NITROGEN AND PHOSPHORUS REMOVAL IN SUBSTRATE-FREE ...

NITROGEN AND PHOSPHORUS REMOVAL IN SUBSTRATE-FREE PILOT CONSTRUCTED WETLANDS WITH HORIZONTAL SURFACE FLOW IN UGANDA JOSEPH KYAMBADDE1,2,∗ , FRANK KANSIIME1 and GUNNEL DALHAMMAR2 1

Makerere University Institute of Environment and Natural Resources, P.O. Box 7298, Kampala, Uganda; 2 Department of Biotechnology, Royal Institute of Technology (KTH), AlbaNova University Centre, S-106 91 Stockholm, Sweden (∗ author for correspondence, e-mail: [email protected]; Tel. +256-41-530135; Fax: +256-41-530134)

(Received 22 September 2004; accepted 10 March 2005)

Abstract. In constructed wetlands (CWs) with horizontal sub-surface flow, nutrient removal, especially phosphorus, is limited because the root biomass fills the pore spaces of the substrate (usually gravel), directing wastewater flow to deeper wetland media; plants are not regularly harvested; the litter formed by decomposing vegetation remains on the surface of the substrate and thus does not interact with the wastewater; and the substrate media often used provide only limited adsorption. Effective nutrient removal including rootzone oxidation, adsorption and plant uptake therefore requires sufficient interaction of wastewater with the treatment media. We assessed the feasibility of biological nutrient removal from wastewater using substrate-free CWs with horizontal flow, planted with two tropical macrophytes namely, Cyperus papyrus and Miscanthidium violaceum. The objectives were to evaluate the system treatment efficiency under semi-natural conditions, and to assess microbial and plant biomass contributions to nutrient removal in the CWs. Results showed high removal efficiencies for biochemical oxygen demand, ammonium-nitrogen (NH4 –N) and phosphorus (P) fractions in papyrus-based CWs (68.6–86.5%) compared to Miscanthidium (46.7–61.1%) and unplanted controls (31.6–54.3%). Ammonium oxidizing bacteria in CW root–mats (108 –109 cells/gram dry weight) and residual nitrite and nitrate concentrations in the water phase indicated active system nitrification. Papyrus showed higher biomass production and nutrient uptake, contributing 28.5% and 11.2%, respectively, of the total N and P removed by the system compared to 15% N and 9.3% P removed by Miscanthidium plants. Compared to literature values, nitrification, plant uptake and the overall system treatment efficiency were high, indicating a high potential of this system for biological nutrient removal from wastewaters in the tropics. Keywords: constructed wetlands, Cyperus papyrus, horizontal flow, miscanthidium violaceum, nitrification, nutrient uptake, plant biomass production, substrate-free, Uganda

1. Introduction Worldwide, wetland ecosystems are increasingly being used for the treatment and disposal of wastewater because they have been recognized as low-cost and effective treatment systems (Brix, 1994). Their ability to improve the water quality from inflow to outflow has been demonstrated (Hammer, 1989; Kadlec and Knight, 1996). Constructed wetland systems use a consortium of physical, chemical and biological Water, Air, and Soil Pollution (2005) 165: 37–59

C

Springer 2005

38

J. KYAMBADDE ET AL.

processes (such as filtration, sedimentation, adsorption, bio-conversion and uptake by micro-organisms and wetland macrophytes) to remove different pollutants (Ran et al., 2004). Studies have shown that the soil/litter compartment is the major long-term storage pool for phosphorus (P) in wetlands (Nichols, 1983; Verhoeven, 1986; Cooke, 1992; Hiley, 1995). Nitrogen removal on the other hand is effected through sediment accumulation, adsorption of ammonium onto the organic sediments (Howard-Williams, 1985), plant uptake and nitrification-denitrification processes. Different designs of constructed wetlands (CWs) have been developed such as vertical flow and horizontal sub-surface flow systems. Horizontal sub-surface flow systems, however, do not usually remove substantial amounts of P because (1) the root biomass fills the pore spaces of the substrate (gravel), directing wastewater flow to deeper wetland media (Breen and Chick, 1995), (2) the litter formed by decomposing vegetation remains on the surface of the substrate and thus does not interact with the wastewater, (3) the substrate media usually used does not contain sufficient concentrations of Ca2+ , Fe3+ , or Al3+ to actively adsorb P (Brix et al., 2001; Vymazal, 2004). Effective nutrient removal mechanisms including rootzone oxidation, adsorption and plant uptake therefore require adequate interaction of wastewater with the treatment media. Small-scale constructed wetlands for rural domestic wastewater treatment are a relatively new technology, and the physical, chemical and biological processes which facilitate treatment are still not well understood (Coleman et al., 2001). Several researchers (Reed et al., 1995; Kadlec and Knight, 1996) have discussed the progress made in our understanding of how these systems function. However, the inconsistency in treatment efficiencies suggests that more studies be carried out to optimize the system functioning. In particular, an understanding of the roles played by plants in these treatment systems is still lacking, and little research has compared different tropical plant species in promoting wastewater treatment under tropical environments. In Uganda, wastewater treatment by natural wetlands has been in use for several decades. On the other hand, the use of CWs is a relatively new technology with very few studies carried out to investigate its potential application for wastewater treatment. Presently, several municipalities treat municipal sewage up to the secondary level using stabilization pond systems because conventional advanced tertiary treatment is expensive, requiring large capital investments to build and operate (Reed et al., 1995). In order for resource-scarce developing countries to adopt wastewater treatment, the treatment technologies must be cost-effective and easy to adopt, require less energy input, and maintenance costs and be capable of meeting effluent discharge standards. This paper reports on the feasibility of biological nutrient removal from wastewater using substrate-free CWs with horizontal flow, planted with two tropical macrophytes namely, Cyperus papyrus and Miscanthidium violaceum. The objectives were: (a) to evaluate the system treatment efficiency under semi-natural conditions, (b) to assess microbial and plant biomass contributions to biological nutrient removal processes in the wetlands.

NUTRIENT REMOVAL IN SUBSTRATE-FREE CONSTRUCTED WETLANDS IN UGANDA

39

2. Methods and Materials 2.1. DESIGN

OF THE CONSTRUCTED WETLANDS

Dye tracer studies have shown that in horizontal sub-surface flow constructed wetlands the wastewater–rootzone contact is reduced due to root biomass that fills the pore spaces of the gravel and directs the flow to deeper wetland media (Breen and Chick, 1995). Therefore, the design used in this study was substrate-free to allow sufficient mixing of wastewater and optimal contact between wastewater, micro-organisms and macrophyte root systems. The beds and walls were lined with a 1mm polythene-plastic material to eliminate seepage and mixing of wastewater with underground water (dilution). The experimental design included three replicated treatments: unplanted control lines (C1 and C2), Miscanthidium-based treatment lines (M1 and M2) and papyrus-based treatment lines (P1 and P2). Each line is composed of four horizontal surface-flow constructed wetland (HSFCW) cells (HSFCW 1–4) arranged in series. HSFCW 2–4 for treatment lines M1 and M2 were planted with Miscanthidium violaceum whereas those for treatment lines P1 and P2 were planted with Cyperus papyrus. The respective length, width and depth dimensions were 3 × 2.5 × 0.2 m for HSFCW 1 and 3 × 2.5 × 0.35 m for HSFCW 2–4 giving a total surface area of 60 m2 for each replicated treatment. Using effluent quality data for the stabilization ponds (Kyambadde et al., 2004a), sizing of the wetland units was done using the basic relationship for plug flow reactor for the BOD5 loading (Equation 1), as recommended by Kickuth (1977). Ah = Q d (ln Co − ln Ci )/K BOD5 ,

(1)

where Ah is the surface area of the wetland bed (m2 ), KBOD5 the BOD5 area-based rate constant at 20 ◦ C, (m/per day), Qd the average flow rate of wastewater (m3 /day), Co the average BOD5 of the influent (mg/L), and Ci the average BOD5 of the effluent (mg/L). The KBOD5 value of 0.19 adopted in this study was that proposed by Kickuth, (1977). The hydraulic retention time (HRT) was established based on the following equation; HRT = hlw/q (Trepel and Palmeri, 2002),

(2)

where h is the effective depth of the HSFCW units, l and w are the length and width of HSFCW units respectively, and q is the average flow rate (m3 /day). 2.2. EXPERIMENTAL

SET UP

The constructed wetlands (Figure 1a) which were planted late January, 2004 were fed with effluent wastewater from stabilization ponds at Kampala, Uganda, under a continuous flow regime. All HSFCW cells for treatment lines C1 and C2 were

40

J. KYAMBADDE ET AL.

(a)

(b) Figure 1. (a) Arrangement of the treatment wetlands. Treatment lines C1 and C2 are unplanted controls; M1 and M2 are planted with Miscanthidium violaceum, while P1 and P2 are planted with Cyperus payrus. The shaded constructed wetland cells (HSFCW 2–4) are vegetated, (b) A schematic representation of a substrate-free CW cell with horizontal surface flow regime. 1 – Impermeable plastic liner, 2 – vegetation, 3 – root system, 4 – inflow and outflow drainage pipes.

used as controls and therefore no macrophytes were planted in them. HSFCW cells 2–4 for treatment lines M1 and M2 were planted with Miscanthidium violaceum while P1 and P2 were planted with Cyperus papyrus (Figure 1b) collected from a neighbouring wetland. For both macrophyte-based treatments, the first horizontal surface-flow constructed wetland cell (HSFCW 1) was not planted to allow for oxygenation to occur at the air-water interface due to air currents and wind. Clones of young plants were suspended in the water column by tying them to a network of wooden pegs before establishing thick-interlaced root mats that later supported the


41

macrophytes. The planting densities per square meter, which were based on plant weight, were 63 and 36 plants for Miscanthidium violaceum and Cyperus papyrus, respectively. To ensure a uniform wastewater flow to the wetland units, 24 mm gate valves were fitted to the flow distribution pipe. Flow rates were adjusted manually and checked regularly using a measuring cylinder and stopwatch. In this study, the hydraulic loading rate for each treatment averaged to 0.115 m/day giving a theoretical HRT of 2.71 days which was in the range reported for CWs in temperate environments (Kadlec and Knight, 1996). 2.3. WATER

QUALITY MONITORING

In order to allow the vegetation and bio-film to establish, sampling was started after three months of constructed wetland plant growth. Waste water entering and exiting from the CW units was monitored between April and August 2004 and average in situ measurements for temperature, pH, electrical conductivity (EC) and dissolved oxygen (DO) determined. Similar measurements were made along the length of the wetlands. Water samples were analyzed for NH4 –N, NO2 –N, NO3 –N, total nitrogen (TN), ortho-phosphate (o-PO4 –P), total phosphorus (TP), and biochemical oxygen demand (BOD). Analyses were conducted in accordance with standard methods for the examination of water and wastewater (APHA, 1995). 2.4. NITRIFICATION

ACTIVITY

Potential nitrification activity of microbial biomass was carried out on water, peat and macrophyte root samples from HSFCW3 and HSFCW4 cells over a period of six hours. Samples were taken once a week for two months (June and July, 2004) and average values used to determine the microbial nitrification potential. The test media used [composition: 1.25 mL trace elements solution/L (Schmidt and Belser, 1982), 0.59 g (NH4 )2 SO4 , 1.25 g NaHCO3 , 0.06 g KH2 PO4 , 0.3 g CaCl2 ·2H2 O, 0.2 g MgSO4 , 0.00625 g FeSO4 , 0.00625 g EDTA, and 1.06 g NaClO3 ] contained ammonium to establish the activity of ammonium-oxidizing microbial biomass under aerobic conditions. Sodium chlorate is used to inhibit the oxidation of nitrite to nitrate (Belser and Mays, 1982; Hynes and Knowles, 1983). Replicate samples of roots (10 g) in 250 mL Erlenmeyer flasks containing 100 mL of test media were incubated at room temperature in an orbital shaker (Gerhardt Thermoshake, Germany) running at 100 rpm. For water and peat, replicate 50 mL samples were added to 100 mL of test media. Ammonium oxidation was monitored by analyzing 30 µL samples for nitrite at 540 nm over time using a Merck Spectroquant test kit and a microplate reader (Tecan Sunrise, Ges.mb.H, Austria). Nitrite concentrations were calculated from absorbance values and standard nitrite (5 mg/L) using BeerLambert’s law. Ammonium oxidation rates were determined as slopes from plots of NO2 –N concentration versus time.

42

J. KYAMBADDE ET AL.

2.5. PLANT

BIOMASS PRODUCTION AND NUTRIENT STORAGE

In order to determine nutrient storage in the different plant parts and to estimate the amount of nutrients that could be removed by harvesting plants, biomass was harvested during the 6th and 7th months (n = 3) after planting. Whole plant samples were harvested from 0.25 m2 sections of each planted CW cell. Plants were separated into roots/rhizomes, culms plus sheathing scales, umbels for papyrus; leaves, stalk and roots/rhizome for Miscanthidium violaceum. Each sample was cut into small pieces, mixed, sun-dried for one week to reduce the moisture content and later oven-dried to constant weight at 103 ◦ C for dry weight determination. The biomass per m2 was calculated as an average of the values obtained on three sampling occasions for each macrophyte species. The concentration of N in plant tissues was determined using the Kjeldahl method, while P was measured by the molybdateascorbic acid method according to Novozamsky et al. (1983). The average nutrient content of above ground biomass was calculated as: X = [(D1 × C1 /Y ) + (D2 × C2 /Y )]/2

(3)

where X is the average nutrient content (g DW/m2 ), D1 the average dry weight (DW) of plant part 1 (g DW/m2 ), D2 the average DW of plant part 2 (g DW/m2 ), C1 the nutrient concentration (g DW/m2 ) of plant part 1, C2 the nutrient content of plant part 2 (g DW/m2 ), and Y the total DW of above ground biomass (g DW/m2 ). 2.6. STATISTICAL

ANALYSES

Statistical analysis was performed with the package MINITAB Release 13.31 for Windows and included analysis of variance (ANOVA), Bartlett’s and Levine’s test for homogeneity of variance and normality, and Tukey’s multiple comparisons for differences between means.

3. Results 3.1. PLANT

BIOMASS PRODUCTION AND NUTRIENT STORAGE

The above and below ground biomass values for papyrus and Miscanthidium in the constructed wetland recorded 6 months after planting are presented in Figure 2. Papyrus-based CWs showed higher aerial and below ground biomass productivity than Miscanthidium ( p = 0.007 and 0.041, respectively). However, one-way ANOVA did not detect significant differences between the above and below ground biomass of papyrus ( p = 0.09) despite papyrus having a two-fold lower below ground than the aerial biomass. Similarly, the above ground and below ground biomass of Miscanthidium was not significantly different ( p = 0.77).


43

Figure 2. Above ground (AG) and below ground (BG) biomass productivity and nutrient (N & P) content of Cyperus papyrus and Miscanthidium violaceum (n = 3). Bars indicate standard error of the mean.

As shown in Figure 2, nutrient assimilation and storage properties varied between the two macrophyte species. One-way ANOVA detected significantly higher N in above ground biomass of papyrus per unit area compared to Miscanthidium ( p = 0.033). However, the differences in concentrations of P in the aerial standing stocks of the two plant species were not statistically significant ( p = 0.119). In addition, the concentrations of N found in the aerial parts of papyrus (Figures 3a and 3b) were significantly higher than P ( p = 0.004). Furthermore, ANOVA showed that the storage of N in all the three papyrus plant parts varied significantly ( p = 0.000). Significantly higher N concentrations ( p = 0.000) were sequestered in papyrus umbels compared to culms and roots/rhizomes portions (Figure 3a). However, the culms and roots/rhizomes portions of papyrus did not show significant differences in their N content ( p = 0.691). Similarly, no significant differences ( p > 0.05) were detected in P content of papyrus plant tissues (Figure 3b) despite lower P levels in papyrus culms (4.7 ± 0.2 g P/kg DW, Figure 3b). As in papyrus, the concentration of N was significantly higher ( p = 0.000) than P in the aerial parts of Miscanthidium (Figure 4a and 4b). However, unlike in papyrus, the differences in N content of the different parts of Miscanthidium were not significant ( p > 0.05). Furthermore, a comparison of N and P content of papyrus culms and roots/rhizomes portions with that of Miscanthidium stalks and roots/rhizomes parts, respectively (Figures 3 and 4) did not detect significant differences ( p > 0.05 for both nutrients). Rather, more N was bound into papyrus umbels than Miscanthidium leaves ( p = 0.000). Whereas a significant increment in N content of papyrus umbels was noticed ( p = 0.000) in comparison to reference samples taken at the time of planting (Figure 5a), significant reductions in the N content of papyrus culms ( p = 0.001) and root/rhizome portions ( p = 0.007)

44

J. KYAMBADDE ET AL.

Figure 3. Concentrations of TN (a) and TP (b) in reference (t = 0) and CWs Cyperus papyrus organs. Bars indicate standard error of the mean (n = 3).

were also recorded in this study. However, no significant increments in P content of papyrus storage parts were detected ( p > 0.05 for all plant parts). On the other hand, one-way ANOVA detected a significant increment in N content of Miscanthidium leaves ( p = 0.001) compared to its stalks ( p = 0.17) and roots/rhizomes portions ( p = 0.687, Figure 5b). Similarly, the increment in P content of all Miscathidium plant parts was highly significant ( p < 0.02). 3.2. N ITRIFICATION

ACTIVITY

Potential nitrification activities of the different matrices of the CWs are presented in Figures 6a–c. Comparatively, the differences in potential nitrification activities between peat (0.44 ± 0.18, 0.21 ± 0.01 and 0.32 ± 0.01 mg NO2 –N/L/h) and the water phase (0.029 ± 0.09, 0.013 ± 0.00 and 0.027 ± 0.002 mg NO2 –N/L/h) of unplanted controls, papyrus and Miscanthidium–based treatments, respectively were highly significant ( p ≤ 0.001). Multiple comparisons detected significantly higher activity in peat sample from unplanted controls than CWs planted with papyrus ( p = 0.000) and Miscanthidium ( p = 0.009). In contrast, the differences


45

Figure 4. Concentrations of TN (a) and TP (b) in reference (t = 0) and CW Miscanthidium violaceum organs. Bars indicate standard error of the mean (n = 3).

in nitrification activities between the controls and Miscanthidium-based CWs were not significant ( p > 0.05). However, lower nitrification activity was detected in the water phase of papyrus-based treatment units compared to that of unplanted controls ( p = 0.003) and Miscanthidium-based treatments ( p = 0.001). In addition, higher ammonium oxidation rate ( p = 0.000) was associated with the Miscanthidium root mat (2.96 ± 0.2 mg NO2 –N/L/h) compared to that of papyrus (0.20 ± 0.02 mg NO2 –N/L/h). 3.3. PHYSICAL ,

CHEMICAL AND BIOCHEMICAL PARAMETERS

Results of the of the measured physico-chemical and biochemical variables for the unplanted controls, Miscanthidium and papyrus-based treatments are presented in Figure 7 and Table I. Water balance data (Figure 7) was used to estimate the

46

J. KYAMBADDE ET AL.

Figure 5. Percentage increment/reduction in TN and TP concentrations of Cyperus papyrus (a) and Miscanthidium violaceum (b) plant organs after six months of planting.

HRT for each treatment. The data shows that effluent wastewater loading was significantly lower than influent loading in all treatments ( p ≤ 0.01). The changes in influent and effluent loading rates (IE) were recorded as 5.5 ± 0.7 × 10−3 , 18.8±0.3×10−3 , and 24.5±0.6×10−3 m/day for the controls, Miscanthidium and papyrus-based treatments, respectively. The IE values were significantly larger for papyrus treatments than Miscanthidium and unplanted control treatments ( p = 0.000). Using water balance data and bearing in mind water losses, the HRT (mean ± standard deviation) were calculated as 2.85 ± 0.17, 3.02 ± 0.13 and 3.44 ± 0.11 days for the controls, Miscanthidium and papyrus-based constructed wetlands, respectively. Furthermore, the differences in water pH and electrical conductivity between the vegetated treatments were highly significant ( p ≤ 0.004). Papyrusbased treatments exhibited significantly lower water temperature ( p = 0.038), conductivity ( p = 0.016), dissolved oxygen ( p = 0.009), and pH ( p = 0.000) than the unplanted controls. Similarly, pH was significantly lower in treatments


47

planted with Miscanthidium in comparison to the unplanted controls ( p = 0.000). However, the differences in water temperature, electrical conductivity and dissolved oxygen (DO) between the unplanted controls and Miscanthidium-based treatments were not significant ( p ≥ 0.06). In addition, the levels of DO in both vegetated treatments were not statistically different ( p = 0.16) despite higher DO levels in Miscanthidium-based treatments.

Figure 6. Potential nitrification activities of water, peat and/or root-mat samples from CWs planted with Miscanthidium violaceum (a) and Cyperus papyrus (b) and unplanted controls (C). (Continued on next page)

48

J. KYAMBADDE ET AL.

Figure 6. (Continued)

Figure 7. Changes in inflow and outflow values of the physical parameters determined for the constructed wetlands (n = 12). Bars indicate standard error of the mean.

To account for the effects of evapo-transpiration, influent and effluent mass loading rates were calculated from concentrations values. The removal efficiencies were then calculated from corresponding mass removal rates (Table I). Mass loading data showed a significant build-up of NO2 –N in controls ( p = 0.001) and treatment M ( p = 0.007). However, the differences between the influent and effluent loads of NO2 –N in systems planted with papyrus were not significant ( p > 0.05). In addition, significantly higher loads of NO2 –N were detected in the control effluent

10.7 ± 0.5 0.52 ± 0.09 1.61 ± 0.61 14.9 ± 0.7 10.9 ± 0.7 13.3 ± 0.7 232 ± 73

19.1 ± 0.6 0.04 ± 0.01 0.41 ± 0.1 31.0 ± 1.4 19.0 ± 0.5 23.8 ± 1.5 320 ± 29

NH4 –N NO2 –N NO3 –N TN o-PO4 –P TRP BOD

C

8.4 ± 0.3 2.2 ± 0.1 1.2 ± 0.05 0.48 ± 0.09a 0.0046 ± 0.001 0.060 ± 0.01 1.20 ± 0.52a 0.047 ± 0.01 0.180 ± 0.07 16.1 ± 1.5 3.57 ± 0.16 1.63 ± 0.08 8.1 ± 0.8 2.19 ± 0.06 1.20 ± 0.08 10.5 ± 0.9 2.74 ± 0.17 1.46 ± 0.08 88 ± 38 37.0 ± 3.3 25.3 ± 8.0

9.0 ± 0.5 0.13 ± 0.04a 0.78 ± 0.53a 17.6 ± 1.7 8.4 ± 0.7 9.7 ± 1.4 153 ± 19

1.0 ± 0.1 0.055 ± 0.001a 0.133 ± 0.03a 1.94 ± 0.11 0.99 ± 0.13 1.28 ± 0.12 11.7 ± 3.1

1.16 ± 0.08 0.013 ± 0.005a 0.073 ± 0.04a 2.18 ± 0.09 1.09 ± 0.07 1.28 ± 0.13 19.7 ± 3.1

1.51 ± 0.07 0.001 ± 0.000 0.007 ± 0.001a 2.64 ± 0.11 1.61 ± 0.04 1.91 ± 0.09 32.0 ± 2.4

45.5 91.7a 73.9a 54.3 45.2 46.7 31.6

52.7 72.2a 60.8a 61.1 49.8 46.7 53.2

68.6 20.0 13.0a 73.9 73.5 69.7 86.5

D D D D

E E E E

C C C C

E E E E

C E C

B B B B

B E B

A A A A

A D A

F

F

G

% Removal efficiency Significance

There was an increment in the measured parameter. Parameters with the same upper case letters are not significantly different between treatments, while those with different upper case letters are significantly different between treatments ( p ≤ 0.05).

a

10.1 ± 0.4 0.17 ± 0.04 1.19 ± 0.63 13.4 ± 0.5 10.6 ± 0.6 14.1 ± 0.8 167 ± 34

19.1 ± 0.6 0.04 ± 0.01 0.41 ± 0.1 31.0 ± 1.4 19.0 ± 0.5 23.8 ± 1.5 320 ± 29

NH4 –N NO2 –N NO3 –N TN o-PO4 –P TP BOD

M

2.2 ± 0.1 1.04 ± 0.04 0.005 ± 0.001 0.018 ± 0.004 0.047 ± 0.01 0.120 ± 0.07 3.57 ± 0.16 1.39 ± 0.05 2.19 ± 0.06 1.10 ± 0.06 2.74 ± 0.17 1.46 ± 0.08 37.0 ± 3.3 17.3 ± 3.5

2.2 ± 0.1 0.69 ± 0.04 0.005 ± 0.001 0.004 ± 0.001 0.047 ± 0.01 0.054 ± 0.03 3.57 ± 0.16 0.93 ± 0.05 2.19 ± 0.06 0.58 ± 0.08 2.74 ± 0.17 0.83 ± 0.08 37.0 ± 3.3 5.0 ± 1.6

11.5 ± 0.2 0.00 ± 0.01 0.19 ± 0.02a 20.7 ± 1.5 12.6 ± 0.8 14.7 ± 1.1 266 ± 43

7.6 ± 0.4 0.04 ± 0.01 0.60 ± 0.37 10.3 ± 0.5 6.4 ± 0.9 9.1 ± 0.9 54 ± 18

19.1 ± 0.6 0.04 ± 0.01 0.41 ± 0.10 31.0 ± 1.4 19.0 ± 0.5 23.8 ± 1.5 320 ± 29

NH4 –N NO2 –N NO3 –N TN o-PO4 –P TP BOD

P

Outflow loading Removal

Inflow loading

Inflow conc. Outflow conc. Removal

Treatment Variable

TABLE I Mean ± standard error values of the variables determined in pilot constructed wetlands (n = 12). Concentrations are in mg/L, loadings in g/m2 /day NUTRIENT REMOVAL IN SUBSTRATE-FREE CONSTRUCTED WETLANDS IN UGANDA

49

50

J. KYAMBADDE ET AL.

than Miscanthidium ( p = 0.006) and papyrus constructed wetland ( p = 0.001). By comparing the two vegetated treatment systems, NO2 –N was significantly higher ( p = 0.006) in Miscanthidium than papyrus-based CWs. Results further showed that despite higher effluent NO3 –N levels in the effluent from the controls and Miscanthidium-based treatments, no significant differences in effluent NO3 –N were detected between two treatments ( p > 0.05). As shown in Table I, reductions in NH4 –N, TN, o-PO4 –P, TP, and BOD, respectively, were achieved in all treatments. The differences in the longitudinal removal rates of NH4 –N, TN, o-PO4 –P and TP between the three treatments were highly significant ( p ≤ 0.001) except for BOD ( p = 0.06). Multiple comparisons detected significantly higher NH4 –N, TN, o-PO4 –P, TP, and BOD removal rates in papyrus ( p ≤ 0.05) than Miscanthidium ( p > 0.1) CWs in comparison to unplanted controls. However, no significant variations in removal rates were detected between unplanted controls and Miscanthidium-based treatments ( p > 0.05 for NH4 –N, TN, o-PO4 –P, TP, and BOD).

4. Discussions Good performance of wetlands for wastewater treatment depends on the growth potential and ability of macrophytes to develop sufficient root systems for microbial attachment and material transformations (Eriksson and Weisner, 1999; Körner, 1999; IWA, 2000; Heidenwang et al., 2001), and to translate nutrients into plant biomass that can be subsequently harvested for nutrient removal (Brix, 1997; IWA, 2000). In the present study, the differences in biomass productivity between the two macrophyte species depicted important consequences for the degradation of wastewater components and uptake of nutrients. The higher below-ground biomass productivity of papyrus favoured more uptake and translocation of nutrients to growing aerial parts compared to Miscanthidium. This resulted in sequestration of more nutrients into plant tissues and hence higher aerial biomass productivity. Moreover, more N was found bound into papyrus umbels as opposed to Miscanthidium leaves (Figures 3a and 4a). Even though the N and P contents of papyrus culms and roots/rhizomes portions were statistically not different from that of Miscanthidium stalks and roots, respectively, calculations from area-based biomass productivity demonstrated that more nutrients were bound into papyrus than Miscanthidium tissues. This further shows that papyrus removed more nutrients per unit area of the constructed wetlands and therefore explains the differences in biomass productivity. Investigations have shown that wetland plants transfer photosynthetic oxygen, at different rates, to the rhizosphere where it influences root growth (Brix et al., 1992; Hiley, 1995). In this study, the two macrophyte species were observed to develop shorter roots when grown on nutrient-rich wastewater compared to pilot-scale


51

container experiments (Kyambadde et al., 2004b). In addition, Miscanthidium grew more slowly than papyrus as observed in other studies (Kipkemboi et al., 2002; Kyambadde et al., 2004b). Therefore the high BOD loading of the influent wastewater exerted an oxygen demand at varying degrees to plant roots thus explaining the observed differences in below ground biomass productivity recorded in this study. Albeit the average below-ground biomass value obtained for papyrus was two-fold lower than its aerial counterpart, it was in the range reported for floating papyrus swamps (Kipkemboi et al., 2002). Moreover, high rates of aerial primary production have been reported for papyrus growing under natural environments (6607 g DW/m2 /year; Muthuri et al., 1989) which is attributed to its high photosynthetic activity resulting from its characteristic C4 photosynthesis (Jones, 1986). Contrary to temperate and cold region wetlands where maximal plant biomass and maximal nutrient concentration do not occur at the same time of the year due to seasonal translocation of nutrients to different plant parts (Boyd, 1970; Vymazal, 2004), the tropical environment of this study indicates that N and P in plant tissues would increase bearing in mind that plants had not yet attained their maximum growth and that both plants grow throughout the year. Moreover, more N than P was bound into the aerial biomass of both plants species which was in concordance with earlier investigations (Kyambadde et al., 2004). In papyrus tissues, P was lowest in the culms, a trend similar to literature findings on samples from natural swamps receiving domestic wastewater (Muthuri and Jones, 1997; Kansiime et al., 2003). However, the concentrations of N in papyrus umbels and roots/rhizomes parts reported in this study were lower than values reported by Chale (1987) but in the range obtained by Muthuri and Jones (1997). Besides, the concentration of N in papyrus culms was higher than literature values (Muthuri and Jones, 1997; Kansiime et al., 2003). Thus, the high nutrient content of the aerial biomass of both plant species particularly papyrus is an indication of active translocation and storage of nutrients to sites where they are needed for primary growth (e.g synthesis of amino acids and enzymes) and thus presents a fairly good potential for biological nutrient removal through plant harvesting. Compared to literature values reported for Phragmites [2.0–2.6 g/m2 , Haberl and Perfler (1990); 3.8 g/m2 , Obarska-Pompkowiak (1999)] and Phalaris arundinacea [1.5 g P/m2 , Vymazal, 1999)] growing in horizontal sub-surface flow beds, the levels of P in standing plant stocks obtained in this study are 2–5 times higher indicating better nutrient uptake and storage performance of the two macrophytes in our CWs. It is interesting to note that both macrophyte species did not show significant differences in P assimilation and storage properties but differed in their abilities to store nitrogen. This probably indicates that N uptake is a critical factor affecting productivity of both macrophytes in wetland systems and influences P storage by these plants, but this needs further investigations. For effective removal of nutrients from wetland systems and to avoid nutrient recycling when plants die (Howard-Williams, 1985; Vymazal, 2004), periodic harvesting of plant biomass from systems with high biomass productivity is not only

52

J. KYAMBADDE ET AL.

desirable but a requirement. Using the area under papyrus vegetation (22.5 m2 ) with average above-ground and below ground nutrient content of 45.3 ± 2.4 and 14.9 ± 1.9 g DW/m2 , respectively, for N and 10.6 ± 2.2 and 6.6 ± 0.4 g DW/m2 respectively for P (Figure 2), mass balance calculations showed that plant uptake and storage of nutrients contributed 28.5% N and 11.2% P of the total nitrogen and phosphorus removed in papyrus-based treatments. Considering an average above-ground biomass of 2219 g DW/m2 , the total above ground papyrus biomass would be 49568 g. Using the average papyrus nutrient content of 13.7 g N/kg DW and 2.5 g P/kg DW obtained in this study, harvesting of the above-ground papyrus biomass would remove 0.68 kg N and 0.12 kg P every six month period. Interestingly, Okurut, (2001) reported plant uptake and storage contributions of 33% of the total P removed from CWs planted with Cyperus papyrus as a floating rooted mat in Uganda. This value is higher than 11.2% reported in this study due to (1) the high P content of our influent wastewater (19.1 ± 0.6 mg o-PO4 –P/L, Table I), compared to 3.7 ± 0.8 mg o-PO4 –P/L for Okurut (2001), (2) differences in treatment designs (a depth of 0.35 m used in this study compared to 1 m for Okurut, 2001), and (3) hydraulic loading regimes (theoretical HRTs were set to 2.71 days under continuous flow regime contrary to intermittent feeding and theoretical HRTs of 3–12 days by Okurut, 2001). However, when the same area (22.5 m2 ) covered by Miscanthidium with an average above ground and below-ground nutrient content of 22.1 ± 1.3 and 5.3 ± 0.5 g DW/m2 , respectively, for N and 6.8±0.7 and 2.8±0.3 g DW/m2 , respectively, for P are considered, plant uptake and storage contributed 15.7% N and 9.3% P of the total nitrogen and phosphorus removed by treatment line M. With an average above-ground biomass of 442 g DW/m2 for Miscanthidium, the total biomass above ground is calculated to be 9936 g DW/m2 . Therefore using average nutrient content of its above-ground biomass (7.4 g N/kg DW; 2.6 g P/kg DW), harvesting the above ground biomass would remove only 0.074 kg N and 0.026 kg P every sixth month period indicating a higher potential for papyrus. In treatment wetlands, surfaces on which nitrifying micro-organisms attach include litter, suspended particles, macrophytes and algae all of which interact with the flowing wastewater (Bastviken et al., 2003). The control and Miscanthidium-based treatment wetlands were often characterized by algae growth which maintained higher populations of nitrifiers in suspension relative to papyrus CWs where the shading effect limited the growth of algae. In addition, nitrifiers in the water column of papyrus CWs were reduced through attachment to plant roots, litter, and settling suspended particles (Brix, 1997; Körner, 1999; Bastviken et al., 2003; Kyambadde et al., 2004b). This observation therefore explains the lower nitrification activity of the water phase in papyrus CWs. The higher activity detected in peat compared to the water column in all treatments was due to attachment and sedimentation of nitrifiers with suspended particles and/or plant litter as explained above. Moreover, nitrifying bacteria have an obligate requirement for oxygen and inhibition occurs under anoxic conditions (Stensel


53

and Barnard, 1992). Therefore, in addition settlement with particulate matter, the relatively high DO (above 0.5 mg/L) in Miscanthidium and unplanted control treatments supported their metabolic requirements (Cloete and Muyima, 1997), thus further explaining the high nitrification activity detected in peat. Wetland plants are reported to have different abilities to maintain an oxygen supply to their roots in order to create a locally aerobic environment (Armstrong and Armstrong, 1988; Brix et al., 1992; Brix, 1994). The oxygen released is harnessed by both heterotrophic and autotrophic bacteria for aerobic organic matter degradation and nitrification (Brix, 1997). However, nitrifiers have a lower affinity for oxygen than aerobic heterotrophs (Bitton, 1994) and the two macrophyte species have different growth and root development properties (Azza et al., 2000; Kyambadde et al., 2004b), resulting in different abilities to maintain an oxygen supply to their roots. Using specific-activity values for various species of Nitrosomonas (0.023 pmol of NO2 –N produced/cell/h; Belser and Mays, 1982) and the potential nitrification activities measured for the different root mat phases, calculations show that papyrus root mats harbored 1.7 × 108 ± 6.3 × 107 cells/g DW compared to 2.8 × 109 ± 1.4 × 108 cells/g DW thriving in the Miscanthidium root mat. Furthermore, by considering the average below ground biomass of papyrus (1227 ± 147 g DW/m2 ) and Miscanthidium (425 ± 190 g DW/m2 ; Figure 2), nitrifying bacteria in the papyrus root mat (2.1 × 1011 ± 9.3 × 109 cells/m2 ) appeared to be significantly lower ( p = 0.000) than those in the root mat of Miscanthidium (1.2 × 1012 ± 2.7 × 1010 cells/m2 ). This implies that the high biomass productivity (see Figure 2) and photosynthetic activity of papyrus accompanied by its larger root surface area (Kyambadde et al., 2004b) provided more oxygen and attachment sites which were conducive for the proliferation of heterotrophic bacteria in the papyrus root mat than that of Miscanthidium. In addition, the dense vegetation cover of papyrus limited atmospheric aeration and oxygen production by algal photosynthesis. Also the extensively interlaced but permeable root mat of papyrus effectively retained suspended organic particles which provided sufficient substrates for the proliferation of heterotrophic bacteria (Azza et al., 2000). In fact, the BOD removal rate was much higher in papyrus-based treatment line P than Miscanthidium-based treatment line M (Table I). All these observations therefore indicate that the papyrus root mat experienced a stronger heterotrophic competition for the little available DO with autotrophic nitrifying bacteria, and thus explain the discrepancy in nitrification potential of the two macrophyte root mats. The nitrification activities we report in this study are higher than values we obtained in earlier pilot-scale container investigations (Kyambadde et al., 2004b) due to the scale of the system, and therefore represented an important component of biological nitrogen removal in our CWs treatment system. Water balance studies showed that in all treatments, differences in flows existed between the wastewater entering and exiting the CWs (IE) due to evapotranspiration and evaporation effects bearing in mind the tropical environment of this study. The higher water losses and hence HRTs in the vegetated treatments

54

J. KYAMBADDE ET AL.

particularly in papyrus-based treatments are explained by the high net biomass productivity accompanied by a higher transpiration rate. Similar observations are reported for vegetated subsurface flow systems treating domestic wastewater in Tanzania (Kaseva, 2004). Furthermore, the higher aerial biomass productivity of papyrus resulted in a higher transpiration rate and shading effect which together led to lower water temperatures (Chale, 1985). Water pH increased in Miscanthidium and unplanted control treatments due to the less-dense vegetation cover and absence of macrophytes, respectively, thereby favouring atmospheric aeration, and algal growth and photosynthesis. During oxygenic algal photosynthesis, the bicarbonate (HCO3 − ) ions present in the wastewater due to carbon dioxide (CO2 ) introduction through atmospheric aeration are consumed (Bitton, 1994). Consequently, hydroxyl (OH− ) ions which have a stronger basic power than the resultant hydrogen ions (H+ ) are released into the system thereby causing a rise in water pH (IWA, 2000; DeBusk and DeBusk, 2001) which explains the higher pH values in controls and treatment line M. The decrease in pH as wastewater flows through the papyrus CWs however, is explained by CO2 production from decomposing plant litter and other wastewater components trapped in the root mat (Chale, 1985; Verhoeven, 1986), and nitrification of ammonia (Bitton, 1994; IWA, 2000; Gerardi, 2002). During autotrophic nitrification, the carbon requirements are met by CO2 , bicarbonate, or carbonate present in the wastewater (Bitton, 1994). Besides, the oxidation of ammonia to nitrite is an acidic process that generates hydrogen ions (Bitton, 1994; IWA, 2000; Gerardi, 2002) which, in poorly buffered wastewater as that of papyrus CWs where surface aeration and algal photosynthesis were minimal, lower the water pH (Bitton, 1994; IWA, 2000; Gerardi, 2002). This therefore accounts for the lower pH recorded in papyrus-based treatments. Electrical conductivity generally reduced in all treatments with higher reductions observed in papyrus-based treatments as earlier reported for pilot-container experiments (Kyambadde et al., 2004b). The decrease in conductivity despite significant water losses is explained by uptake of micro and macro elements and ions by plants and bacteria, and their removal through adsorption to plant roots, litter and settleable suspended particles (Bitton, 1994; IWA, 2000; DeBusk and DeBusk, 2001). Even though the wastewater entering the CWs was practically deoxygenated, the wastewater exiting all the three treatments was characterized by higher DO levels especially wastewater exiting the unplanted controls (Table I). The higher effluent DO levels in unplanted control and Miscanthidium-based CWs relative to papyrusbased treatments are attributed to atmospheric aeration and oxygen release during algal photosynthesis. Similar trends between unplanted and vegetated CWs have been reported else where (Okurut et al., 1999). Albeit DO concentrations below 1.5 mg/L are reported to limit nitrification (Gerardi, 2002), effective nitrification has been reported in systems with DO as low as 0.5 mg/L (Cloete and Muyima, 1997). Additionally, nitrification effectively occurs in systems with pH and temperature ranges of 6.5–8.6 and 5–30 ◦ C, respectively (Schmidt and Belser, 1982; Kadlec and Knight, 1996; Cloete and Muyima, 1997;


55

Grundtz et al., 1998; Im et al., 2001). In this study, the rise in NO2 –N and NO3 – N loading in the effluent wastewater compared to the influent wastewater was a manifestation of active system nitrification. In treatment wetlands, nitrification occurs in aerobic regions of the water column, soil-water interface, and root zone (Reddy and D’Angelo 1997). The oxygen required for the nitrification process is supplied by diffusion from the atmosphere and release by macrophyte roots (Armstrong and Armstrong, 1988; Brix et al., 1992; Brix, 1994, 1997). In addition, nitrifying bacteria have less affinity for oxygen than aerobic heterotrophic bacteria (Bitton, 1994) and as results showed, the activity and numbers of nitrifying bacteria were lower in the papyrus root mat compared to that of Miscanthidium. Therefore, the lower effluent loading of NO2 –N and NO3 –N for vegetated treatments compared to unplanted controls is attributed to a higher competition for oxygen between aerobic heterotrophic and autotrophic nitrifying bacteria (Forcht and Verstraete, 1977; van Benthum et al., 1997; Nogueira et al., 2002), plant uptake of NO3 –N (IWA, 2000; DeBusk and DeBusk, 2001), and denitrification (IWA, 2000; DeBusk and DeBusk, 2001). Although denitrification takes place preferably under anoxic conditions, there is accumulating evidence that some bacteria also denitrify aerobically (Patureau et al., 1998; Plessis et al., 1998). Thus, the relatively low DO levels notably in papyrus CWs were favourable for denitrification. Interestingly, however, the effluent concentrations of NO2 –N and NO3 –N in all treatments were far below the Uganda regulatory discharge standards of 2 and 20 mg/L, respectively, for NO2 –N and NO3 –N (NEMA, 1999). Longitudinally, the NH4 –N, TN, o-PO4 –P, TP, and BOD removal rates obtained in vegetated systems were high and demonstrated a positive influence of macrophytes on nutrient removal processes, particularly papyrus. Papyrus-based treatments yielded effluent NH4 –N, TN and TP concentrations that were below the Uganda discharge limit of 10 mg/L (NEMA, 1999). However, with the exception of NH4 –N, the effluent TN, o-PO4 –P, TP, and BOD from Miscanthidium-based treatments were still above the recommended discharge standards (NEMA, 1999). In the present investigation, the NH4 –N and o-PO4 –P removal rates obtained in papyrusbased treatments (1.51 and 1.61 g/m2 /day, respectively) were much higher than 1.01 and 0.05 g/m2 /day, respectively, reported in literature for CWs in which Cyperus papyrus grown as a floating rooted mat treated pre-settled municipal wastewater (Okurut et al., 1999). As earlier explained, the discrepancy in removal rates is attributed to differences in treatment designs and hydraulic loading regimes which were higher in this study. In addition, the depth of their CW units was 1 m compared to 0.35 m used in this study, and their theoretical HRTs were set to 3–12 days contrary to 2.71 days used in this study. Furthermore, our design included an unplanted constructed wetland unit (HSFCW 1) within each macrophyte-based treatment to allow for oxygenation to occur at the air-water interface and improve on nitrification of the system; and was subjected to a continuous flow hydraulic regime whereas their CWs were intermittently loaded with wastewater either once or twice a week.

56

J. KYAMBADDE ET AL.

5. Conclusions Biological nutrient removal in horizontal sub-surface flow systems tends to be limited by the prevailing anaerobic – anoxic conditions with little NH4 –N removal via the nitrification–denitrification processes while phosphorus removal is reduced by the low mineral content of the substrate media. Plant uptake on the other hand is limited because the root biomass blocks the flow-paths of wastewater in the substrate media resulting in poor interaction of wastewater with plant root systems and the associated microorganisms. Apparently, our system design exhibited substantial nitrification capacity coupled with quite a large nitrifying biomass and fairly high nutrient uptake and plant biomass production. This was due to better contact of wastewater with plant litter, root systems and the associated micro-flora and fauna. The high removal efficiencies for BOD and nutrients in papyrus-based CWs indicate the system’s ability to treat high oxygen demanding and nutrient rich wastewater. The amount of phosphorus sequestered in above ground biomass of the two plants species was 2–5 times higher than literature values for sub-surface flow systems indicating a high system treatment potential. In addition, the effluent quality of papyrus-based CWs was within the national discharge limits of 10 mg/L for NH4 – N, TN and TP as well as 2 mg/L and 20 mg/L, respectively, for NO2 –N and NO3 – N. With plant uptake contributing 28.5% N and 11.2% P for papyrus, and 15.7% N and 9.3% P for Miscanthidium clearly shows that plant uptake, together with other removal mechanisms such as nitrification–denitrification, microbial uptake, adsorption to plant roots and sediments, played a crucial role in biological nutrient removal from these CWs.

Acknowledgments This work was financially supported by the Swedish International Development Cooperation Agency (Sida)/Department of Research Cooperation (SAREC) under the East African Regional Programme and Research Network for Biotechnology, Biosafety and Biotechnology Policy Development (BIO-EARN). The authors would like to thank Mr Joel Opio, Mr Bright Twesigye and Mr Alex Wacco for the help during construction and operation of the experimental system.

References APHA (American Public Health Association): 1995, Standard Methods for the Examination of Water and Wastewater, 19th ed., American Public Health Association Inc., Washington, DC, USA. Armstrong, J. and Armstrong, W.: 1988, ‘Phragmites australis – a preliminary study of soil-oxidising sites and internal gas transport pathways’, New Phytol. 108, 372–382.


57

Azza, N. G. T., Kansiime, F., Nalubega, M. and Denny, P.: 2000, ‘Differential permeability of papyrus and Miscanthidium root mat in Nakivubo swamp, Uganda’, Aquat. Bot. 67, 169– 178. Bastviken, K. S., Eriksson, P. G., Martins, I., Neto, J. M., Leonardson, L. and Tonderski, K.: 2003, ‘Potential nitrification and denitrification on different surfaces in a constructed treatment wetland’, J. Environ. Qual. 32, 2414–2420. Belser, L. W. and Mays, E. L.: 1982, ‘Use of nitrifier activity measurements to estimate the efficiency of viable nitrifier counts in soils and sediments’, Appl. Environ. Microbiol. 43, 945–948. Bitton, G.: 1994, Wastewater Microbiology, Wiley-Liss, New York. Boyd, C. E.: 1970, ‘Production, mineral accumulation and pigment concentrations in Typha Latifolia and Scirpus americanus’, Ecology 51, 285–290. Breen, P. F. and Chick, A. J.: 1995, ‘Rootzone dynamics in constructed wetlands receiving wastewater: A comparison of vertical and horizontal flow systems’, Water Sci. Technol. 32, 281–290. Brix, H.: 1994, ‘Use of constructed wetlands in water pollution control: Historical development, present status and future perspectives’, Water Sci. Technol. 30, 209–223. Brix, H.: 1997, ‘Do macrophytes play a role in constructed treatment wetlands?’, Water Sci. Technol. 35, 11–17. Brix, H., Arias, C. A. and del Bubba, M.: 2001, ‘Media selection for sustainable phosphorus removal in subsurface flow constructed wetlands’, Water Sci. Technol. 44, 47–54. Brix, H., Sorrell, B. K. and Orr, P. T.: 1992, ‘Internal pressurization and convective gas flow in some emergent freshwater macrophytes’, Limnol. Oceanogr. 37(7), 1420–1433. Chale, F. M. M.: 1985, ‘Effects of a Cyperus papyrus L. swamp on domestic wastewater’, Aquat. Bot. 23, 185–189. Chale, F. M. M.: 1987, ‘Plant biomass and nutrient levels of a tropical macrophyte (Cyperus papyrus L) receiving domestic wastewater’, Hydrobiol. Bull. 21(2), 167–170. Cloete, T. E. and Muyima, N. Y. O.: 1997, ‘Microbial community analysis: The key to the design of biological wastewater treatment systems’, in T.E. Cloete and N.Y.O. Muyima (eds), International Association on Water Quality (IAWQ) Scientific and Technical report No. 5, Cambridge University Press, Cambridge, pp. 10. Cooke, J. G.: 1992, ‘Phosphorus removal processes in a wetland after a decade of receiving a sewage effluent’, J. Environ. Qual. 21, 733–739. Coleman, J., Hench, K., Garbutt, K., Sexstone, A., Bissonnette, G. and Skousen, J.: 2001, ‘Treatment of domestic wastewater by three plant species in constructed wetlands’, Water, Air, and Soil Pollut. 128, 283–295. DeBusk, T. A. and DeBusk, W. F.: 2001, ‘Wetlands for water treatment’, in D.M. Kent (ed.), Wetlands Science and Technology, 2nd edn, CRC Press, Boca Raton, USA. Eriksson, P. G. and Weisner, S. E. B.: 1999, ‘An experimental study on effects of submersed macrophytes on nitrification and denitrification in ammonium-rich aquatic systems’, Limnol. Oceanogr. 44(8), 1993–1999. Focht, D. D. and Verstraete, W.: 1977, ‘Biochemical ecology of nitrification and denitrification’, Adv. Microb. Ecol. 1, 135–214. Gerardi, M. H.: 2002, ‘Nitrification and Denitrification in Activated Sludge Processes’, Wastewater Microbiology Series, John Wiley and Sons, Inc, New York. Grunditz, C., Gumaelius, L. and Dalhammar, D.: 1998, ‘Comparison of inhibition assays using nitrogen-removing bacteria: application to industrial wastewaters’, Water Res. 32, 2995–3000. Haberl, R. and Perfler, R.: 1990, ‘Seven years of research work and experience with wastewater treatment by a reed bed system’, in P.F. Cooper and B.C. Findlater (eds), Constructed Wetlands in Water Pollution Control, Pergamon Press, Oxford, UK, pp. 205–214. Hammer, D. A. (ed.): 1989, ‘Constructed Wetlands for Wastewater Treatment: Municipal, Industrial and Agricultural’, Lewis Publishers Inc., Chelsea, Michigan, pp. 831.

58

J. KYAMBADDE ET AL.

Heidenwang, I., Langheinrich, U. and Luderitz V.: 2001, ‘Self-purification in upland and lowland streams’, Acta Hydroch. Hydrobiol. 29, 22–33. Hiley, P. D.: 1995, ‘The reality of sewage treatment using wetlands’, Water Sci. Technol. 32(3), 329–338. Howard-Williams, C.: 1985, ‘Cycling and retention of nitrogen and phosphorus in wetlands: A theoretical and applied perspective’, Freshwater Biol. 15, 391–431. Hynes, R. K. and Knowles, R.: 1983, ‘Inhibition of chemoautotrophic nitrification by sodium chlorite: a reexamination’, Appl. Environ. Microbiol. 45, 1178–1182. Im, J. H., Woo, H. J., Choi, M. W., Hi, K. B. and Kim, C. W.: 2001, ‘Simultaneous organic and nitrogen removal from municipal landfill leachate using an anerobic–aerobic system’, Water Res. 35, 2403–2410. Jones, M. B.: 1986, ‘Wetlands’, in N.R. Baker and S.P. Long (eds), Photosynthesis in Contrasting Environments, vol. 7, Topics in Photosynthesis, Elsevier, Amsterdam, pp. 103–138. Kadlec, R. H. and Knight, R. L.: 1996, ‘Treatment wetlands’, Lewis–CRC Press, Boca Raton, FL, U.S.A, pp. 893. IWA.: 2000, ‘Constructed wetlands for pollution control: processes, performance, design and operation’, in. R.H. Kadlec, R.L. Knight, J. Vymazal, H. Brix, P. Cooper and R. Haberl (eds), IWA specialist group on the use of macrophytes in water pollution control, Scientific and Technical Report No. 8, IWA Publishing, London, UK: pp. 1–156. Kansiime, F., Nalubega, M., van Bruggen, J. J. A. and Denny, P.: 2003, ‘The effect of wastewater discharge on biomass production and nutrient content of Cyperus papyrus and Miscanthidium violaceum in the Nakivubo wetland, Kampala, Uganda’, Water Sci. Technol. 48(5), 233– 240. Kaseva, M. E.: 2004, ‘Performance of a sub-surface flow constructed wetland in polishing pretreatment wastewater–a tropical case study’, Water Res. 38(3), 681–687. Kickuth, R.: 1977, ‘Degradation and incorporation of nutrients from rural wastewaters by plant rhizosphere under limnic conditions’, in Proceedings of the International Conference on Utilization of Manure by Land Spreading, Commission of the European Community, EUR 5672e, London, U.K., pp. 335–343. Kipkemboi, J., Kansiime, F. and Denny, P.: 2002, ‘The response of Cyperus papyrus (L.) and Miscanthidium violaceum (K. Schum.) Robyns to eutrophication in natural wetlands of Lake Victoria, Uganda’, Afr. J. Aquat. Sci. 27(1), 11–20. Körner, S.: 1999, ‘Nitrifying and denitrifying bacteria in epiphytic communities of submerged macrophytes in a treated sewage channel’, Acta Hydrochim. Hydrobiol. 27(1), 27–31. Kyambadde, J., Kansiime, F. and Dalhammar, G.: 2004a, ‘Hydraulic loading, stability and water quality of Nakivubo wetland in Uganda’, Afr. J. Aquat. Sci. 29(2). Kyambadde, J., Kansiime, F., Gumaelius, L. and Dalhammar, G.: 2004b, ‘A comparative study of Cyperus papyrus and Miscanthidium violaceum-based constructed wetlands for wastewater treatment in a tropical climate’, Water Res. 38(2), 475–485. Muthuri, F. M. and Jones, M. B.: 1997, ‘Nutrient distribution in a papyrus swamp: Lake Naivasha, Kenya’, Aquat. Bot. 56, 35–50. Muthuri, F. M., Jones, M. B. and Imbamba, S. K.: 1989, ‘Primary productivity of papyrus (Cyperus papyrus) in a tropical swamp: Lake Naivasha, Kenya’, Biomass 18, 1–14. NEMA (National Environment Management Authority): 1999, ‘Environmental standards and preliminary environment impact assessment for water quality and discharge of effluent into water and on land in Uganda’, National Environment Management Authority– Ministry of Natural Resources–Government of the Republic of Uganda, Kampala, Uganda, pp. 89. Nichols, D. S.: 1983, ‘Capacity of natural wetlands to remove nutrients from wastewater’, J. Water Pollut. Control Fed. 55, 495–505.


59

Nogueira, R., Melo, L. F., Purkhold, U., Wuertz, S. and Wagner, M.: 2002, ‘Nitrifying and heterotrophic population dynamics in biofilm reactors: Effects of hydraulic retention time and the presence of organic carbon’, Water Res. 36, 469–481. Novozamsky, I., Houba, V. J. G., van Eck, R. and van Vark, W.: 1983, ‘A novel digestion technique for multi-element plant analysis’, Commun. Soil Sci. Plant Anal. 14(3), 239–248. Obarska-Pompkowiak, H.: 1999, ‘Nutrient cycling and retention in constructed wetland system in Darzlubie near Puck Bay, Southern Baltic Sea’, in J. Vymazal (ed.), Nutrient Cycling and Retention in Natural and Constructed Wetlands, Backhuys Publishers, Leiden, The Netherlands, pp. 41–48. Okurut, T. O.: 2001, ‘Plant growth and nutrient uptake in a tropical constructed wetland’, in J. Vymazal (ed), Transformations of Nutrients in Natural and Constructed Wetlands, Backhuys Publishers, Leiden, The Netherlands, pp. 451–462. Okurut, T. O., Rijs, G. B. J. and van Bruggen, J. J. A.: 1999, ‘Design and performance of experimental constructed wetlands in Uganda, planted with Cyperus papyrus and Phragmites mauritianus’, Water Sci. Technol. 40(3), 265–271. Patureau, D., Godon, J. J., Dabert, P., Bouchez, T., Bernet, N., Delgenes, J. P. and Moletta, R.: 1998, ‘Microvirgula aerodenitrificans gen. nov., sp., a new gram-negative bacterium exhibiting co-respiration of oxygen and nitrogen oxides up to oxygen saturated conditions’, Int. J. Syst. Bacteriol. 48, 775–782. Plessis, C. A. du., Kinney, K. A., Schroeder, E. D., Chang, D. P. Y. and Scow, K. M.: 1998, ‘Denitrification and nitric oxide reduction in an aerobic toluene-treating biofilter’, Biotechnol. Bioeng. 58, 408–415. Ran, N., Agami, M. and Oron, G.: 2004, ‘A pilot study of constructed wetlands using duckweed (Lemna gibba L.) for treatment of domestic primary effluent in Israel’, Water Res. 38(9), 2241– 2248. Reddy, K. R. and D’Angelo, E. M.: 1997, ‘Biogeochemical indicators to evaluate pollutant removal efficiency in constructed wetlands’, Water Sci. Technol. 35(5), 1–10. Reed, S. C., Crites, R. W. and Middlebrooks, E. J.: 1995, ‘Natural Systems for Wastewater Management and Treatment’, McGraw-Hill Inc., New York, pp. 308. Schmidt, E. L. and Belser, L. W.: 1982, ‘Nitrifying bacteria’, in A.L. Page, R.H. Miller and D.R. Keeney (eds), Methods of soil analysis, Part 2: chemical and microbiological properties, vol. 9, American Society of Agronomy Monograph, Madison, W.I., pp. 1027–1042. Stensel, H. D. and Barnard, J. L.: 1992, ‘Principles of biological nutrient removal’, in C.W. Randal, J.L. Barnard and H.D. Stensel (eds), Design and Retrofit of Wastewater Treatment Plants for Biological Nutrient Removal, Technomic Pub. Co. Inc., Lancaster, pp. 25–45. Trepel, M. and Palmeri, L.: 2002, ‘Quantifying nitrogen retention in surface flow wetlands for environmental planning at the landscape-scale’, Ecol. Eng. 19, 127–140. van Benthum, W. A. J., van Loosdrecht, M. C. M. and Heijnen, J. J.: 1997, ‘Control of heterotrophic layer formation on nitrifying biofilms in a biofilm airlift suspension reactor’, Biotechnol. Bioeng. 53, 397–405. Verhoeven, J. T. A.: 1986, ‘Nutrient dynamics in minerotrophic peat mires’, Aquat. Bot. 25, 117–167. Vymazal, J.: 1999, ‘Removal of phosphorus in constructed wetlands with horizontal sub-surface flow in the Czech Republic’, in J. Vymazal (ed), Nutrient Cycling and Retention in Natural and Constructed Wetlands, Backhuys Publishers, Leiden, The Netherlands, pp. 73–83. Vymazal, J.: 2004, ‘Removal of phosphorus in constructed wetlands with horizontal sub-surface flow in the Czech Republic’, Water, Air, Soil Pollut. 4, 657–670.