and shifts in r value can be caused by the succession of phytoplankton species. Since this succession is an outcome of the competition between phytoplankton species co-occurring in similar ecological niches, a dramatic change in r is unlikely to occur. Actually, the succession from Scenedesmus to the cyanobacterium sp. observed in this study does not appear to have affected the kinetics of algal growth. On the contrary, K can be influenced strongly by inflow. When N0 is large and approaches K, growth of the phytoplankton is likely to be limited by the resource that determines K. If the limiting resource is a nutrient, amount of the nutrient in inflow can make K fluctuate. Therefore, an AAFW system is not robust if K for phytoplankton is set by a limiting nutrient, the level of which fluctuates in inflow water. Also, K can be set by a factor not related to inflow. For example, phytoplankton growth is limited by the availability of sunlight. In this case, phytoplankton growth fluctuates independently from the inflow but depends on climate factors. These fluctuations, however, can be prevented to some extent by careful modulation of HRT. Setting HRT to a value that can result in phytoplankton abundance below the carrying capacity can achieve stability of the AAFW system from stochastic variation of inflow and climate factors. In the zooplankton chamber, growth kinetics of a single species, D. magna, is the key process affecting the nutrient removal rate. In the system, range of abundance of D. magna (N0) is set and modulated by the operator of the system while r and K of D. magna growth may depend on the inflow from the phytoplankton tank. K can be determined by the availability of space, which does not vary, or abundance of phytoplankton prey, which may fluctuate with time. According to Lotka–Volterra predator–prey model, the growth rate of predator depends on the product of prey and predator densities. Since the density of D. magna is controlled to remain within a range set by the system operator, r of D. magna is likely to depend more on the abundance of phytoplankton flowing into the zooplankton chamber. Therefore, when the prey density is low, both r and K of D. magna may fluctuate along with the conditions in the phytoplankton tank. When K is determined by a factor other than prey density, Lotka–Volterra model depicts that r is determined mostly by the abundance of D. magna. In conclusion, the growth kinetics of D. magna is influenced by fluctuations in inflow

123

Hydrobiologia (2009) 630:149–159

water from the phytoplankton tank when the inflow yields low density of phytoplankton. Therefore, proper operation of the phytoplankton tank can support stability of D. magna growth and nutrient removal efficiency. Dynamic modelling for the automation of monitoring and operation Automated monitoring and manipulation of parameters in each compartment of the AAFW system, based on dynamic models for algal production and D. magna grazing, can be an ideal solution to establish robustness of the system. Growth rate and carrying capacity for phytoplankton or D. magna population can be modeled as logistic growth model or Lotka–Volterra model, respectively. Real-time estimation of model parameters can be achieved by automated monitoring of population densities of plankton. Algal biomass can be measured in situ by fluorometers which detect auto-fluorescence of chlorophylls. Image analysis techniques for non-destructive determination of demography and biomass of Daphnia populations are also available (Færøvig et al., 2002). Automated methods for harvesting of sludge in the phytoplankton tank and D. magna biomass in the zooplankton chamber may complete automation of the entire processes of the AAFW system. Deployment of the food web system Kim et al. (2003) compiled lists of advantages and disadvantages of the artificial food web system. Being a solar-energy-based system, the most significant advantage is its low cost and convenience for construction and operation. In addition, use of sediment sludge as fertilizer or D. magna biomass as fish feed is another advantage. Disadvantages are the need for a greenhouse for operation during cold seasons, risk of collapse of the system due to occasional contamination by toxic chemicals, and the need for space to support both phytoplankton and zooplankton. We believe careful maintenance of the system is required because unattended long-term operation of the system may have increased risk of system collapse, such as development of non-grazeable large algal species.

Hydrobiologia (2009) 630:149–159

Hereby, we also suggest flexibility of the system as another advantage. Phytoplankton tank and zooplankton chamber can be separately and independently operated depending on field condition. For example, only zooplankton chamber can be operated for treatment of waters with blooms of algal community grazeable by D. magna. If nutrient concentrations in effluent from phytoplankton tanks are in compliance with the water quality standard, one may choose to operate phytoplankton tanks only. In conclusion, this aquatic food web system can be applied not only to waste water but also to treatment of slightly polluted stream waters, consequently preventing first pollution and after eutrophication of rivers and lakes from largevolume non-point source pollutions. Acknowledgments This research was supported by the Ecotechnopia 21 Project, Ministry of Environment, Korea, and by the 2nd phase of Brain Korea 21 Project in 2008–2009.

References APHA, 2001. Standard Methods for the Examination of Water and Wastewater, 20th ed. American Public Health Association, Washington, DC. Azza, N., P. Denny, J. V. D. Koppel & F. Kansiime, 2006. Floating mats: their occurrence and influence on shoreline distribution of emergent vegetation. Freshwater Biology 51: 1286–1297. Braskerud, B. C., 2003. Clay particle retention in small constructed wetlands. Water Research 37: 3793–3802. Coveney, M. F., D. L. Stites, E. F. Lowe, L. E. Battoe & R. Conrow, 2002. Nutrient removal from eutrophic lake water by wetland filtration. Ecological Engineering 19: 141–159. Dawidowicz, P., Z. M. Gliwicz & R. D. Gulati, 1988. Can Daphnia prevent a blue green algal bloom in hypertrophic lakes? A laboratory test. Limnologica (Berlin) 19: 21–26. Emerson, K. W. & R. H. Peters, 1978. Some size-dependent inhibitions of larger cladoceran filterers in filamentous suspensions. Limnology and Oceanography 23: 1238–1245. Environmental Protection Agency, 1987. Design Manual Phosphorus Removal. EPA/625/1-87/001: 6–49. Færøvig, P. J., T. Andersen & D. O. Hessen, 2002. Image analysis of Daphnia populations: non-destructive determination of demography and biomass in cultures. Freshwater Biology 47: 1956–1962. Greenway, M. & J. S. Simpson, 1996. Artificial wetlands for wastewater treatment, water reuse and wildlife in Queensland, Australia. Water Science and Technology 33: 221–229. Hawkins, P. & W. Lampert, 1989. The effect of Daphnia body size on filtering rate inhibition in the presence of a filamentous cyanobacterium. Limnology and Oceanography 34: 1084–1089. Hur, J. G., J. H. Nam, Y. J. Kim, I. S. Kim, K. S. Choi, S. I. Choi & T. S. Ahn, 2007. Analysis of efficiency of artificial

159 wetland for waste water treatment past six year operation. Journal of the Korean Society for Environmental Restoration and Revegetation Technology 10: 1–7. Jiang, C., X. Fan, G. Cui & Y. Zhang, 2007. Removal of agricultural non-point source pollutants by ditch wetlands: implications for lake eutrophication control. Hydrobiologia 581: 319–327. Kawasaki, L. Y., E. Tarifeno-Silva, D. P. Yu, M. S. Gordon & D. J. Chapman, 1982. Aquacultural approaches to recycling of dissolved nutrients in secondarily treated domestic wastewaters—I Nutrient uptake and release by artificial food chains. Water Research 16: 37–49. Kim, S.-R., S.-S. Woo, E.-H. Cheong & T.-S. Ahn, 2003. Nutrient removal from sewage by an artificial food web system composed of phytoplankton and Daphnia magna. Ecological Engineering 21: 249–258. Leeben, A., I. To˜nno, R. Freiberg, V. Lepane, N. Bonningues, N. Makaro˜tsˇeva, A. Heinsalu & T. Alliksaar, 2008. History of anthropogenically mediated eutrophication of Lake Peipsi as revealed by the stratigraphy of fossil pigments and molecular size fractions of pore-water dissolved organic matter. Hydrobiologia 599: 49–58. Mizuno, T., 1975. Illustrations of the Freshwater Plankton of Japan. Hoikusha Publishing, Osaka. Mourelatos, S. & G. Lacroix, 1990. In situ filtering rates of cladocera: effect of body length, temperature, and food concentration. Limnology and Oceanography 35: 1101–1111. Olin, M., M. Rask, J. Ruuhija¨rvi, J. Keskitalo, J. Horppila, P. Tallberg, T. Taponen, A. Lehtovaara & I. Sammalkorpi, 2006. Effects of biomanipulation on fish and plankton communities in ten eutrophic lakes of southern Finland. Hydrobiologia 553: 67–88. Persson, J., M. T. Brett, T. Vrede & J. L. Ravet, 2007. Food quantity and quality regulation of trophic transfer between primary producers and a keystone grazer Daphnia in pelagic freshwater food webs. Oikos 116: 1152–1163. Powell, N., A. N. Shilton, S. Pratt & Y. Chisti, 2008. Factors influencing luxury uptake of phosphorus by microalgae in waste stabilization ponds. Environmental Science & Technology 42: 5958–5962. Shutes, R. B. E., 2001. Artificial wetlands and water quality improvement. Environment International 26: 441–447. Somodi, I. & Z. Botta-Duka´t, 2004. Determinants of floating island vegetation and succession in a recently flooded shallow lake, Kis-Balaton (Hungary). Aquatic Botany 79: 357–366. Tam, N. F. Y. & Y. S. Wong, 1989. Wastewater nutrient removal by Chlorella pyrenoidosa and Scenedesmus sp. Environmental Pollution 58: 19–34. Work, K. A. & K. E. Havens, 2003. Zooplankton grazing on bacteria and cyanobacteria in a eutrophic lake. Journal of Plankton Research 25: 1301–1306. Xiaolian, W., P. Yongzhen, W. Shuying, F. Jie & C. Xuemei, 2006. Influence of wastewater composition on nitrogen and phosphorus removal and process control in A2O process. Bioprocess and Biosystems Engineering 28: 397–404. Zeiss, F. R. Jr., 1963. Effects of population densities on zooplankton respiration rates. Limnology and Oceanography 8: 110–115.

123

Suggest Documents