A set of experiments were conducted in Brazil in a pilot-scale waste stabilization pond (WSP) system (a four-maturation-pond series) treating an upflow anaerobic sludge blanket (UASB) reactor effluent. Over a year and a half the pond series was monitored under two flow rate conditions, hence also different hydraulic retention times and surface loading rates. On-site and laboratory trials were carried out to assess: (i) ammonia losses by volatilization using acrylic capture chambers placed at the surface of the ponds; (ii) organic nitrogen sedimentation rates using metal buckets placed at the bottom of the ponds for collecting settled particulate matter; (iii) nitrogen removal by algal uptake based on the nitrogen content of the suspended particulate matter in samples from the ponds' water column. In addition, nitrification and denitrification rates were measured in laboratory-based experiments using pond water and sediment samples. The pond system achieved high nitrogen removal (69% total nitrogen and 92% ammonia removal). The average total nitrogen removal rates varied from 10,098 to 3,849 g N/ha·d in the first and the last ponds, respectively, with the following fractions associated with the various removal pathways: (i) 23.5–45.6% sedimentation of organic nitrogen; (ii) 13.1–27.8% algal uptake; (iii) 1.2–3.1% ammonia volatilization; and (iv) 0.15–0.34% nitrification-denitrification.

Some of the most widely accepted models for predicting nitrogen and ammonia removal in waste stabilization ponds (WSPs) were developed on the basis that ammonia volatilization is the major pathway for nitrogen removal. This is the underpinning assumption of, for instance, the total nitrogen removal models proposed by Pano & Middlebrooks (1982) for ammonia removal, and by Middlebrooks et al. (1999), and Reed (1985), both of them for total nitrogen removal.

However, as pointed out by Camargo Valero & Mara (2010), these models are actually simple first-order equations developed for either plug-flow or complete-mix conditions; they reflect statistically significant relationships between nitrogen/ammonia removal and variables like pH, water temperature and hydraulic loading rate or hydraulic retention time (HRT), but do not provide any information on the nature of the nitrogen removal mechanisms. Indeed, high pH and temperature values (as well as in-pond favourable mixing conditions) favour free ammonia (NH3) stripping (Middlebrooks et al. 1999), which may as well be modelled as a first-order reaction in a completely mixed reactor. This is why the Pano & Middlebrooks model and slight variations of it (Silva et al. 1995; Soares et al. 1996; Bastos et al. 2007) still make good predictions of ammonia removal in WSPs, but they do not confirm ammonia volatilization as the main mechanism for nitrogen removal (Camargo Valero & Mara 2010). Actually, since pH fluctuates as a result of the algae-carbonate interactions in WSPs, high pH values may simply reflect intense algal activity, and thus algal uptake of ammonium (the NH4+ ionic form). In turn, high temperatures will also favour other biological removal mechanisms such as nitrification-denitrification. In effect, there have been more and more reports that ammonia volatilization makes only a small contribution to the overall performance of nitrogen removal in WSPs compared to other mechanisms such as biological uptake, nitrification-denitrification and sedimentation of dead biomass (Senzia et al. 2002; Assunção & von Sperling 2012; Camargo Valero & Mara 2007; Camargo Valero et al. 2010a, 2010b).

In this work, pilot-scale experiments were conducted in order to contribute to the understanding of the relative importance of the aforementioned mechanisms and pathways of nitrogen removal in WSPs in warm-climate regions.

This study was undertaken in an experimental pilot-scale WSP system (a four-maturation-pond series) fed by an upflow anaerobic sludge blanket (UASB) reactor (field scale) in Viçosa, Minas Gerais State, Brazil (latitude: 20°45′14″S, longitude: 42°52′53″W). All ponds had a surface area of 16.3 m2 and a length/width ratio equal to 2.0. Over the study period, the four ponds were operated under the following conditions: flow rate (Q) = 2 and 1.5 m3/d, pond depth (h) = 1 m; HRT = 7.05 and 9.4 d, hydraulic surface loading (HSL) = 0.123 and 0.093 m3/m2·d.

Based on the work of Camargo Valero & Mara (2007), to initially assess ammonia losses by volatilization, 12 trials of laboratory-controlled experiments were carried out using an apparatus comprising: (i) a capture chamber made of an acrylic 25 L container with a lid and an air inlet and outlet on opposite sides; the chamber was filled with 18 L of ammonium chloride solution adjusted to a concentration around 20 mg N/L, and pH adjusted to 10 with NaOH solution; (ii) an absorption system consisting of three flasks in series containing a 2% boric acid solution (Figure 1). An air flow of 3.5 L/min was pumped through the head space in the capture chamber, forcing the gases in there to bubble up through the absorption system. Water samples were collected in duplicate from the capture chamber at the beginning of the experiment and after 24 hours, along with samples from the three absorption flasks; the water and boric acid samples were processed for NH4+ (APHA 1998) in order to determine the amount of volatilized ammonia that was actually captured by the acid solution. Three extra 48 hour trials were undertaken in order to compare the recovery fraction with that of the 24 hour experiments. pH and temperature were continuously monitored using a data logger.

Figure 1

Experimental set-ups for measuring ammonia loss by volatilization: laboratory (left) and on-site (right) apparatus.

Figure 1

Experimental set-ups for measuring ammonia loss by volatilization: laboratory (left) and on-site (right) apparatus.

Close modal

Using an apparatus and experimental procedures similar to those employed in the laboratory tests, 29 on-site trials were carried out simultaneously in each of the four maturation ponds. An acrylic capture chamber, with air inlet and outlet on opposite sides, and two equally spaced plastic baffles in order to avoid dead zones in the air flow, were placed on the surface and at the center of each pond (Figure 1). At the end of each 24 hour trial, boric acid solution samples from the absorption system were collected and processed for NH4+. The results were expressed in terms of volatilization rates (kg N-NH3/ha·d), taking into account the capture chamber area and the recovery efficiency previously determined in the laboratory experiments. Temperature and pH were measured every minute at the water surface and next to the capture chamber using a data logger.

In addition, in order to assess the overall pond system performance, samples from raw sewage and from the UASB and pond effluents were collected and analysed for organic nitrogen, ammonia, nitrite, and nitrate (APHA 1998).

As in Camargo Valero & Mara (2007) and Assunção & von Sperling (2012), the field measurements of ammonia volatilization were compared to theoretical rates calculated from an equilibrium-based mass transfer equation (Equation (1)).
formula
(1)
where λNH3 is the ammonia volatilization rate (g N/ha·d); Kl, the mass transfer coefficient in the liquid phase (d−1); [NH3], the free ammonia concentration (g N/m3); V, the pond volume (m3); and A, the pond surface area (ha).
Kl values were calculated from Stratton's equation (Equation (2)) (Zimmo et al. 2003); the free ammonia concentration was obtained from the ammonia–ammonium equilibrium expression described elsewhere (Zimmo et al. 2003; Camargo Valero & Mara 2010).
formula
(2)
where d is the depth of the water column in the pond (m); T, the water temperature (°C).

Six field trials were carried out simultaneously in each pond in order to estimate the sedimentation rate of organic nitrogen. Settled particulate matter samples were collected in two 10 L metal buckets, placed at the bottom of each pond as shown in Figure 2. The buckets were taken out at the end of each trial (after time periods ranging from 58 to 90 days) and the collected samples were processed as described by Camargo Valero et al. (2010a) for solids, moisture and nitrogen content. These results were expressed in terms of sedimentation rate (kg N/ha·d).

Figure 2

On-site experimental set up for estimating organic nitrogen sedimentation rate: positioning of 10 L metal buckets (left) at the bottom of a pond (right).

Figure 2

On-site experimental set up for estimating organic nitrogen sedimentation rate: positioning of 10 L metal buckets (left) at the bottom of a pond (right).

Close modal

In addition, using an adaptation of the most probable number (MPN) method described by Alexander & Clark (1982) for soil samples, nitrifying and denitrifying bacteria were enumerated in, respectively, water column and sediment samples; in both cases six grab samples were collected near the exit of each pond. Using these same samples and the approach described by Zimmo et al. (2004), nitrification and denitrification rates (g N-NOx/ha·d) were determined, respectively, in the water and in the sediment samples. Nitrification and denitrification rates were calculated on the basis of NOx (nitrite + nitrate) increase and decrease, respectively. Water and sediment samples were incubated for 24 hours at a temperature resembling that of the ponds at the time of sampling. The water samples were exposed to fluorescent lamps for a 12 hour period followed by another 12 hours in the dark, in order to simulate day and night conditions in the field. The sediment samples were kept in the dark. Finally, the assessment of nitrogen removal by algal uptake was based on the suspended organic nitrogen concentration in the ponds' water column, i.e., the difference between unfiltered and filtered total Kjeldahl nitrogen (TKN) values.

Nitrogen removal in the sewage treatment system

The total nitrogen loading rates in the pond systems varied widely during the experimental period: 36,635–64,318 g Ntotal/ha·d in the first pond (P1), 27,130–43,936 in the second pond (P2), 17,660–33,580 in the third pond (P3) and 10,639–25,345 g Ntotal/ha·d in the fourth pond (P4). The total nitrogen removal rates (g Ntotal/ha·d) in the ponds varied as follows: 9,505–20,632 in P1, 5,943–16,097 in P2, 5,454–8,999 in P3 and 1,398–6,785 in P4 (average values are presented in Table 4).

As expected, the ammonia content increased from raw sewage to the UASB effluent, but gradually decreased along the pond series. There were no marked variations of nitrate concentration in the pond system, remaining always below 5 mg/L; nitrite was found only in very low concentrations (0.005–0.007 mg/L) and are not shown in Figure 3. Overall, the pond system achieved high removal efficiency of ammonia, higher than that of total nitrogen, while the organic fraction increased in P1 (in relation to the UASB reactor) and remained at rather constant values along the pond series (average values are given in Table 1 and Figure 3). All this suggests that algal uptake may have played an important role in ammonia removal.

Table 1

Average ammonia and total nitrogen removal efficiencies in the pond system

Nitrogen formRemoval in each pond
Cumulative removal in the pond series
P1P2P3P4P2P3P4
N-NH3 40.6% 39.0% 49.4% 53.0% 63.8% 81.7% 91.4% 
Ntotal 24.3% 26.7% 28.0% 20.4% 44.5% 60.0% 68.2% 
Nitrogen formRemoval in each pond
Cumulative removal in the pond series
P1P2P3P4P2P3P4
N-NH3 40.6% 39.0% 49.4% 53.0% 63.8% 81.7% 91.4% 
Ntotal 24.3% 26.7% 28.0% 20.4% 44.5% 60.0% 68.2% 
Figure 3

Average concentrations of nitrogen species in the sewage treatment plant. RW = raw wastewater; UASB = anaerobic reactor effluent, Pi = 1–4 pond effluents.

Figure 3

Average concentrations of nitrogen species in the sewage treatment plant. RW = raw wastewater; UASB = anaerobic reactor effluent, Pi = 1–4 pond effluents.

Close modal

Laboratory experiments of ammonia capture

Based on the twelve 24 hour laboratory-controlled experiments, an average 53.7% recovery of the volatilized ammonia was recorded. However, in the three extra 48 hour trials the average recovery dropped to 38.5%, indicating increasing losses with longer trials. In similar experiments, Camargo Valero & Mara (2007) recovered 54–60% of the volatilized ammonia, whereas, surprisingly, Assunção & von Sperling (2012) reported 95.8% recovery. The experimental setup of the present work (as described above) and that of the Camargo Valero & Mara study were rather similar one another, whereas that of Assunção & von Sperling (2012) differed in some aspects: a smaller capture chamber (2 L); higher concentration NH4Cl in the capture chamber (40 L); longer trials (48 hours); higher air flow rate (4 L/min as opposed to 2.6 L/min in Camargo Valero & Mara and 3.5 L/min in the present work). The prevailing average temperatures during the three studies were: 21.6 °C (the present work), 30 °C (Assunção & von Sperling) and 17.1 °C (Camargo Valero & Mara). The procedures used here (as described above) and by Camargo Valero & Mara for measuring NH4+ in the capture chamber and in the absorption flasks were essentially the same, but this is not so clear in Assunção & von Sperling (2012). However, these experimental differences may help but are not enough to explain the much higher recovery recorded by Assunção & von Sperling (2012).

Field experiments

In the field, ammonia volatilization rates varied widely, with average and standard deviations values of 211.5 ± 194.5; 211.3 ± 158.0; 97.3 ± 89.7 and 38.4 ± 47.9 g N/ha·day in ponds 1, 2, 3 and 4, respectively. Marked seasonal differences were found (Figure 4) and the volatilization rates seemed to be associated with ammonia loading rates, both along the pond series and in each pond. Overall, the volatilization rates decreased along the pond series (more clearly from pond 2 onwards), probably following the decreasing loading rates (Figure 4). Also, in all ponds, the volatilization rates were higher when the operational conditions (e.g. flow rate, thus HRT) led to higher ammonia loading rates (Table 2). During the summer (with higher temperatures and higher ammonia production in the UASB reactor), the volatilization rate in the first pond (with higher pH values – Figure 5) was noticeably higher than in the other seasons, and was followed by a gradual decrease along the subsequent ponds. In the other seasons, the relatively low rates recorded in pond 1 were followed by an increase in pond 2 (rather intense over the spring), decreasing thereafter (Figure 4). Over the experimental period, in-pond temperatures varied in the following ranges: 26–30 °C, 19–25 °C, 15–23 °C and 23–27 °C during summer, autumn, winter and spring, respectively.

Table 2

Average and standard deviation (in brackets) values of ammonia loading and volatilization rates in the pond system for different hydraulic retention times

HRT (d)P1P2P3P4
 Ammonia loading rate (kg N–NH3/ha·d) 
7.1 44.5 (22.6) 26.9 (13.8) 18.4 (9.9) 10.0 (5.9) 
9.4 23.3 (12.9) 13.5 (9.8) 6.7 (5.6) 2.8 (2.4) 
 Ammonia volatilization rate (g N–NH3/ha·d) 
7.1 272.8 (196.2) 277.4 (134.6) 126.7 (88.7) 49.3 (52.1) 
9.4 50.7 (24.0) 37.8 (24.6) 20.1 (17.5) 9.8 (11.9) 
HRT (d)P1P2P3P4
 Ammonia loading rate (kg N–NH3/ha·d) 
7.1 44.5 (22.6) 26.9 (13.8) 18.4 (9.9) 10.0 (5.9) 
9.4 23.3 (12.9) 13.5 (9.8) 6.7 (5.6) 2.8 (2.4) 
 Ammonia volatilization rate (g N–NH3/ha·d) 
7.1 272.8 (196.2) 277.4 (134.6) 126.7 (88.7) 49.3 (52.1) 
9.4 50.7 (24.0) 37.8 (24.6) 20.1 (17.5) 9.8 (11.9) 
Figure 4

Seasonal variations of the ammonia volatilization rates in the pond system.

Figure 4

Seasonal variations of the ammonia volatilization rates in the pond system.

Close modal
Figure 5

Seasonal variations of pH values in the pond system.

Figure 5

Seasonal variations of pH values in the pond system.

Close modal

In summer experiments in a maturation pond fed by a facultative pond in Bradford (UK), Camargo Valero & Mara (2007) found volatilization rates ranging from 0–27 g N/ha·d (mean rate = 15 g N/ha d, HRT = 17.5 d, an average flow rate of 0.6 m3/d and loadings of 9.3 kg BOD/ha·d and 6.0 kg N/ha d). In the experiments of Assunção & von Sperling (2012) in three maturation ponds fed by a UASB reactor in Brazil, the volatilization rates varied from 3 to 821 g N/ha·d, with mean rates of 112; 151 and 98 g N/ha·d in ponds 1, 2 and 3, respectively (HRT = 4.3 d in P1 and P2 and 1.5 d in P3, average surface loading rates of 70.6 and 76 kg BOD/ha·d in ponds 1, 2 and 3, respectively). Thus, the volatilization rates found here are quite similar to those reported by Assunção & von Sperling in a nearby location, under similar experimental conditions to those of the present study, whereas much lower volatilization rates were found in England in ponds with rather lower loading rates. Finally, it is noteworthy that, like both in Camargo Valero & Mara (2007) and in Assunção & von Sperling (2012), the ammonia volatilization field measurements were much lower than the theoretical figures calculated using Equations (1) and (2): 822–3,050 g NH3⁄ha·d. This reinforces the understanding that equilibrium-based mass transfer models are not adequate for estimating ammonia volatilization in WSPs, as they overestimate field values.

The water column samples revealed the presence of nitrifying bacteria in relatively high numbers: 3.62 × 104–3.31 × 105Nitrosomonas per 100 mL and 4.24 × 103–1.73 × 104Nitrobacter per 100 mL (geometric means). Similarly, denitrifying bacteria were found in sediment samples in the range of 2.89 × 104–2.88 × 105 organisms per gram (geometric means). The fact that nitrate and nitrite concentrations do not increase in the pond effluents does not negate the possibility of nitrification and denitrification being important intermediate steps in nitrogen transformation and removal in WSPs because, as hypothesized by Camargo Valero et al. (2010b), nitrification may be masked by simultaneous biological nitrate uptake and/or denitrification. However, judging by the recorded nitrification and denitrification rates, in relation to the ammonia loading removal rates (Table 3), in this particular study, nitrification-denitrification did not seem to have played a major role in nitrogen removal.

Table 3

Nitrification, denitrification and ammonia loading rates in the pond system

PondNitrification rate (g N-NOX/ha·d)Denitrification rate (g N-NOX/ha·d)Ammonia loading rate (g N-NH3/ha·d)
75 11 42,123 
86 28,444 
90 15,471 
52 6,903 
PondNitrification rate (g N-NOX/ha·d)Denitrification rate (g N-NOX/ha·d)Ammonia loading rate (g N-NH3/ha·d)
75 11 42,123 
86 28,444 
90 15,471 
52 6,903 

The balances shown in Table 4 and Figure 6 revealed the following percentage removal estimates associated with each removal pathway in relation to total nitrogen removal along the four ponds in series: (i) sedimentation of organic nitrogen: 36.3% (P1), 23.5% (P2), 38.5% (P3) and 45.6% (P4); (ii) algal uptake: 13.1% (P1), 16.2% (P2), 17.6% (P3) and 27.8% (P4); (iii) ammonia volatilization: 1.8% (P1), 3.1% (P2), 1.5% (P3) and 1.2% (P4); and (iv) denitrification: 0.24% (P1), 0.15% (P2), 0.34% (P3) and 0.27% (P4). Interestingly, the mass balance becomes more precise along the pond series. Nitrification-denitrification and ammonia volatilization never exceeded figures as low as 0.3 and 3%, respectively. Conversely, algal uptake and sedimentation of organic nitrogen were by far the most important mechanisms for nitrogen removal; this trend increased noticeably along the pond series, probably due to the increasingly favourable in-pond conditions for phytoplankton activity. On the whole, these findings reinforce those previously reported by Zimmo et al. (2003), Camargo Valero & Mara (2007), Camargo Valero et al. (2010a), and Assunção & von Sperling (2012).

Table 4

Nitrogen loading and removal rates in the pond system

ParameterP1P2P3P4
Total N loading rate (g Ntotal/ha·d) 41,641 31,543 23,745 17,380 
Total N load removal rate (g Ntotal/ha·d) 10,098 7,798 6,365 3,849 
Volatilization rate (g N-NH3/ha·d) 184 203 93 39 
Sedimentation rate (g Ntotal/ha·d) 4,044 1,618 2,230 1,797 
Uptake rate (g suspended Norg/ha·d) 1,156 1,080 1,036 931 
ParameterP1P2P3P4
Total N loading rate (g Ntotal/ha·d) 41,641 31,543 23,745 17,380 
Total N load removal rate (g Ntotal/ha·d) 10,098 7,798 6,365 3,849 
Volatilization rate (g N-NH3/ha·d) 184 203 93 39 
Sedimentation rate (g Ntotal/ha·d) 4,044 1,618 2,230 1,797 
Uptake rate (g suspended Norg/ha·d) 1,156 1,080 1,036 931 
Figure 6

Nitrogen removal percentage due to the various pathways in relation to total nitrogen removed.

Figure 6

Nitrogen removal percentage due to the various pathways in relation to total nitrogen removed.

Close modal

The long-standing belief that ammonia volatilization is the major pathway for nitrogen removal in WSPs (at least one of the major mechanisms and mainly in warm-climate regions) has been challenged in more recent work conducted both in temperate and warm climates. This work adds further evidence that ammonia volatilization is not a significant mechanism for both ammonia and total nitrogen removal in maturation ponds in warm climates, reinforcing the major role of algal nitrogen uptake and the subsequent sedimentation of biologically incorporated organic nitrogen.

The authors would like to acknowledge the Brazilian agencies FINEP, CNPq and CAPES for funding this work and providing student scholarships.

Alexander
,
M.
&
Clark
,
F. E.
1982
Nitrifying bacteria
. In:
Methods of Soil Analysis, Part 2. Chemical and Microbiological Proprieties
(
Black
,
C. A.
, ed.).
American Society of Agronomy
,
Madison, Wisconsin
,
USA
, pp.
1467
1472
.
APHA
1998
Standard Methods for the Examination of Water and Wastewater
, 20th edition.
American Public Health Association
,
Washington, DC
.
Assunção
,
F. A. L.
&
von Sperling
,
M.
2012
Importance of the ammonia volatilization rates in shallow maturation ponds treating UASB reactor effluent
.
Water Science and Technology
66
(
6
),
1239
1246
.
Bastos
,
R. K. X.
,
Rios
,
E. N.
,
Dornelas
,
F. L.
,
Assunção
,
F. A. L.
&
Nascimento
,
L. E.
2007
Ammonia and phosphorus removal in polishing ponds. A case study in southeast Brazil
.
Water Science and Technology
55
(
11
),
65
71
.
Camargo Valero
,
M. A.
&
Mara
,
D. D.
2007
Nitrogen removal via ammonia volatilization in maturation ponds
.
Water Science and Technology
55
(
11
),
87
92
.
Camargo Valero
,
M. A.
&
Mara
,
D. D.
2010
Ammonia volatilisation in waste stabilisation ponds: a cascade of misinterpretations?
Water Science and Technology
61
(
3
),
555
561
.
Camargo Valero
,
M. A.
,
Mara
,
D. D.
&
Newton
,
R. J.
2010a
Nitrogen removal in maturation WSP ponds via biological uptake and sedimentation of dead biomass
.
Water Science and Technology.
61
(
4
),
1027
1034
.
Camargo Valero
,
M. A.
,
Read
,
L. F.
,
Mara
,
D. D.
,
Newton
,
R. J.
,
Curtis
,
T. P.
&
Davenport
,
R. J.
2010b
Nitrification-denitrification in WSP: a mechanism for permanent nitrogen removal in maturation ponds
.
Water Science and Technology.
61
(
5
),
1137
1145
.
Middlebrooks
,
E. J.
,
Reed
,
S. C.
,
Pano
,
A.
&
Adams
,
V. D.
1999
Nitrogen removal in wastewater stabilization lagoons
. In:
Proceedings of the 6th National Drinking Water and Wastewater Treatment Technology Transfer Workshop
,
Kansas City, Missouri
,
2–4 August 1999
.
38
pp. ).
Pano
,
A.
&
Middlebrooks
,
E. J.
1982
Ammonia nitrogen removal in facultative ponds
.
Journal of the Water Pollution Control Federation
4
(
54
),
344
351
.
Reed
,
S. C.
1985
Nitrogen removal in wastewater stabilization ponds
.
Journal of the Water Pollution Control Federation
57
(
1
),
39
45
.
Senzia
,
M. A.
,
Mayo
,
A. W.
,
Mbwette
,
T. S. A.
,
Katima
,
J. H. Y.
&
Jorgensen
,
S. E.
2002
Modeling nitrogen transformation and removal in primary facultative ponds
.
Ecological Modeling
154
,
207
215
.
Silva
,
S. A.
,
de Oliveira
,
R.
,
Soares
,
J.
,
Mara
,
D. D.
&
Pearson
,
H. W.
1995
Nitrogen removal in pond systems with different configurations and geometries
.
Water Science and Technology
31
(
12
),
321
330
.
Soares
,
J.
,
Silva
,
S. A.
,
de Oliveira
,
R.
,
Araújo
,
A. L. C.
,
Mara
,
D. D.
&
Pearson
,
H. W.
1996
Ammonia removal in a pilot-scale WSP complex in northeast Brazil
.
Water Science Technology
33
(
7
),
165
171
.
Zimmo
,
O. R.
,
van der Steen
,
N. P.
&
Gijzen
,
H. J.
2004
Quantification of nitrification and denitrification rates in algal and duckweed based wastewater treatment systems
.
Environmental Technology
25
(
3
),
273
282
.