ARTICLE IN PRESS

Journal of Environmental Management 81 (2006) 420428

www.elsevier.com/locate/jenvman

Application of Eichhornia crassipes and Pistia stratiotes for treatment of

urban sewage in Israel
Y. Zimmels, F. Kirzhner, A. Malkovskaja
Environmental, Water Resources and Agricultural Engineering, Department of Civil and Environmental Engineering, Technion, Haifa 3200, Israel
Received 18 April 2005; received in revised form 6 September 2005; accepted 18 November 2005
Available online 4 April 2006

Abstract

The effectiveness of sewage purication by aquatic plants, such as water hyacinth (Eichhornia crassipes) and water lettuce (Pistia
stratiotes), was tested on laboratory and pilot scales. Cascade and semi-continuous pilot experiments veried that the plants are capable
of decreasing all tested indicators of water quality to levels that permit the use of the puried water for irrigation of tree crops. This
applies to biochemical oxygen demand (BOD), chemical oxygen demand (COD), total suspended solids (TSS), and turbidity. The
laboratory-scale tests conrm the capacity of the plants to reach and hold reasonably low levels of BOD (57 mg L 1) and COD
(4050 mg L 1) and very low levels of TSS (35 mg L 1) and turbidity (12 NTU).
In the experimental pilot setup, with circulation, COD decreased from 460 to 100 mg L 1 after 2.54 days of treatment, while 67 days
were required to this end without circulation. This doubled the active pond area and provided a two-level hydraulic loading (8 and
12 L min 1) with circulation that proved to be effective during the summer as well as the winter season. The outow concentrations were
5085 mg L 1 of COD and 46 mg L 1 of BOD. The results show that the use of this free water surface ow system (FWS) and its low
maintenance system for treatment of urban and agricultural sewage is a viable option.
r 2006 Elsevier Ltd. All rights reserved.

Keywords: Aquatic plants; Eichhornia crassipes; Pistia stratiotes; Sewage; Pollution; Natural treatment; Phytoremediation; Free water surface ow system
(FWS)

1. Introduction good conditions for physical ltration and a large surface

Wetlands, which are being constructed worldwide, are
designed and operated for wastewater treatment at the
secondary and tertiary level (Campbell and Ogden, 1999;
Gopal, 1999; Kadlec and Knight, 1995; Kadlec, 1995).
Several physical, chemical and biological processes are
involved in the transformation and consumption of organic
matter and plant nutrients within the wetland (Cooke,
1994; Gale et al. 1993). The most important functions of
the macrophytes in the treatment of wastewater relate to
physical effects which they induce therein (Brix, 1997). For
example, wetlands involve settling of suspended particulate
matter, which is the prime cause for reduction of BOD
levels in the treated wastewaters. The macrophytes provide
Corresponding author. Tel.: +972 4 8293774; fax: 972 4 8228898.
E-mail address: kfelix@tx.technion.ac.il (F. Kirzhner).
area for attached microbial growth and activity (Brix,
1997).
Over the years, the use of articially constructed
wetlands for wastewater treatment has been increasing
considerably (Bhamidimarri et al., 1991; Chescheir et al.,
1991, Hammer, 1990; Moshiri, 1993; Reuter et al., 1992;
Schwartz and Boyd, 1995; Tripathi and Shukla, 1991). The
general practice provides evidence that wetlands remove
contaminating nutrients and solids from the wastewater.
The problem is to maximize their efciency at the lowest
possible cost.
Aquatic plants, such as water hyacinth (Eichhornia
crassipes), pennywort (Hydrocotyle umbellate), and duck-
weeds (Lemna spec.) (US EPA, 1988) are known to be
effective in single pond wastewater treatment. Circulation
of the treated stream enhances the treatment capacity,
facilitates operation at higher organic loading, and helps to

0301-4797/$ - see front matter r 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jenvman.2005.11.014
 ARTICLE IN PRESS
Y. Zimmels et al. / Journal of Environmental Management 81 (2006) 420428
reduce the required pond area (US EPA, 1988, 1993).
Furthermore, circulation can increase the dissolved oxygen
concentration near the inlet. This, in turn, can help
suppress inlet odors, lower the BOD, and enhance
nitrication potential in open-water zones (US EPA, 2000).
The major characteristics of Eichhornia crassipes, that
make them an attractive biological support media for
bacteria, are their extensive root system and rapid growth
rate. The major advantage of duckweeds is their lower
sensitivity to cold climates, while their shallow root systems
and sensitivity to wind (US EPA, 1988) are considered to
be drawbacks.
Wetlands are affected by seasonal growth of wild plants
and rising temperatures (Campbell and Ogden, 1999). The
rates of biological decomposition slow down as the
temperature drops (US EPA, 2000).
For this research, Eichhornia crassipes and Pistia
stratiotes were selected in view of their high growth rates.
These plants grow rapidly under favorable conditions in a
nutrient-rich environment such as domestic sewage. Effec-
tive water management of FWS aquatic plant systems
requires understanding of the design and operational
features of cascade, semi-continuous and batch treatment
modes. The scarcity of data concerning the effect of
circulation and ow rate on articial closed FWS systems
calls for the conduction and analysis of related tests.
The aim of this research was to evaluate the effect of
sewage dosage, in laboratory pulsed cascade experiments,
and wastewater circulation on the process kinetics and
quality of the treated water, under eld conditions of the
pilot unit. The results provide estimates of the efcacy of
wastewater circulation in cascade and semi-continuous
experiments carried out in our two-story articial wetland
systems, which were operated under the conditions
prevailing in the greenhouse (in winter). The research
hypothesis, that ow rate and circulation of wastewater
can enhance the BOD and COD removal kinetics, is
established experimentally. To this end a monitoring
system for the operating variables and water quality
indicators was set up.
2. Methods
Two types of experiments were performed: laboratory
cascade and in the eld. The aquatic plants Eichhornia
crassipes and Pistia stratiotes were obtained from natural
specimens grown in fresh water ponds in Israel. Sample
collections and handling procedures were performed
according to the proceedings recommended by Standard
Methods (APHA, AWWA and WPFC, 1998).
The purication process was monitored by standard
measurements of BOD, COD, TSS and turbidity levels, in
the control and treated sewage. Each result is given as an
average of two or three measurements taken from the same
sample.
421
2.1. Laboratory pulsed cascade experiments (set nos. 1
and 2)
The rst and second sets of laboratory experiments were
both performed with ve identical containers
(0.39  0.56 m2 and 0.45  0.7 m2 oor area, respectively).
These containers were operated at either 40 (rst set) or
30 L (second set) levels. Fresh water (for 7 days in the rst
set and 3 days in the second set) was used in all containers
to support the oating plants with clean roots. Subse-
quently, 50% of the water in each container was replaced
(over a period of 1 h) by sewage. The system was built as a
cascade, and operated in a pulsed mode of sewage dosage.
The mixed wastewater was circulated at a xed ow rate of
1 L min 1. This gave a residence time of 40 and 30 min in
the two sets of containers, respectively. The sewage was
added only to the fth container and water mixtures were
withdrawn from the other three containers (nos. 4, 3 and 2,
Fig. 1). The rst container, which served as a control, was
lled with a mixture comprising 50% fresh water and 50%
raw sewage. The tests were performed in each container
with 0.4 L min 1 aeration (applied for 1516 h/day), and
30 min/h of articial light at 46004700 lux, produced by
special lamps for plant growing. Each result is given as an
average of two measurements taken from the second and
fth containers. Cascade tests were performed during a
period of 11 and 15 days.
Set no. 1: In four containers, nos. 2, 3, 4 and 5, which
were set as a cascade, fresh water was used to support the
oating Eichhornia crassipes plants (12 pieces in each) and
one container (no. 1, without plants) was used as the
control. On day 1, 80 L of water were withdrawn (20 L per
container) and 80 L of sewage gradually added, through
container no. 5, so as to set the mixture volume back to the
operating level of 40 L in all containers. This procedure was
repeated on day 4.
Set no. 2: In this set of experiments, the working volume
per container was set at 30 L. In four (cascade) containers,
fresh water was used to support the oating Pistia
stratiotes plants (15 pieces in each) and one container
(without plants) was used as the control. On day one 60 L
of water were withdrawn (15 L per cascade container) and
60 L of sewage added through container no. 5, so as to set
the mixture volume back to the operating level of 30 L in
all containers. This procedure was repeated on days 3, 5, 7,
9 and 11.
2.2. Laboratory experiments with circulation
Set no. 3: Six batch experiments were performed in
identical containers (0.39  0.56 m2 oor area), in order to
simulate the level of circulation. The working volume of
the containers was set at 46 L. The tests were performed
using 30 min/h of articial light at 46004700 lux. In ve
containers, sewage was used to support the oating Pistia
stratiotes plants (15 pieces in each), and one container
 ARTICLE IN PRESS
422 Y. Zimmels et al. / Journal of Environmental Management 81 (2006) 420428
Fig. 1. Schematic layout of cascade laboratory setup with circulation.
(sixth) (without plants) was used as the control. In these
experiments, the circulation conditions were as follows:
Test number Circulation, ow rate (L h 1)
1 10
2 20
3 30
4 40
5 0
6control 0
2.3. Outdoors experimental setup
The experimental setup, which was constructed on the
outskirts of the Civil and Environmental Engineering
Department, was used to test the performance of the
oating plants Eichhornia crassipes and Pistia stratiotes
under different feed conditions of sewage and fresh water.
To this end, use was made of a pilot unit built as a
greenhouse. The pilot unit comprised two 2  5 m2 pools
connected by pipes and set one on top of the other. (No
analog was found in the literature). The volume of water
(0.4 m deep) in the lower pool was xed at 4.0 m3. The
volume of water (0.1 m deep) in the upper pool was xed at
1.0 m3. The water temperature inside the greenhouse (in
winter) was 1516 1C, an increase of 45 1C compared to
the outside air. The tests were performed with aeration of
the pools. Two sets of experiments extended for 29 and 40
days were performed as follows:
Set no. 4: This set of experiments was performed in order
to test the effect of daily dosage of sewage on the quality of
the treated water under eld conditions in a pilot unit (in
summer). Initially, fresh water was used to support the
oating plants and then, following the dosage of sewage
and mixing, the excess water was withdrawn so as to set the
mixture volume back at the operating level of 4.0 m3 in the
lower pool, and 1.0 m3 in the upper pool. Sewage was
added daily during a 36-day period in different cycles. The
daily dosage varied between cycles. The mixed wastewater
was circulated at a xed ow rate of 6 L min 1.
Set no. 5: This set of experiments was performed in order
to test the effect of sewage circulation on the quality of the
treated water under eld conditions of the pilot unit (in
winter). The circulation load consisted of most of the total
stream and was directed from the efuent port back to the
feed port (Fig. 2). Operating conditions were set identical
to set no. 3. Sewage was added in four daily steps of
0.25 m3 day 1, and then it was stopped for two days. This
cycle was repeated for 29 days. The mixed wastewater was
circulated at a ow rate of 12 and 8 L min 1 for 29 days.
The sewage characteristics were as follows:
BOD (mg L 1) 200250
COD (mg L 1) 270550
Turbidity (NTU) 130250
TSS (mg L 1) 250350
3. Results
3.1. Set no. 1. Cascade
Fig. 3 depicts results of (a) BOD, (b) COD, (c) TSS, and
(d) turbidity vs. time. In Fig. 3(ad) the BOD, COD, TSS
and the turbidity are given as the percent of their
corresponding value as specied in the preceeding dosage
of sewage. Thus, reference levels of 100% were set for day 1
and 4, following the respective sewage dosage.
Fig. 3a shows that the plants were capable of decreasing
the BOD level, after the rst and second dosage of sewage,
to nearly the same levels after 4 days. The rate of BOD
 ARTICLE IN PRESS
Y. Zimmels et al. / Journal of Environmental Management 81 (2006) 420428 423

Aeration

Mixer Effluent
Feed

Circulating load
Primary
precipitation

Fig. 2. Schematic layout of outdoor facility with circulation.

System with aquatic plants Control tank Sewage addition

100 60

Total sewage dosage, L

System with aquatic plants Control tank Sewage addition
100 60

Total sewage dosage, L

80
80 40

TSS, %
60
40
BOD, %

60
40
40 20
20
20 20

0 0 0 0
1 2 3 4 4 5 6 7 8 9 10 11 1 2 3 4 4 5 6 7 8 9 10 11
(a) Time, day (c) Time, day

System with aquatic plants Control tank Sewage addition

100 60

Total sewage dosage, L

System with aquatic plants Control tank Sewage addition
Total sewage dosage, L %

100 60 80
Turbidity, %

80 40
60
40
COD, %

60
40
40 20
20
20
20

0 0 0 0
1 2 3 4 4 5 6 7 8 9 10 11 1 2 3 4 4 5 6 7 8 9 10 11
(b) Time, day (d) Time, day

Fig. 3. Results of cascade test in the presence of Eichhornia crassipes: (a) BOD (%) and total sewage dosage vs. time; (b) COD (%) and total sewage
dosage vs. time; (c) TSS (%) and total sewage dosage vs. time; (d) turbidity (%) and total sewage dosage vs. time.

removal with time (at the higher initial BOD levels on day increased from 1 (on day 1) to approximately 3 (on day 4)
160 mg L 1), was about 96%, as compared to that on day 4 and from 1 (on day 4) to approximately 2.5 (on days 7 and
(2.4 mg L 1). The treatment with plants was more efcient, as 8), see Fig. 3b. On day 4, the COD level dropped from its
compared to the control. The increase in efciency becomes initial 237 mg L 1 level to 65 mg L 1 (72.6% removal) in
more evident with an increase in time. This can be seen by the control, and from 242 mg L 1 down to 23.6 mg L 1
comparing the percent residual BOD after the rst and (90.25% removal) in the presence of plants. However, on
second dosage, to that of the control (Fig. 3a). For example, days 711, no decrease below 60.6 mg L 1 (52.3% removal,
the ratio obtained by the percent BOD in the control and that from the initial 127 mg L 1 level) in the control and
in the presence of Eichhornia crassipes plants, increased from 48 mg L 1 (79.04% removal, from the new level of
1 (on day 1) to approximately 4 (on day 4). Thus on day 4, 229 mg L 1) in the presence of plants was observed. This
there was a 4-fold decrease in the percent of residual BOD indicated that the second cycle was less efcient than the
due to removal by the plants, as compared to the control. rst one. Similar behavior was also observed for the TSS
Similar results were obtained after the second dosage on day (Fig. 3c). Here the corresponding ratios increased from 1
4, where the ratio increased again from 1 (on day 4) to nearly (on days 1 and 4) to a peak of 9 (on day 4) and
4 (on day 9), see Fig. 3a. approximately 3 on days 811. The capacity of the plants
Similar behavior was observed with the other water- to decrease the TSS level seemed to be unaffected by the
quality indicators. The ratio obtained by the percent COD second dosage of sewage on day 4. This observation applies
in the control, and in the presence of Eichhornia plants, also to the turbidity (Fig. 3d).
 ARTICLE IN PRESS
424 Y. Zimmels et al. / Journal of Environmental Management 81 (2006) 420428

3.2. Set no. 2. Cascade absence of plants (control container), the turbidity level
reached 8 NTU.
The results of 6 cycles of sewage addition, in 15 L equal Ammonia assimilation occurs during biological oxida-
doses, each followed by a monotonic decrease in the water- tion of organic components in the presence of plants. The
quality indicators, BOD, COD, TSS and turbidity, are decrease in ammonia due to the presence of Pistia (from
shown in Fig. 4ad, respectively. In this set, the water plant 56.1 down to 0.46 mg L 1) produced an increase in the
was Pistia. The 15 L dosages of sewage were administered nitrite level from 0 up to 11 mg L 1 (graphic data not
on days 1, 3, 5, 7, 9 and 11. shown).
Each dosage produced an increase in the levels of all
measured water quality indicators (a) BOD, (b) COD, (c) 3.3. Set no. 3. Circulation
TSS and (d) turbidity. In all cycles the removal of
pollutants with plants was greater compared to the control. Fig. 5 depicts results of (a) BOD and (b) COD vs. time.
This is reected by the lower values of BOD, COD TSS Application of circulation (in the presence of plants)
and turbidity as compared to the control. After 14 days the increased the organic loading rate. The effect of circulation
BOD level, in the presence of Pistia, decreased from 50.8 on the BOD removal (from its 61 mg L 1 initial value) in
down to 5.3 mg L 1, as compared to 18.6 mg L 1 (Fig. 4a) the rst 4 days depended on the ow rate. At 10 L h 1 it
in their absence. The COD level showed a trend of decreased to 7.4 mg L 1, whereas at 40 L h 1 it dropped
monotonic decrease from cycle to cycle (Fig. 4b). At the further to 5.3 mg L 1. In the absence of circulation the
end of the rst cycle (day 3), the Pistia plants decreased the results were 10.2 and 20.1 mg L 1 with plants and control,
COD to 50 mg L 1 and at the end of the fth and sixth respectively. With circulation, a level of 7.613.3 mg L 1
cycles it was lowered below 40 mg L 1. In this respect the was reached after 4 days of treatment, while 6 days were
performance of the Pistia plants was not affected by required to this end without circulation.
the sewage dosages and it showed consistency, and After 8 days (in the presence of plants and circulation at
even improvement, as the cycles advanced. The response 30 L h 1), the BOD reached 3 mg L 1, as compared to
of the TSS and turbidity to the dosage cycles was 4.8 mg L 1 in the presence of plants, and 8.8 mg L 1 in the
similar (except for a uctuation on day 5). After 14 days control, both without circulation. The COD removal
(Fig. 4d), the initial level of 56 NTU (Nephelometric followed a similar pattern (Fig. 5b). After 6 days in the
Turbidity Units) was brought down to 1.9 NTU. In the presence of plants, the COD removal (from its

System with aquatic plants Control tank Sewage addition

350 105

System with aquatic plants Control tank Sewage addition 300 90

80 105

Total sewage addition, L

250 75
70 90
Total sewage addition, L

COD, mgL-1

60 200 60
75
BOD, mgL-1

50
60 150 45
40
45
30 100 30
30
20
50 15
10 15

0 0 0 0
1 2 3 3 4 5 5 6 7 7 8 9 9 10 11 11 12 13 14 1 2 3 3 4 5 5 6 7 7 8 9 9 10 11 11 12 13 14
(a) Time, day (b) Time, day

System with aquatic plants Control tank Sewage addition

System with aquatic plants Control tank Sewage addition
80 105
70 105
70 90
Total sewage addition, L

60 90
Total sewage dosage, L

60
75
50 75
Turbidity, NTU
TSS, mgL-1

50
60 40 60
40
45 30 45
30
30 20 30
20

10 15 10 15

0 0 0 0
1 2 3 3 4 5 5 6 7 7 8 9 9 10 11 11 12 13 14 1 2 3 3 4 5 5 6 7 7 8 9 9 10 11 11 12 13 14
(c) Time, day (d) Time, day

Fig. 4. Results of cascade test in the presence of Pistia stratiotes: (a) BOD (mg L 1) and total sewage dosage vs. time; (b) COD (mg L 1) and total sewage
dosage vs. time; (c) TSS (mg L 1) and total sewage dosage vs. time; (d) turbidity (NTU) and total sewage dosage vs. time.
 ARTICLE IN PRESS
Y. Zimmels et al. / Journal of Environmental Management 81 (2006) 420428
5 1
100.0
BOD, mgL-1
10.0
1.0
0 1 2 3
(a)
5 1
100.0
COD, mgL-1
10.0
1.0
0 1 2 3
(b)
Fig. 5. Results of circulation test in the presence of Pistia stratiotes: 110 L h 1; 220 L h 1; 330 L h 1; 440 L h 1; 5no circulation; 6control (no
plants). (a) BOD (mg L 1) vs. time; (b) COD (mg L 1) vs. time.
460.5 mg L 1 initial value), with circulation at 40 L h 1,
was estimated at 404.5 mg L 1, as compared to
340.5 mg L 1 without circulation, and 370.6 mg L 1 in the
control (with no plants and circulation). After 8 days with
plants and the same circulation, the COD decreased to
42.958.9 mg L 1, the corresponding result with plants and
without circulation being 72.9 mg L 1, as compared to
78.8 mg L 1 in the control. The circulation had the effect of
decreasing the COD by 30 to 35.9 mg L 1. Note the
enhanced COD removal kinetics due to circulation; for
example, with circulation, 100 mg L 1 was reached after
2.54 days of treatment, while 67 days were required to
this end without circulation.
A similar response of turbidity to circulation was
observed. After 6 days with plants and circulation at
1040 L h 1 (graphic data not shown), the initial level of
212 NTU was brought down to 2.75.5 NTU. In the
absence of circulation (control containers), the turbidity
level reached 16.429 NTU. Thus, the effect of circulation
on the daily rate of turbidity removal (in the presence of
425
2 3 4 6
4 5 6 7 8
Time, day
2 3 4 6
4 5 6 7 8
Time, day
Pistia stratiotes) was estimated at 2.34.4 NTUday 1, or
10.520.1 NTU day 1 m 2. The turbidity removal kinetics
were similar to that of the COD. For example, with
circulation at 40 L h 1, 6.8 NTU was reached after 4 days
of treatment, while 8 days were required to this end without
circulation.
3.4. Set no. 4. Outdoors setup
The daily dosage of sewage was increased every 5 days
from 0.2 up to 1.5 m3/day (Fig. 6a). An estimated total of
28.8 m3 of wastewater was treated in this test. The plants
were capable (under summer conditions, with air tempera-
ture up to 33 1C) of holding the BOD below 10 mg L 1 and
the COD below 70 mg L 1 in both (lower and upper) pools
(Fig. 6b and c). The TSS level was kept below 15 mg L 1, as
compared to its 100150 mg L 1 initial content. The
optimal daily dosage in this set was estimated as
1.2 m3 day 1. Notwithstanding the considerable changes
and uctuation of the feed and treated sewage quality, the
 ARTICLE IN PRESS
426 Y. Zimmels et al. / Journal of Environmental Management 81 (2006) 420428

1.6 35 ability to continue with an effective treatment using plants

1.4 30 also under the Israeli winter conditions.
1.2
25
Daily dosage, m3

Total dosage, m3
4. Discussion and conclusions
1
20
0.8 Wastewater treatment systems using water hyacinth have
15
0.6 been in practice in the states of California, Florida,
0.4
10 Massachusetts and Texas (US EPA, 1988). The treatment
5
efciency was generally good. For example, BOD removal
0.2
efciencies ranged from 37% to 91%, while that of TSS
0 0 ranged from 21% to 92%. DeBusk and Reddy (1987)
1 6 11 16 21 26 31 36
(a) Time, day performed experiments in tanks with semi-continuous
sewage ow in the presence and absence of water hyacinths
100 and pennywort. In these xed ow rate experiments
80 (1.5 L min 1), removal of BOD from 130 down to
Content, mgL-1

10 mg L 1 was established.
60
Natural-growing aquatic macrophytes can be used to
COD, mg/l
40 remove nitrates, phosphates and heavy metals, by con-
BOD, mg/l
20 suming them in the form of plant nutrients (Reddy et al.,
0
1991; Elifantz and Tel-Or, 2002; Tripathi et al., 1991). An
0 5 10 15 20 25 30 35 40 integrated pond system, consisting of duckweed and algae
(b) Time, day ponds, was investigated for treatment of anaerobically
80
treated domestic wastewater (Van der Steen et al., 1998).
Fifty-six percent of the pond inuent nitrogen, mainly
60
ammonium, was removed. Results of studies on wastewater
Content, mgL-1

COD, mg/l purication by aquatic plants as well as of coliform tests

40
BOD, mg/l
20
0 articial wetland systems with the option of circulation, for
0 5 10 15 20 25 30
(c) Time, day
Fig. 6. Results of semi-continuous treatment with variable daily dosage of
sewage. (a) Daily and total dosage vs. time; (b) BOD and COD vs. time for
upper pool. (c) BOD and COD vs. time for lower pool.
pilot-scale test conrmed the capacity of the plants to hold
reasonably low levels of BOD and COD and a very low
level of TSS. This demonstrated the effectiveness of the
method for purication of sewage during the summer
period.
3.5. Set no. 5
The circulation rate affected the performance of
the purication system in the presence of plants, see
Fig. 7(ac). At 12 L min 1 the BOD uctuated between 1
and 4 mg L 1, COD between 50 and 70 mg L 1 and
turbidity between 2 and 6 NTU, for 18 days. Decreasing
the circulation rate to 8 L min 1 resulted in a shift in the
BOD, COD and turbidity to higher levels of 3 to 6 mg L 1,
50 to 85 mg L 1 and 3 to 8.5 NTU, respectively. On
average, the shift upwards was 4 mg L 1 BOD, nearly
10 mg L 1 of COD and 4.5 NTU of turbidity. Recalling
that winter conditions prevailed, with temperature ranging
in the greenhouse from 16 to 18 1C, the results conrm the
were reported by Zimmels et al. (2004). They tested
wastewater from the Kishon River and urban sewage
under laboratory and eld conditions without circulation.
The potential use of aquatic plants in Israel, as FWS
35 40 the purpose of improving water quality of urban sewage
was studied in this work. Circulation enhances the kinetics
of the processes that lower the levels of BOD, COD and
turbidity, see Fig. 5(a, b). This means that the organic
loading rate may be increased without the need to extend
the active treatment area, while maintaining the same
quality of the puried water.
The cascade laboratory system operated in pulsed mode
improved the BOD, COD, TSS and turbidity levels.
Aquatic plants and the imposed circulation produced
better kinetic features for sewage purication as compared
to the control. This difference became more pronounced
when the range of low levels of the water-quality indicators
were reached. In summary, the present results conrm the
capacity of a pulsed cascade of aquatic plants, Eichhornia
crassipes and Pistia stratiotes, with circulation and a well-
developed root system, to purify sewage in shorter periods
of time. Notwithstanding the considerable uctuation of
the feed and treated sewage quality, the laboratory-scale
tests conrm the capacity of the plants to reach and hold
reasonably low levels of BOD (57 mg L 1) and COD
(4050 mg L 1) and very low levels of TSS (35 mg L 1)
and turbidity (12 NTU).
In general, this agrees with results available from the
literature. For example, the BOD level was reduced from
123 to 25 mg L 1 in efuents obtained from wetlands
 ARTICLE IN PRESS
Y. Zimmels et al. / Journal of Environmental Management 81 (2006) 420428 427

Flow rate BOD

14 7 Flow rate COD
14 90
12 6
80
Flow rate, Lmin-1

12
5

Flow rate, Lmin-1

10 70

BOD, mgL-1
10

COD, mgL-1
8 4 60
8 50
6 3 40
6
4 2 4 30
20
2 1 2 10
0 0 0 0
3 4 5 8 10 15 17 18 22 23 24 25 29 3 4 5 8 10 15 17 18 22 23 24 25 29
(a) Time, day (b) Time, day

Flow rate Turbidity

14 9

8
12
7
10
Flow rate, Lmin-1

Turbidity, NTU
8 5

6 4

3
4
2
2
1

0 0
3 8 10 11 15 17 18 22 23 24 25 29
(c) Time, day

Fig. 7. Results of continuous treatment with xed sewage dosage (1 m3 day 1) at 130 mg L 1
BOD and 425 mg L 1
COD. (a) BOD and ow rate vs. time;
(b) COD and ow rate vs. time; (c) turbidity and ow rate vs. time.

(Nelson et al., 1999). Treatment of dirty water from dairy were held below 4 and 70 mg L 1, respectively. A decrease
farms, using a soil-based batch circulation system was in the circulation ow rate, from 12 to 8 L min 1, was
analyzed during a period of 25 days (Turrel and Leeds- found to reduce the BOD and COD levels to 6 and
Harrison, 2004). The results indicated that using a 7086 mg L 1, respectively, see Fig. 7(a, b). In this context,
circulation system holds some promise in the context of the increased circulation, from 8 to 12 L min 1, was found
cost savings. In this case, 90% removal of BOD and TSS to enhance the kinetics of average BOD and COD removal.
was achieved after 3 weeks. Macrophytes mediated transfer This translates to an increase in the allowed organic
of oxygen to the rhizosphere, by leakage from roots, and loading per unit pond area. The improvement of BOD and
increased aerobic degradation of organic matter and COD removal kinetics can be attributed to increased mass
nitrication (Brix, 1997). The attendant increase in aerobic transfer which is associated with the ow, and the
degradation of organics is attributed to the effect of enhanced aerobic degradation of organic matter due to
circulation. the effect of circulation. These results are in agreement with
The plants (Eichornia crassipes and Pistia stratiotes) were those reported in the literature (Turrel and Leeds-Harrison,
capable (within the test time) of lowering major quality 2004; US EPA, 1988, 1993). Furthermore, the plants
variables to levels required by national and local guidelines eliminate the disturbing smell of sewage water which poses
for irrigation water. This was veried on laboratory as well serious problems in several locations across the country,
as on pilot scale. The articial wetland systems, that and circulation can help reduce inlet odors. In addition to
performed inside a greenhouse structure with efuent its being a research facility, the new pilot unit, designed and
circulation, continued the effective treatment of water constructed as a closed circuit ecological system in the yard
during the winter season. In this context, the closed of the Civil and Environmental Engineering Department of
greenhouse FWS system helps to safeguard the environ- the Technion, is also designated to supply puried
ment from aggressive spreading of the aquatic plants, e.g. irrigation water to the Technion Ecological Garden.
to other water reservoirs.
It was shown that the plants were capable (even under Acknowledgements
winter conditions) of overcoming changes in the feed and
still maintaining the required levels of the measured water This research was supported by the Keren Kayemeth
quality indicators. At 12 L min 1, the BOD and COD levels LeIsrael, the Center for Absorption in Science, Ministry of
 ARTICLE IN PRESS
428 Y. Zimmels et al. / Journal of Environmental Management 81 (2006) 420428
Immigrant Absorption, State of Israel, and The Grand
Water Research Institute. The contributions of Dr. G.
Papkov, A. Kadmon and Ing. S. Roitman in performing
laboratory and eld tests are hereby acknowledged.
References
APHA, AWWA and WPFC, 1998. Standard Methods for the Examina-
tion of Water and Waste water, 19th ed. American Public Health
Association, Washington.
Bhamidimarri, R., Shilton, A., Armstrong, I., Jacobson, P., Scarlet, D.,
1991. Constructed wetlands for wastewater treatment: the New
Zealand experience. Water Science Technology 24 (5), 247253.
Brix, H., 1997. Do macrophytes play a role in constructed treatment
wetlands? Water Science Technology 35 (5), 1117.
Campbell, C.S., Ogden, M., 1999. Constructed Wetlands in the Sustain-
able Landscape. Wiley, New York.
Chescheir, G.M., Gilliam, J.W., Skaggs, R.W., Broadhead, R.G., 1991.
Nutrient and sediment removal in forested wetlands receiving pumped
agricultural drainage water. Wetlands 11, 7103.
Cooke, L.G., 1994. Nutrient transformation in a natural wetland receiving
sewage efuent and the implications for waste treatment. Water
Science and Technology 29 (4), 209217.
DeBusk, T.A., Reddy, K.R., 1987. BOD removal in oating aquatic
macrophyte-based wastewater treatment systems. Water Science and
Technology 19 (12), 273279.
Elifantz, H., Tel-Or, E., 2002. Heavy metal biosorption by plant biomass
of the macrophyte Ludvigia Stolonifera. Water, Air and Soil Pollution
141, 207218.
Gale, P.M., Devai, L., Reddy, K.R., Graetz, D.A., 1993. Denitrication
potential of soil from constructed and natural wetlands. Ecological
Engineering 2, 119130.
Gopal, Br., 1999. Natural and constructed wetlands for wastewater
treatment: potentials and problems. Water Science and Technology 40
(3), 2735.
Hammer, D.A. (Ed.), 1990. Constructed Wetlands for Wastewater
Treatment: Municipal, Industrial, and Agricultural. Lewis Publishers,
Chelsea, MI, p. 831.
Kadlec, R.H., 1995. Overview: surface ow constructed wetlands system.
Water Science and Technology 32 (3), 112.
Kadlec, R.H., Knight, R.L., 1995. Treatments Wetlands. Lewis Publisher,
Boca Raton, USA.
Moshiri, G.A. (Ed.), 1993. Constructed Wetlands for Water Quality
Improvement. Lewis Publishers, Boca Raton, FL, p. 632.
Nelson, M., Finn, M., Wilson, C.H., et al., 1999. Bioregenerative recycling
of wastewater in Biosphere 2 using a constructed wetland: 2-year
results. Ecological Engineering 13 (14), 189197.
Reddy, K.R., Agami, M., DAngelo, E.M., Tucker, J.C., 1991. Inuence
of potassium supply on growth and nutrient storage by water hyacinth.
Bioresource Technology 39 (1), 7984.
Reuter, J.E., Djohan, T., Goldman, C.R., 1992. The use of wetlands for
nutrient removal from surface runoff in a cold climate region of
Californiaresults from a newly constructed wetland at Lake Tahoe.
Journal of Environmental Management 36 (1), 3553.
Schwartz, M.F., Boyd, C.E., 1995. Constructed wetlands for treatment of
channel catsh pond efuents. Progressive Fish-Culturist 57, 255266.
Tripathi, B.D., Shukla, S.A., 1991. Biological treatment of wastewater by
selected aquatic plants. Environmental Pollution 69 (1), 6978.
Tripathi, B.D., Srivastava, J., Misra, K., 1991. Nitrogen and phosphorus
removal-capacity of four chosen aquatic macrophytes in tropical
freshwater ponds. Environmental Conservation 18, 143147.
Turrel, S.F., Leeds-Harrison, P.B., 2004. Treatment of dirty water from
dairy farms using a soil-based batch recirculation system, Proceedings
of ninth International Conference on Wetland Systems for Water
Pollution Control, Avignon (France), vol. 1, pp. 133140.
US Environmental Protection Agency, 1988. Constructed Wetlands and
Aquatic Plant Systems for Municipal Wastewater Treatment: Design
Manual. Ofce of Research and Development, Center of Environ-
mental Research Information, Cincinnati, OH 83p.
US Environmental Protection Agency, 1993. Subsurface ow constructed
wetlands for wastewater treatment. A Technical Assessment. Ofce of
Water, Washington, DC.
US Environmental Protection Agency, 2000. Constructed Wetlands
Treatment of Municipal Wastewater: Manual. Ofce of Research
and Development, National Risk Management Research Laboratory,
Cincinnati, OH 166p.
Van der Steen, P., Brenner, A., Oron, G., 1998. An integrated duckweed
and algae pond system for nitrogen removal and renovation. Water
Science and Technology 38 (1), 335343.
Zimmels, Y., Kirzhner, F., Roitman, S., 2004. Use of naturally growing
aquatic plants for wastewater purication. Water Environmental
Research 76 (3), 220230.