Bioresource Technology 96 (2005) 969–976

Cell-to-cell pollution reduction effectiveness of subsurface
domestic treatment wetlands
David N. Steer a,*, Lauchlan H. Fraser b, Beth A. Seibert c
a

Geology Department, The University of Akron, Akron, OH 44325-4101, USA
Biology Department, The University of Akron, Akron, OH 44325-3908, USA
c
Ottawa River Coalition, 3900 Campus Dr., Suite A, Lima, OH 45804-3596, USA
b

Received 8 February 2002; received in revised form 30 July 2004; accepted 5 August 2004
Available online 2 October 2004

Abstract
Quarterly water quality data from 1998 to 2003 for eight single-family domestic systems serving 2–7 people in Ohio, USA, were
studied to determine the cell-to-cell and system wide pathogen reduction efficiency and effectiveness of these systems in meeting compliance standards. Two-cell domestic wastewater treatment systems displayed significant variability in their cell-to-cell performance
that directly impacted the overall ability of systems to meet effluent compliance standards. Fecal coliform was effectively reduced
(99%) in these systems while two-thirds of the input biochemical oxygen demand was mitigated in each of the cells of these systems. Fecal coliform and biochemical oxygen demand were typically reduced below 2000 counts per 100 ml and 15 mg/l (respectively)
before discharge to surface waters. Total suspended solids were reduced by 80% overall with cell one retaining the majority of the
solids (70%). These systems discharged more than 18 mg/l of suspended solids in less than 5% of the samples thus displaying a very
high compliance rate. Ammonia and total phosphorus were less effectively treated (30–40% reductions in each cell) and exceeded
standards (1.5 mg/l) more frequently. Analyses based on the number of occupants indicated that the two-cell design used here was
most effective for smaller occupancy systems. More study is required to determine the value of this design for large occupancy systems. In the future, wetlands should be evaluated based on the total loads delivered to the watershed rather than by effluent
concentrations.
Ó 2004 Elsevier Ltd. All rights reserved.
Keywords: Treatment wetlands; Efficiency; Pathogens; Effectiveness

1. Introduction
The use of wetlands to treat domestic wastewater has
gained global attention in recent years (Bastian et al.,
1987) because they are seen as a low-maintenance management strategy that is an alternative to traditional
treatment systems (Kadlec and Knight, 1996). These
natural systems are known to effectively mitigate a variety of pollutants (Wood, 1995; Nokes et al., 1999;
Mitsch and Gosselink, 2000). The ability of wetland sys*

Corresponding author. Tel.: +1 330 972 2099; fax: +1 330 972
7611.
E-mail address: steer@uakron.edu (D.N. Steer).
0960-8524/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.biortech.2004.08.006

tems to effectively reduce total suspended solids, biochemical oxygen demand (Watson et al., 1990) and
fecal coliform (Nokes et al., 1999; Neralla et al., 2000)
are well established. Nitrogen (ammonia and total nitrogen) and phosphorus are processed with relatively low
efficiency by most wetland systems (Nichols, 1983;
Mann, 1990; Hammer and Knight, 1994; Urbanc-Bercic
and Bulc, 1995; Cronk and Fennessy, 2001). Wastewater
designs employing wetlands that discharge to the surface
may be well suited for areas with poor drainage or hydric soils that limit effective use of traditional leach fields.
The systems discussed in this paper discharge to the
surface and are considered non-point sources of pollution because of the small volumes of effluent and widely

970

D.N. Steer et al. / Bioresource Technology 96 (2005) 969–976

separated geographic locations involved. Surface discharging domestic wastewater wetland treatment systems in the US need to meet National Pollution
Discharge Elimination Systems (NPDES) permit guidelines to be in compliance with pollution reduction goals
implemented under the watershed Total Maximum
Daily Loads (TMDL) program. The TMDL program
involves establishment of maximum pollution loads for
impaired river segments. The NPDES program focuses
on improving water impaired by pollution and promoting wide-ranging solutions to the myriad of problems
threatening water quality (USEPA, 2001). Regulatory
guidelines in Ohio, USA that were established to comply
with NPDESÕ, lower several pathogen limits (OEPA,
1999: Table 1) and discourage off-lot discharge of any
kind (OEPA, 2001). The effluent limits required by the
household sewage disposal permitting system were derived from EPA guidance and regulations (USEPA,
2001). Implementation of the guidelines is delegated to
local health departments under blanket authority
through memoranda of agreements with the US Environmental Protection Agency. Since permit writers must
consider every surface water discharge in their total
management strategy, non-point source systems such
as these household treatment systems may be required
to meet more stringent effluent regulations. This study
addresses performance of domestic treatment wetlands
in meeting specific standards. Such standards are directly applicable in the US, but are relevant worldwide
where similar performance criteria may eventually be
required.
Surface discharging domestic treatment wetlands of
the future must reduce effluent loads effectively in order
to remain a viable alternative for homeowners. Numerous studies in the literature have documented overall
performance and design of domestic systems (Kadlec
and Knight, 1996; Mitsch and Gosselink, 2000). The literature is also replete with US case studies (Alabama—
White and Shirk, 1998; Kentucky—Watson et al., 1990;
Thom et al., 1998; Texas—Neralla et al., 2000; Ohio—
Steer et al., 2002). Several major conclusions can be
drawn from these studies. Calculations of system efficiencies using pathogen concentrations indicate that
these systems display similar load reduction ranges despite widely differing climatic and geomorphologic conditions. Individually the data display large variances in
the actual values while limited examples with flow meas-

Table 1
EPA surface discharge limits (maximum concentrations)
Pathogen

Concentration

Units

Fecal Coliform
BOD5
TSS
Ammonia

2000
15
18
1.5

counts/100 ml
mg/l
mg/l
mg/l

urements reduce their utility for quantitative kinetics
modeling (see Mitchell and McNevin, 2001 for a review
of BOD kinetics).
Design improvements or modifications necessary to
meet new effluent guidelines can only be accomplished
by detailed documentation of the operational characteristics of subsurface treatment wetland systems. Steer
et al. (2002) reported initial results for these systems
while focusing on their overall ability to reduce pathogens as compared to more mature wetlands. This study
uses two additional years of data to explore the cell-tocell performance of eight domestic treatment wetlands
(IDs 13–17, 19–21 from Steer et al., 2002). Since multicell wetlands are rarely reported on a cell-by-cell basis,
these data serve as an example of system performance
during the first 5 years of wetland development. As these
systems become more widely used, compliance information will be valuable for scientists and engineers preparing watershed management plans.

2. Methods
2.1. System description
The systems cited in this study incorporated a single
septic tank for primary treatment and two subsurface
flow wetlands cells connected in series (Fig. 1). These
wetland designs all consisted of two 25 m2 treatment
cells with a depth of 0.46 m that employed gravity feed
using a 1° slope. Water level was maintained 6 cm below the surface through manual adjustment of inlet
pipes. The substrate consisted of 3 cm diameter clean
riverbed gravel. Runoff was prevented from entering
the cells by surrounding earthen berms. The same design
was used for all systems regardless of the number of
occupants. The design is consistent with a conservative
design for a four-occupant home based on the EPA design manual (USEPA, 2000). These systems also have
similar characteristics and fall within the typical 10–
100 m2 size reported in a review of domestic wetland
treatment systems (Mankin and Powell, 1998).
Several physical and operational features of these systems were unique to this study. These wetlands were
constructed with the first cell sealed from the surrounding substrate to prevent infiltration into the soil column.
Seven of the systems were sealed with compacted clay
and one had a rubber liner. The second cells were intentionally not isolated from the natural soil, as percolation
and use of the soil column was deemed as a positive element of treatment and as a means of reducing off-lot discharge. These systems are all located in the Ottawa
River watershed of western Ohio, serve from 1 to 7 people, were constructed and began operation in 1998. Wetlands were planted in 1998 or 1999 using Scirpus
(bulrushes) or Sagittaria (arrowhead) in the first cell

D.N. Steer et al. / Bioresource Technology 96 (2005) 969–976

971

Fig. 1. Schematic diagram of wetland systems used in this study. Quarterly water quality samples were collected at the outlet to te septic tank (water
control box 1), the distribution box between cells 1 and 2 and at the outlet box to cell 2.

and ornamental wetland plants such as Acorus calamus,
Lobelia cardinalis, Asclepias incarnata and Pontederia
cordata in the second treatment wetland.
2.2. Sample collection and analyses
Water samples were collected quarterly from each
wetland. Sampling began in 1999 when the wetland systems were installed as part of a local watershed-monitoring program and continued through summer 2003 (21
sample dates). Samples were collected from the septic
tank, distribution box between the first and second cell,
and the outlet distribution box that connects to the box
that is connected to drainage tile (see Fig. 1). Grab samples were collected and analyzed to determine concentrations of fecal coliform, BOD5, ammonia, NO3/NO2,
total nitrogen, total suspended solids and total phosphorus using standard protocols (USEPA, 1983; APHAAWWA-WPCF, 1998). Hydraulic loads of these systems
were not measured. See Steer et al. (2002) for a complete
description of the specific sampling procedures used to
collect data for this study.

system. In several cases, flow was conducted through
one or more cells, but not through the final cell. In these
cases (43/108), a value of zero was entered for the final
effluent concentration for the particular cell (usually
stage 3). Such cases were considered optimum performance since no effluent was released to the environment.
Average inlet, intermediate and outlet concentrations
were calculated for individual systems and for groups of
systems. Because of the large variance in these data, confidence intervals were reported using the standard error
of the mean, σN (where N is the sample size and r is the
standard deviation). Pathogen reduction efficiencies
were determined using 1—(outlet concentration/inlet
concentration) with probabilities calculated using simple
t-tests assuming equal variance. Outlet counts or concentrations were compared to suggested USEPA effluent
limits for discharge to surface waters (Table 1). Per EPA
sampling guidance, a sample met compliance standards
if the effluent exceeded the limit during one period, but
was below the limit during the next consecutive sampling period.

2.3. Data reduction

3. Results

Data were binned into two groups based on household occupancy (Group 1: 1–3 or Group 2: 5–7 persons)
and were edited prior to analysis. Fecal coliform counts
were underestimated in these data (particularly pertinent
for the inlet) because they were generally counted to a
maximum 20,000. BOD5 samples that unusually exceeded 300 mg/l (4 of 124 samples) were removed from
analyses because such high levels likely indicated a sampling error or malfunctioning septic tank. Other BOD5
samples (18/124) were removed from analyses because
of dilution-related lab errors (concentrations annotated
as ‘‘>’’). Total suspended solid samples that were two
or more orders of magnitude higher than the average
value for the grouped systems (5 of 151 samples) were
also removed as sampling errors. Data were neglected
if ‘‘no flow’’ or ‘‘no access’’ were recorded for the entire

3.1. Fecal coliform
Fecal coliform bacteria were reduced significantly
in these systems and typically met effluent standards
(Table 2). The first treatment cell reduced fecal concentrations by 83% for both occupancy groups
(p < 0.00001). The second cell reduced the remaining
concentration by 97% (p < 0.00001). Total system
pathogen reduction efficiencies for the systems were
99% for both groups. Average output concentrations
of 1248 ± 326 counts/100 ml for Group 1 (low occupancy systems) systems were below the 2000 counts
per 100 ml standard. One system (ID 13) exceeded standards in 35% of the samples. Other low occupancy systems routinely met standards (in 88% or more of the
samples). The average 2494 ± 590 counts per 100 ml

972

D.N. Steer et al. / Bioresource Technology 96 (2005) 969–976

Table 2
Average fecal coliform count (log # per 100 ml) at each sampling point, cell pathogen reduction efficiencies and system compliance rate with EPA
standards
System ID
(Occupants)

Inlet from
Septic

Outlet Cell 1

Reduction
Cell 1 (%)

Cell 2
Outlet

Reduction
Cell 2 (%)

Reduction
Aggregate (%)

Compliance
<2000 (%)

Observations N

13(2)
14(1)
17(2)
21(3)
Combined

4.27 ± 0.049
4.27 ± 0.02
4.25 ± 0.02
4.33 ± 0.01
4.28 ± 0.02

3.59 ± 0.42
3.53 ± 0.69
3.47 ± 0.48
3.26 ± 0.29
3.47 ± 0.47

79
82
84
91
84

2.89 ± 0.93
1.19 ± 2.08
0.69 ± 1.86
2.14 ± 1.05
1.71 ± 2.17

80
99
99
92
98

96
99
99
99
99

65
100
100
89
88

21
21
18
19
79

15(5)
16(6)
19(7)
20(5)
Combined

4.24 ± 0.04
4.21 ± 0.18
4.28 ± 0.04
4.17 ± 0.28
4.22 ± 0.14

3.58 ± 0.30
3.59 ± 0.41
3.47 ± 0.72
3.24 ± 0.68
3.47 ± 0.52

78
76
85
88
82

1.91 ± 1.75
2.41 ± 1.27
2.01 ± 2.99
2.02 ± 2.21
2.10 ± 2.01

98
93
96
94
96

99
98
99
99
99

89
86
95
95
91

18
21
22
20
81

releases of Group 2 systems (higher occupancy) exceeded the 2000 counts per100 ml standard. Two of these
systems (IDs 15 and 16) displayed average effluent levels
below limits but exceeded the fecal standard in approximately 15% of the samples. Two (IDs 19 and 20) exceeded standards on average, but met the effluent
standard in 95% of the samples.

(24.4 ± 12.2 and 20.9 ± 3.0 respectively) and failed to
meet standards in more than 45% of the individual samples. Average output BOD5 concentrations for the high
occupancy systems of 23.9 ± 3.9 mg/l exceeded the suggested limit. One system (ID 16) displayed output levels
nearly 3 times higher than all other systems.
3.3. Total suspended solids

3.2. BOD5
Biochemical oxygen demand reduction and compliance with standards varied significantly from system to
system in the wetlands (Table 3). The first wetland cell
reduced BOD5 by 66% (p < 0.00001) for both low
and high occupancy systems. The second cell reduced
the remaining concentrations by 74% (p < 0.00001)
in Group 1 systems and 59% (p < 0.0005) in the Group
2 systems. Overall these systems reduced 91% (Group 1:
p < 0.00001) and 86% (Group 2; p < 0.00001) of the
BOD5. The Group 1 average output biochemical oxygen
demand of 13.8 ± 3.2 mg/l was below the 15.0 mg/l recommended standard. Average BOD5 concentrations
for two of these systems (IDs 14 and 17) always met
the standards (100% compliance). Two other systems
(IDs 13 and 21) averaged well above the standard

Total suspended solids (TSS) were reduced effectively
enough to routinely meet effluent limits (Table 4). The
first cell processed 70% (p < 0.00001) of the input
TSS based on concentrations. The second cell played little or no role in water quality improvement as it reduced
39% (low occupancy; p < 0.1) and 25% (high occupancy;
p < 0.1) of the remaining solids. In some systems, solids
appeared to be remobilized (e.g. IDs 20 and 21). Overall
these systems removed 83% (Group 1: p < 0.00001) and
77% (Group 2; p < 0.00001) of the total suspended solids. Average outlet total suspended solids (TSS) of
6.37 ± 1.42 (Group 1) and 8.83 ± 1.46 (Group 2) mg/l
met OEPA/EPA standards of 18 mg/l (Table 4). With
the exception a single system (ID 21 at 78% compliance),
these wetlands met standards in almost all cases (95–
100% compliance).

Table 3
Average BOD5 concentrations (mg/l) at each sampling point, cell pathogen reduction efficiencies and system compliance rate with EPA standards
System ID
(Occupants)

Inlet from
Septic

Outlet Cell 1

Reduction
Cell 1 (%)

Cell 2 Outlet

Reduction
Cell 2 (%)

Reduction
Aggregate (%)

Compliance
<15 mg/l (%)

Observations N

13(2)
14(1)
17(2)
21(3)
Combined

162.2 ± 20.8
131.4 ± 16.4
125.6 ± 17.6
249.5 ± 27.1
162.1 ± 11.8

71.3 ± 14.8
23.7 ± 3.8
83.4 ± 10.0
45.2 ± 8.0
53.8 ± 5.6

56
82
34
82
67

24.4 ± 12.2
2.6 ± 1.1
4.1 ± 2.3
20.9 ± 3.0
13.8 ± 3.2

52
89
95
54
74

79
98
97
92
91

55
100
100
36
76

11
16
13
11
51

15(5)
16(6)
19(7)
20(5)
Combined

137.9 ± 15.5
173.9 ± 18.9
244.2 ± 25.8
141.6 ± 15.1
171.2 ± 10.7

45.2 ± 5.5
89.5 ± 10.8
63.2 ± 7.9
39.3 ± 8.8
58.9 ± 4.9

67
49
74
72
66

14.0 ± 3.0
53.3 ± 9.1
10.3 ± 4.3
16.7 ± 7.4
23.9 ± 3.9

69
40
84
58
59

90
69
96
88
86

73
29
83
79
65

15
14
12
14
55

D.N. Steer et al. / Bioresource Technology 96 (2005) 969–976

973

Table 4
Average TSS Concentrations (mg/l) at each sampling point, cell pathogen reduction efficiencies and system compliance rate with EPA standards
System ID
(Occupants)

Inlet
from Septic

Outlet Cell 1

Reduction
Cell 1 (%)

Cell 2
Outlet

Reduction
Cell 2 (%)

13(2)
14(1)
17(2)
21(3)
Combined

47.28 ± 9.63
25.17 ± 5.84
27.84 ± 6.82
54.41 ± 13.65
38.17 ± 4.67

11.86 ± 3.69
8.14 ± 3.18
10.15 ± 2.05
11.17 ± 2.71
10.36 ± 1.48

75
68
64
79
73

5.72 ± 0.97
3.25 ± 1.81
5.40 ± 4.61
12.58 ± 3.60
6.37 ± 1.42

15(5)
16(6)
19(7)
20(5)
Combined

24.81 ± 3.09
45.49 ± 5.33
38.62 ± 6.49
42.13 ± 4.35
37.92 ± 2.59

12.75 ± 4.70
12.73 ± 3.98
10.95 ± 3.06
10.60 ± 1.80
11.76 ± 1.76

49
72
72
75
69

7.78 ± 2.69
7.51 ± 1.85
9.01 ± 4.07
10.98 ± 3.02
8.83 ± 1.46

Reduction
Aggregate (%)

Compliance
<18 mg/l (%)

Observations N

52
60
47
13
39

88
87
81
77
83

100
100
100
78
94

20
18
17
19
74

39
41
18
4
25

69
83
77
74
77

100
100
100
95
99

18
19
20
20
77

Table 5
Average ammonia concentrations (mg/l) at each sampling point, cell pathogen reduction efficiencies and system compliance rate with EPA standards
System ID
(Occupants)

Inlet from
Septic

Outlet Cell 1

13(2)
14(1)
17(2)
21(3)
Combined

36.84 ± 3.95
31.86 ± 2.21
23.32 ± 2.34
43.67 ± 3.08
33.95 ± 1.73

28.94 ± 2.70
17.29 ± 2.03
26.09 ± 2.15
29.61 ± 6.73
25.33 ± 1.94

15(5)
16(6)
19(7)
20(5)
Combined

29.52 ± 2.80
40.85 ± 2.68
42.01 ± 4.64
102.01 ± 10.54
53.04 ± 4.53

21.36 ± 2.17
33.16 ± 2.30
21.79 ± 2.23
64.93 ± 7.64
35.07 ± 2.99

Reduction
Cell 1 (%)

Cell 2 Outlet

Reduction
Cell 2

Reduction
Aggregate (%)

Compliance
<1.5 mg/l (%)

Observations N

21
46
12
32
25

18.41 ± 2.21
1.96 ± 1.17
3.42 ± 1.90
12.47 ± 1.78
9.14 ± 1.23

36
89
87
58
64

50
94
85
71
73

12
88
93
13
52

17
17
15
15
64

28
19
48
36
34

8.81 ± 1.95
23.38 ± 1.98
9.13 ± 2.22
33.50 ± 7.48
18.63 ± 2.35

59
29
58
48
47

70
43
78
67
65

41
0
50
50
35

17
17
16
16
66

3.4. Ammonia
Ammonia reduction efficiencies varied greatly from
wetland to wetland and none of the systems routinely
met standards (Table 5). Both cells appeared to contribute to ammonia reduction with 25–34% (p < 0.001) reduced in the first cell and 47–64% (p < 0.00001) of the
remaining amount mitigated in the second cell. Overall
these systems reduced 70% of the ammonia that entered
the systems from the septic tanks. The average ammonia
effluent discharged by Group 1 systems (9.14 ± 1.23 mg/l)
was over six times the 1.5 mg/l ammonia standard. Systems serving larger numbers of occupants (Group 2)
had average ammonia discharge concentrations double
those of Group 1 systems. Both groups of systems routinely exceeded the 1.5 mg/l ammonia standard and all
but two systems (IDs 14 and 17) even exceeded the total
nitrogen limit of 4.5 mg/l (OEPA, 1999) allowed for these
systems. The highest performing system (ID 17) met the
standard in 94% of the cases while the worst performing
system (ID 16) never met the standard.
3.5. Total phosphorus
Total phosphorus reduction efficiencies and compliance with suggested standards varied extensively one

wetland to another (Table 6). Both cells appeared to
contribute to phosphorus reduction with 32–38% reduced in the first (p < 0.000001) and second (p < 0.001)
cell. Overall these systems reduced 55% of the total
phosphorus that entered the systems from the septic
tanks. The average total phosphorus discharged by
Group 1 systems (2.79 ± 0.40 mg/l) was 70% of that
discharged by the systems serving larger numbers of
occupants of systems. Both groups of systems routinely
exceeded the 1.5 mg/l standard. The highest performing
system (ID 17) met the standard in 94% of the cases
while the worst performing system met the standard in
only 10% of the samples (ID 16).

4. Discussion
4.1. Fecal coliform
Wastewater treatment data indicated the number of
residents did not impact pathogen reduction efficiency
or compliance rates in these systems. The higher cell 2
processing efficiency (97%) compared to cell 1 (83%)
was not significant because cell 1 inlet concentrations
were under-reported (counts terminated at 20,000). Cell
2 outlet concentrations from low occupancy systems were

974

D.N. Steer et al. / Bioresource Technology 96 (2005) 969–976

Table 6
Average total phosphorus concentrations (mg/l) at each sampling point, cell pathogen reduction efficiencies and system compliance rate with EPA
standards
System ID
(Occupants)

Inlet from
Septic

Outlet Cell 1

Reduction
Cell 1 (%)

Cell 2 Outlet

Reduction
Cell 2

Reduction
Aggregate (%)

Compliance
<1.5 mg/l (%)

Observations N

13(2)
14(1)
17(2)
21(3)
Combined

9.96 ± 0.69
5.02 ± 0.39
5.80 ± 0.57
7.80 ± 0.60
7.22 ± 0.36

6.94 ± 0.40
2.70 ± 0.39
4.94 ± 0.37
3.36 ± 0.46
4.49 ± 0.29

30
46
15
57
38

6.09 ± 0.87
1.61 ± 0.57
0.79 ± 0.42
2.11 ± 0.54
2.79 ± 0.40

12
40
84
37
38

39
68
86
73
61

19
81
94
79
66

21
21
16
19
77

15(5)
16(6)
19(7)
20(5)
Combined

5.08 ± 0.52
8.02 ± 0.62
6.84 ± 0.68
15.41 ± 1.1
8.92 ± 0.58

3.45 ± 0.49
6.32 ± 0.56
3.42 ± 0.51
11.30 ± 1.67
6.08 ± 0.59

32
21
50
27
32

1.58 ± 0.37
6.02 ± 0.60
2.89 ± 0.82
5.47 ± 1.26
4.08 ± 0.46

54
5
15
52
33

69
25
58
65
54

78
10
55
60
49

18
21
20
20
79

slightly lower than those observed from the higher occupancy systems (p < 0.05) despite having nearly identical
(p = 0.5) cell 2 inlet concentrations. Closer examination
of the data (Table 2) indicated that these differences were
minimal and had little impact on the compliance rate.
Any differences in output loads are likely attributed to
changes in total hydraulic loads due to evaporation and
transpiration that occurred as effluent moved through cell
2. Approximately 70% of the low occupancy samples recorded no off-lot discharge from cell 2 compared to 10%
of those from Group 2 systems. Evapotranspiration is a
mechanism that is known to be a factor in other wetlands
(Kadlec and Knight, 1996; Mitsch and Gosselink, 2000)
and may have had more impact on the systems with smaller hydraulic loads (Group 1 systems). Compliance with
the 2000 counts per 100 ml standard did not appear to
be related to occupancy since all systems generally met
the 2000 counts per 100 ml fecal standard. Options such
as aeration in cell 2 or chlorination at the end of the cell
might serve to improve system performance to the levels
required to meet fecal coliform effluent standards in all
cases albeit at higher cost (Steer et al., 2003).
4.2. BOD5
Treatment efficiency and compliance with BOD5
standards did not appear to be related to input cell concentrations or household size. The actual average BOD5
loads delivered to the two groups of wetlands were 160–
170 mg/l (Table 3). The input BOD5 concentrations to
cell 1 reported here are well above the 100 mg/l upper
limit for effective removal based on EPA sizing criteria
(USEPA, 2000). Curiously, the percentage reduction of
BOD5 from the septic tank to the outlet of cell 1 and
the outlet of cell 1 to the outlet of cell 2 were similar despite much lower concentrations entering the second
cells (Table 3). The similar treatment efficiencies in both
cells may indicate that cell one did not reach the maximum capacity to effectively treat BOD5 even when well
above the design threshold. Low occupancy system

BOD5 final effluent levels were much lower than those
observed for the higher occupancy systems (p < 0.05).
However, one system (Group 2: ID 16) had significantly
higher cell 2 concentrations than any of the other systems. When that system was removed from the analyses,
there were no significant differences between the groups
(p = 0.5) for outlet BOD5. System 21 (Group 1) and system 16 (Group 2) routinely exceeded standards. System
21 had some of the highest BOD5 inlet concentrations
while ID 16 had some of the highest cell 2 inlet concentrations. With these systems removed from analyses, the
remaining systems complied with the 15 mg/l BOD5 limit
in 80% of the cases. These low effluent concentrations
and high compliance rates were expected based on studies from other regions (Bhamidimarri et al., 1991; Maehlum et al., 1995; Neralla et al., 2000).
4.3. Total suspended solids
These systems significantly reduced TSS in the wastewater and effectively met compliance standards for both
groups of systems. There were no significant differences
in input loads or individual cell treatment efficiencies between groups of systems. Inlet concentrations of TSS
were approximately 1/3 the suggested maximum design
concentration of 100 mg/l (USEPA, 2000). Cell 2 efficiencies were influenced by remobilization of solids that
occurred in two systems (Group 1: ID 21; Group 2: ID
20). For the remaining systems, cell 2 suspended solid
outlet concentrations were statistically indistinguishable
between the two groups (p > 0.05). On average, solids
were reduced below the 18 mg/l limit by the time the
effluent exited the first cell (Table 4). As such, systems
met compliance standards in a very high percentage of
the samples.
4.4. Ammonia
Ammonia mitigation efficiencies did not appear to be
related to the number of occupants and the systems were

D.N. Steer et al. / Bioresource Technology 96 (2005) 969–976

not able to consistently meet standards. Two of the systems (ID 17: low occupancy; ID 20: high occupancy) disproportionately influenced average cell 1 reduction rates
for these systems. System ID 17 had the lowest average inlet ammonia concentrations and actually appeared to
show an increase in ammonia through cell 1. Further
analysis indicated that there was no statistically significant difference between cell 1 inlet and outlet ammonia
concentrations for system ID 17 (p = 0.2). Group 2 system ID 20 displayed inlet ammonia concentrations that
were 2–3 times larger than observed for any other systems
(Table 6). When these two systems were removed from
the analyses, there was no difference in ammonia concentrations between systems at the various sample points.
The reduced data for all systems combined indicated that
cell 1 reduced 32% (p < 0.00001) and cell 2 reduced 51%
(p < 0.00001) of the remaining ammonia for an overall
ammonia reduction of 67% (p < 0.00001). These results
were not unexpected based on values reported in other
studies (Gersberg et al., 1986; Koottatep and Polprasert,
1997) and because subsurface treatment wetlands
are known to exhibit low ammonia removal efficiencies (Hammer and Knight, 1994; Koottatep and
Polprasert, 1997; Weaver et al., 2001). The low efficiencies
contributed to the inability of these systems to meet the
1.5 mg/l standard. A design change that includes aeration
could improve performance.
4.5. Total phosphorus
Total phosphorus reductions were not linked to the
number of occupants and the systems were able to meet
standards in the majority of the samples. One of the systems (ID 20: Group 2) displayed input phosphorus concentrations that were nearly double that of other
systems. For the remaining systems, there was no difference in total phosphorus concentrations between systems at the various sample points. The reduced data
for all systems combined indicated that cell 1 reduced
phosphorus 36% (p < 0.00001) and cell 2 reduced 30%
(p < 0.0005) of the remaining amount for an overall
phosphorus reduction of 55% (p < 0.00001). The poor
performance of these systems compared to the 80–97%
efficiency of similar systems in other studies (Maehlum
et al., 1995; Urbanc-Bercic and Bulc, 1995) is difficult
to explain. System efficiency at removing P is related
to the type of substrate (Richter and Weaver, 2003)
and has been shown to decrease over time, occasionally
with remobilization (Mann, 1990). Despite the low total
phosphorus removal rate, systems were able to meet the
1.5 mg/l standard in over 50% of the samples.
4.6. Compliance with standards
There is a discrepancy between current US Environmental Protection Agency compliance standards and

975

the USEPA National Pollution Discharge Elimination
Systems Total Maximum Daily Load policy. Compliance concentration standards (Table 1) that were developed to conform to USEPA (2001) guidelines can be
monitored quickly, at relatively low cost and rapidly
evaluated as pass or fail. However, monitoring concentrations has limited usefulness for water resource
managers because total loads delivered are of key
importance to the overall health of the watershed
(USEPA, 2001). Flows were not measured in these systems, but larger occupancy systems more frequently discharged to surface waters than small occupancy systems.
The true influence of these systems on water quality can
only be fully addressed if hydraulic loading rates are
determined and used in concert with effluent concentrations. A more effective method for evaluating the utility
of using these systems could be developed by comparing output loads of various wastewater treatment
options rather than by simply monitoring output
concentrations.

5. Conclusions
Analyses of effluent concentrations and standards
compliance from eight two-cell treatment wetlands indicated these systems effectively reduced pathogens and
their discharged effluent generally met EPA suggested
standards (Table 1). Fecal coliform concentrations were
reduced 83% in cell one and 97% in cell 2 with high
compliance rates (90%). BOD5 concentrations were reduced with lower efficiencies than those found for fecal
coliform (65% for cell 1 and 70% for cell 2). These systems met the 15 mg/l standard in 2/3 of the samples.
Solids were reduced 70% in cell one with cell two providing much less additional water quality improvement
(30% reduction). The two-cell design effectively met
18 mg/l standards showing higher than 90% compliance.
Ammonia reduction was actually higher in cell 2 (55%)
than it was in cell 1 (30%) for overall reductions of
70%. The low performance of these systems in processing ammonia resulted in frequent failures to meet the
1.5 mg/l standard (>50%). Total phosphorus was also
not well processed in these systems on a cell-to-cell basis
(35% for both cells 1 and 2) or overall (60%) with a
correspondingly low compliance level (55%). Additional research is needed to determine effective methods
for improving ammonia and total phosphorus processing of these systems. The performance documented here
suggests that trends in cell-to-cell pathogen reduction
are important to understanding system performance.
Additional study is needed to determine if systems with
a larger surface area would be more efficient (as suggested by Stecher and Weaver (2003)) or if other design
changes are required for large (5+) occupancy households. Monitoring-based regulations should be guided

976

D.N. Steer et al. / Bioresource Technology 96 (2005) 969–976

by total loads delivered to the watershed rather than by
effluent concentrations.

Acknowledgement
The authors would like to acknowledge the useful reviews of several anonymous reviewers that improved
this submission. We particularly wish to thank reviewer
Richard Weaver of Texas A&M for his willingness to
personally discuss issues related to data reduction and
system performance. The authors would like to thank
the technicians who collected these data that contributed
to this project. We also acknowledge that the funding to
originally collect and process portions of water quality
data were obtained through local governments and
organizations. This publication was financed in part
through a grant from the Ohio Environmental Protection Agency and the United States Environmental Protection Agency, under the provisions of Section 319(h)
of the Clean Water Act.

References
APHA-AWWA-WPCF, 1998. In: Clesceri, L.S., Greenberg, A.E.,
Eaton, A.D. (Eds.), Standard methods for examination of water and
wastewater, 20th ed. Am. Pub. Health Assoc., Washington, DC.
Bastian, R.K., Shanaghan, P., Thompson, B., 1987. Report on the use
of wetlands for municipal wastewater treatment and disposal (EPA
430/09-88-005). USEPA, Washington, DC.
Bhamidimarri, R., Shilton, A., Armstrong, I., Jacobson, P., Scarlet,
D., 1991. Constructed wetlands for wastewater treatment: The New
Zealand experience. Water. Sci. Technel. 24 (5), 247–253.
Cronk, J.K., Fennessy, M.S., 2001. Wetland Plants: Biology and
Ecology. Lewis Publishers, New York, 462 pp.
Gersberg, R.M., Elkins, B.V., Lyon, S.R., Goldman, C.R., 1986. Role
of aquatic plants in wastewater treatment by artificial wetlands.
Water. Res. 20 (3), 363–368.
Hammer, D.A., Knight, R.L., 1994. Designing constructed wetlands
for nitrogen removal. Water. Sci. Technol. 29 (4), 15–27.
Kadlec, R.H., Knight, R.L., 1996. Treatment Wetlands. Lewis
Publishers, New York, pp. 893.
Koottatep, T., Polprasert, C., 1997. Role of plant uptake on nitrogen
removal in constructed wetlands located in the tropics. Water. Sci.
Technol. 36 (12), 1–8.
Mann, R.A., 1990. Phosphorus removal by constructed wetlands:
Substratum adsorption. In: Cooper, P.F., Findlater, B.C. (Eds.),
Constructed wetlands in water pollution: Advances in water
pollution control. Pergamon Press, Great Britain, pp. 97–105.
Maehlum, T., Jenssen, P.D., Warner, W.S., 1995. Cold-climate
constructed wetlands. Water. Sci. Technol. 32 (3), 95–101.
Mankin, K.R., Powell, G.M., 1998. Onsite rock-plant filter monitoring
and evaluation in Kansas. In: Proceedings of 8th National
Symposium on Individual and Small Community Sewage Systems,
ASAE, St. Joseph, Mo.
Mitchell, C., McNevin, D., 2001. Alternative analysis of BOD removal
in subsurface flow constructed wetlands employing Monod kinetics. Water. Res. 35 (5), 1295–1303.

Mitsch, W.J., Gosselink, J.G., 2000. Wetlands, Third ed. Wiley, New
York, pp. 936.
Neralla, S., Weaver, R.W., Lesikar, B.J., Persyn, R.A., 2000.
Improvement of domestic wastewater quality by subsurface flow
constructed wetlands. Biores. Technol. 75, 19–25.
Nichols, D.S., 1983. Capacity of natural wetlands to remove nutrients
from wastewater. J. Water. Poll. Contr. Fed. 55, 495–505.
Nokes, R.L., Gerba, C.P., Karpiscak, M.M., 1999. Reduction of
enteric organisms in small-scale, subsurface flow constructed
wetlands. In: Means, J.L., Hinchee, R.E. (Eds.), Wetlands and
remediation: Wastewater remediation and treatment. Battelle
Press, Columbus Ohio, pp. 195–203.
OEPA, 1999. Ohio Environmental Protection Agency general permit
authorization to discharge wastewater from sewage treatment
systems designed to treat an average flow of 25,000 gallons per day
or less under the Nation Pollution Discharge Elimination System.
Columbus, OH.
OEPA, 2001. Ohio Environmental Protection Agency general permit
authorization to discharge wastewater from new and replacement
household sewage treatment systems under the national pollutant
discharge elimination system. Ohio Environmental Protection
Agency. Columbus, OH.
Richter, A.Y., Weaver, R.W., 2003. Treatment of domestic wastewater
by subsurface flow constructed wetlands filled with gravel and tire
chip media. Envron. Technol. 24, 1561–1568.
Stecher, M.C., Weaver, R.W., 2003. Effects of umbrella palms and
wastewater depth on wastewater treatment in a subsurface flow
constructed wetland. Environ. Technol. 24, 471–478.
Steer, D., Fraser, L., Boddy, J., Seibert, B., 2002. Efficiency of small
constructed wetlands for subsurface treatment of single-family
domestic effluent. Ecol. Eng. 18, 429–440.
Steer, D., Aseltyne, T., Fraser, L., 2003. Life-cycle economic model of
small treatment wetlands for domestic wastewater disposal. Ecol.
Econ. 44 (2–3), 359–369.
Thom, W.O., Gaddy, B.S., Ball, H.L., 1998. Long-term results of
residential constructed wetlands. In: On-site wastewater treatment,
Proc. 8th Annual National Sym. On Ind. and Small Community
Sewage Systems, ASAE. St. Joseph, MO.
Urbanc-Bercic, O., Bulc, T., 1995. Integrated constructed wetland for
small communities. Water Sci. Tech. 32 (3), 41–47.
USEPA, 1983. Methods for chemical analysis of water and wastes.
(EPA 600/4-79-020) rev. Mar. 83. USEPA, Washington, DC.
USEPA, 2000. Manual: Constructed wetlands treatment of municipal
wastewater. (EPA 625/R-99/010). USEPA, Washington, DC.
USEPA, 2001. Protection of the NationÕs Waters through effective
NPDES permits: A strategic plan. (EPA-833-R-01-001) USEPA
Washington, DC.
Watson, J.T., Choate, K.D., Steiner, G.R., 1990. Performance of
constructed wetland treatment systems at Benton, Hardin, and
Pembrook, Kentucky during the early vegetation establishment
phase. In: Cooper, P.F., Findlater, B.C. (Eds.), Constructed
wetlands in water pollution: Advances in water pollution control.
Pergamon Press, Great Britain, pp. 171–182.
Weaver, R.W., Lane, J.J., Johns, M.J., Lesikar, B.J., 2001. Uptake of
15N by macrophytes in subsurface-flow wetlands treating domestic
wastewater. Env. Tech. 22, 837–843.
White, K.D., Shirk, C.M., 1998. Performance and design recommendations for on-site wastewater treatment using constructed wetlands. In: On-site wastewater treatment, Proc. 8th Annual National
Sym. On Ind. and Small Community Sewage Systems, ASAE. St.
Joseph, MO.
Wood, A., 1995. Constructed wetlands in water pollution control:
Fundamentals to their understanding. Water. Sci. Technol. 32 (3),
21–22.