Nitrogen removal in a shallow maturation pond with sludge accumulated during 10 years of operation in Brazil

Polishing ponds are used to improve the quality of urban wastewater after anaerobic treatment, like upflow anaerobic sludge blanket (UASB) reactors, so that the final effluent quality may be compatible with legal standards or desired targets (Cavalcanti et al. 2001). Polishing ponds are designed in a similar way to maturation ponds (that usually follow facultative ponds), and exert approximately the same treatment objectives. This type of wastewater treatment configuration performs well in the removal of organic matter, pathogenic organisms and nitrogen (Dias et al. 2014).

After years of operation, the accumulation of sludge at the bottom of the ponds deserves special attention. In principle, the volume occupied by the sludge could potentially influence treatment efficiency due to the reduction of the hydraulic retention time in the liquid column. However, Keffala et al. (2011) studied a system of ponds in Belgium (Bertrix) and observed that the sediment contributed to the removal of nitrogen through nitrification and denitrification processes. In another study (Yamamoto et al. 2009), the role of the sediment for nitrogen removal in waste stabilization ponds (WSP) was investigated. The results demonstrated that the nitrification capacity was enhanced in the presence of the sediment. Zimmo et al. (2004) also reported higher nitrification rates in wastewater incubations containing sediment compared to those containing only wastewater.

Studies on complementary areas such as hydrodynamics of ponds and molecular biology are important to better understand the biological mechanisms involved in nitrogen removal in ponds. Thermal stratification is usually reported in ponds with depths greater than five metres (Addy & Green 1997). However, due to the turbidity that blocks part of the solar radiation that enters the pond, this phenomenon may occur in systems with shallower depths, especially in the summer (Ukpong et al. 2006). Daily cycles of stratification and mixing can take place as a function of the differences in temperature and liquid density during the day.

The vertical mixing of the liquid transports nutrients and oxygen to the bottom, favoring bacterial growth responsible for nitrogen removal. As reported by Namèche et al. (1997), Chabir et al. (2000) and Keffala et al. (2011), the conditions for the development of nitrifying and denitrifying bacteria are present in the sludge due to the stratified environment in terms of dissolved oxygen (DO) and the hydrodynamic behavior of the ponds.

The objectives of this research were: (i) to determine the nitrogen removal rate by nitrification and denitrification processes measured in sediment cores, (ii) to quantify bacteria involved in the nitrogen cycle and (iii) to evaluate mixing and stratification conditions in a polishing pond with sludge accumulated over 10 years of operation. The major goal was the understanding of possible relevant mechanisms for nitrogen removal in maturation ponds in tropical climates, since there is substantial controversy in the ponds literature on this topic.

The studies were concentrated on Pond 1, which had the following dimensions: length: 25.00 m; width: 5.25 m; depth: 0.80 m. During the experimental period, the median flow was 19.9 m³/d and the median hydraulic retention time was 5.7 d for this pond. This unit was in operation for more than 10 years, and the bottom sludge occupied 40% of the useful volume of the pond (Possmoser-Nascimento et al. 2014). The description of the ponds system can be seen in Rodrigues et al. (2015).

This study investigated the influence of sediments in nitrogen removal (nitrification and denitrification processes). The nitrification and denitrification rates were determined in sediment cores taken from the pond by applying ammonia pulses. Environmental conditions of importance in the nitrogen cycle, represented by pH, DO, temperature (T) and oxidation-reduction potential (ORP), were measured at two different depths: the pond surface and liquid-sediment interface in the pond. Analyses of total solids, total volatile solids (TVS) and total organic carbon (TOC) were performed in samples from the liquid layer and sludge layer (at different depths). TOC was measured in a Shimadzu model TOC-V analyzer. The different bacterial groups involved in the nitrogen cycle were detected by real-time polymerase chain reaction (real-time PCR). The following items show a brief description of the effluent and sludge sampling during the experimental phase.

The monitoring results for the nitrogen fractions (organic, ammonium, nitrite and nitrate) in the treatment line (from the UASB reactor and Pond 1) in this specific study were obtained from May to December 2013. Sampling was done on a predominantly weekly basis. The nitrogen fractions were analyzed according to the methodology described in Standard Methods for the Examination of Water and Wastewater (APHA/AWWA/WEF 2005).

In order to understand the pond's vertical dynamics, DO, T, ORP and pH values were monitored at two different depths, one at 2 cm below the pond surface and the other at the liquid-sediment interface (bottom), in the first polishing pond during a sequence of 7 days. Sampling intervals were 30 minutes. Measurements were made using a multiparametric probe YSI XLM 600 with datalogger.

DNA was extracted from 0.5 g of frozen biomass samples, taken from the sediment core, according to the protocol described by Egli et al. (2003). The DNA samples were extracted from samples of the sediment core, which was constructed based on eight different samples collected from the pond that were pooled together in order to represent one sediment core. This sediment core was prepared in June 2014 (a mild winter in Brazil) and January 2015 (summer in Brazil). In total, four samples were used from each sediment core, from the liquid column, and sludge from the different layers (upper layer – aerobic, middle layer – anoxic and bottom layer – anaerobic). The extracted DNA was purified using the Wizard DNA Kit Clean-Up System (Promega, Madison, USA) and stored at −20 °C. The abundance of total bacteria, ammonia oxidizing bacteria (AOB), nitrite oxidizing bacteria (NOB), anammox and denitrifiers was investigated by real-time quantitative PCR (qPCR) using SyBR green assays on biomass samples taken from the sediment core. qPCR assays were conducted on a real-time PCR thermal cycler (Applied Biosystems 7500 instrument). Each 20 μL reaction mixture contained 10 ng of template DNA, 375 nM of forward and reverse primers (each) and 10 μL of Maxima SYBR Green/ ROX q PCR Master Mix 2X (Thermo Scientific, USA). The total bacterial abundance was quantified using eubacterial 16S rRNA-targeted primers, anammox bacteria were quantified using primers Pla46F and Amx667R, and the abundance of denitrifiers was quantified using primers targeting the nitrous oxide reductase gene nosZF and nosZ1622R (Table 1). PCR conditions were as follows: initial denaturation at 95 °C for 10 min, followed by 40 cycles of 94 °C for 1 min, 53 to 56 °C for 1 min (annealing temperature varied depending on primer used – Table 1), and 72 °C for 1 min. The program finished with the melting curve, which consisted of 15 s at 95 °C, 1 min at 60 °C, 30 s at 95 °C and 15 s at 60 °C. The concentration of the different bacterial groups (as measured by qPCR) was expressed in the number of gene copies (Whelan et al. 2003) per gram of TVS.

Ammonium was added to each core sample in order to make up a concentration of 30 mg of in the supernatant in each of the reconstructed cores, which contained 600 mL of UASB effluent and the sludge layer. The cores were covered with aluminium paper to prevent the growth of algae due to solar radiation incidence. After 4 h of contact between the sediment and the solution (supernatant), samples in each core were taken for initial characterization (t₀). During 5 days, tests were carried out every 24 hours for nitrite ⁠, nitrate ⁠, ammonium ⁠, chemical oxygen demand, DO, T and pH in the supernatant sample. Ammonium was measured with a photometer, Hach DR2800 model 10031 Hach (salicylate method). Nitrite and nitrate were measured with chromatography equipment, Dionex model ICS1000. Ammonium and oxidized nitrogen (N-NO_x = nitrite + nitrate) concentrations (mg·L⁻¹) measured in the supernatant during 5 days of testing were transformed into mass (mg) for calculating the mass balance. The mass of ammonium consumption and the mass of oxidized nitrogen formed were calculated based on the mass balance proposed by Keffala et al. (2011).

As expected, part of the organic nitrogen was converted to ammonium in the UASB reactor due to ammonification. Oxidized nitrogen concentrations in raw sewage were virtually insignificant, but increased slightly in the pond, indicating low nitrification or possibly denitrification. The oxidized forms of nitrogen remained low in the effluent from Pond 1. Regarding the conversion mechanisms, it appears that sediments can play an important role in the global nitrogen mass balance in WSP (Keffala et al. 2011). On the other hand, Camargo Valero & Mara (2007) found that the nitrification process was masked by simultaneous algal nitrate uptake during the peak of algal activity in a maturation pond in the UK. However, in situations in which water quality changes are small, any understanding of the fate of nitrogen is particularly difficult, due to the possibility of having simultaneous processes (e.g. nitrification-denitrification, nitrification-biological uptake).

The pond remained stratified from 6:00 am until 10:00 pm. From 6:00 am, the increase in solar radiation at the surface stimulated algal growth and photosynthetic production, and consequently DO increased, reaching supersaturation at the top layer (∼40 mg·L⁻¹) (Figure 4). The fluctuations in DO are also associated with the oxygen balance between production and consumption in the system (algae and bacteria), leading to the following values: liquid surface: 42.2 mg·L⁻¹ (max), 15.7 mg·L⁻¹ (average) and 0.06 mg·L⁻¹ (min); bottom: 7.10 mg·L⁻¹ (max), 0.79 mg·L⁻¹ (average) and 0.03 mg·L⁻¹ (min). At night, the water temperature decreased, leading to mixing of the liquid masses and oxygenation at the bottom.

The pH and ORP values at the surface presented the same tendency, namely the increase in pH accompanied by the increase in ORP (Figure 4). This behavior was also related to photosynthetic activity of algae that consumed H⁺ ions and increased the pH. The surface pH values ranged from 7.0 (min) to 9.0 (max), with an average of 7.6, while pH values at the bottom showed lower values, ranging from 7.1 (min) to 7.4 (max) and averaging 7.1, probably due to lower photosynthetic activity (Figure 4). pH values above 9.0 enhance free ammonia formation, which may inhibit the activity of nitrifying bacteria (Nitrospira and Nitrobacter) (Blackburne et al. 2007).

Ammonium conversion to nitrite and nitrate followed by their reduction to nitrogen gas depends on environmental conditions (aerobic and anoxic conditions), which also impact bacterial populations responsible for the nitrogen transformations in the ponds. At the surface of the pond, favorable conditions for nitrification could be found (ORP > 300 mV), while in the bottom the ORP values (−50 < ORP < +250 mV) suggest that nitrification and denitrification could take place (in the sludge layer). In addition, anaerobic conditions (with negative ORP values) were observed in the bottom of the pond, late at night and in the early morning, suggesting that nitrogen removal via anammox activity could occur.

The distribution and abundance of bacteria were related to their metabolic capabilities, and their success depends on several abiotic factors, such as temperature, which influences the growth rate of microorganisms. During the day, at the surface of the pond the DO concentration reached 42 mg·L⁻¹ due to intense photosynthetic activity, while in the bottom of the pond, anoxic conditions prevailed over the whole period, with pH around 7.2 and an absence of light. At night, mixing of the water column occurred, and anoxic niches appeared at the surface, while the bottom was oxygenated (around 5 mg·L⁻¹ of DO). Once the light intensity and DO concentrations were limited or practically absent in this region, the anoxic and anaerobic processes (such as fermentation, denitrification and anammox) were predominant.

The abundances of the different bacterial groups remained more or less uniform through the depth of the sediment core. However, the relative abundance of bacteria involved in the nitrogen cycle in relation to total bacteria increased through the depth (from 1.0% in the liquid layer to 7.0% in the anaerobic sludge layer). The concentration of denitrifying bacteria (as measured by nosZ gene abundance) and Nitrobacter (as measured by 16S rRNA gene abundance) were higher than the other groups investigated (AOB, Nitrospira and anammox bacteria).

The presence of denitrifiers through the pond depth was expected, since they can grow under aerobic and anoxic conditions and in the presence of organic matter. In fact, the concentration of denitrifying bacteria correlated well with TOC concentrations observed in each layer. The TOC concentrations were 250, 600, 450 and 500 mg·L⁻¹, respectively, for the liquid layer, aerobic sludge, anoxic sludge and anaerobic sludge. In this pond and in other biological nitrogen removal systems (Mac Conell et al. 2015), denitrifying bacteria were predominant among the nitrogen cycle bacteria, with relative abundances above 91%. The predominance of denitrifying bacteria over the other bacterial groups investigated can be explained by the fact that most of them are heterotrophic bacteria and, therefore, have higher growth rates than the autotrophic bacteria (such as AOB, NOB and anammox).

In relation to AOB, the relative abundance of this group (AOB and anammox) was much lower (ranging from 0.05% to 0.1% in relation to bacteria from the nitrogen cycle) than the abundance of NOB (abundance ranged from 5.5% to 8.5% in relation to bacteria from the nitrogen cycle).

Among the NOB, higher concentrations of Nitrobacter compared with Nitrospira were observed in all layers (Figure 5). Such behavior may be related to changes in nitrite concentration and oxygen availability in the deeper layers. The intense mixing of the water column that was observed might have also contributed to these findings. Several studies have shown that the NOB typically found in sewage treatment systems are Nitrospira and Nitrobacter. Nitrospira are k-strategists, well adapted to low nitrite and oxygen concentrations (half-saturation constant, ⁠; Blackburne et al. 2007), whereas Nitrobacter are r-strategists, and therefore proliferate better under high concentrations of nitrite and oxygen (Schramm et al. 1999). The predominance of Nitrobacter over Nitrospira observed in this study might also be related to the ammonium concentration found in this pond (around 30.4 mg·L⁻¹). According to Blackburne et al. (2007), Nitrospira is affected by free ammonia concentrations of 0.04 to 0.08 mg NH₃-N·L⁻¹, while Nitrobacter is affected by concentrations of 50 mg NH₃-N·L⁻¹.

Since the abundance of ammonia oxidizers was very low, particularly AOB, responsible for supplying nitrite, it is likely that other metabolic routes favored the growth of Nitrobacter. Firstly, nitrite may have been supplied by partial denitrification. Secondly, Nitrobacter may have grown mixotrophically using organic compounds. These two processes were observed and discussed by Winkler et al. (2015) in studies with aerobic granular sludge.

In a previous study, Mac Conell et al. (2015), investigated the microbial community (by 454 pyrosequencing of the 16S rRNA genes) in a trickling filter, also used as post-treatment of anaerobic effluent, receiving wastewater with the same origin as the one in the current study. The authors observed that the relative abundance of denitrifiers ranged from 6% to 16% of total reads; anammox bacteria ranged from 0.5% to 1.0% of the total sequences; and Nitrospira abundance ranged from 0.3% to 1.0% of total reads. The authors concluded that the bacterial community involved in the nitrogen cycle changed with filter depth and during different operational periods. Nitrification occurred in all compartments of the filter, whereas the increase in relative abundance of anammox bacteria occurred when the organic loading rate was lower. The differences observed between these two studies are obviously related to the differences in the treatment processes, but are worth mentioning here because of the similarities in terms of influent sewage and anaerobic pretreatment.

Under the test conditions, part of the removed ammonia was not fully recovered in the form of oxidized nitrogen, suggesting that nitrification was not the only process occurring in the system. This difference in the mass balance could be explained by an alternative process such as denitrification. The loss of ammonia was calculated by mass balance in the system. It was considered that the only nitrogen input was in the form of ammonium and the loss was associated with denitrification. These considerations were made once the algal influence was attenuated (core covered with aluminium paper) and the biological uptake was considered low. The mass balance of the system can be seen in Table 2. The related equations are presented as supplementary material (available with the online version of this paper).

Total nitrogen here represents the sum of ammonium and oxidized nitrogen. The concentrations obtained through mass balance and the final mass measured in each core were not equal. The calculated mass was always greater than the measured mass in practically all cores. It seems that other processes in the system occurred, such as exchange of material between the liquid-sediment interfaces. Moreover, part of the ammonia present in the pond may have come from organic nitrogen degradation in the sediment, deposited in the form of ammonium and not recorded in the mass balance. Unlike the test with samples from sludge and effluent, nitrification and denitrification were close to zero in the test containing only supernatant (the control sample). This result confirmed that nitrogen removal occurred in the sludge layers. Average values of kinetic rates of ammonium consumption were and for formation of oxidized nitrogen were 0.025 g N-NO_x/m² d.

Aerobic and anaerobic conditions were observed in the first polishing pond during 24-hour periods due to mixing and thermal stratification of the liquid column. The bacteria involved in the nitrogen cycle were found in all layers of the sediment core and the concentrations of each bacterial group (AOB, NOB, anammox and denitrifiers) were similar in the liquid layer and in the sludge, which could be explained by the intense mixing of the water column carrying oxygen to the sediment. Denitrifying bacteria were predominant among the nitrogen cycle bacteria, while AOB and anammox were detected in very low concentrations. The availability of oxygen and nitrite might have favored the development of Nitrobacter, since the abundance of this group was much higher than Nitrospira. The mass balance showed that the concentration of oxidized forms of nitrogen was lower than the concentration of the ammonium removed, suggesting the occurrence of denitrification or anammox removal. The results indicate that denitrifiers, nitrifiers and anammox bacteria coexisted in the sludge of this pond, and thus different metabolic pathways were involved in ammonium removal in this wastewater treatment system.

The authors would like to thank the Brazilian agencies CAPES, CNPq and FAPEMIG for their support to the research. Also this research was part of an international program financed by the Bill & Melinda Gates Foundation for the project ‘Stimulating local innovation on sanitation for the urban poor in Sub-Saharan Africa and South-East Asia’ under the coordination of Unesco-IHE, Institute for Water Education, Delft, The Netherlands.