3.1 A and B anaerobic systems
The mean pH values of the effluent inflowing into the two anaerobic treatment systems (Figure 2a, b, and c) were higher than the typical value (7.0) of domestic sewage (von Sperling 2017a), than that found by Silva and Souza (2011) in the studied ST-ANF (7.3), and than the optimal pH range for the development of methanogenic archaea (6.6 - 7.4) (Chernicharo 2007). Therefore, the system may not be performing well in terms of microbiological degradation, requiring pH control interventions.
In B system, however, the pH remained within the range reported by von Sperling (2017a), 6.7 to 8.0, while in A system, the mean pH was above the upper limit. This may be due to the presence of some chemical substances at higher concentrations in the effluent generated by the washing of the stalls, which is directed to P1A. The system in question covers more animals, in addition to stalls sheltering dogs that are undergoing treatment and receiving medication or that underwent surgical procedures. In addition to this hypothesis, the mean COD concentrations (Figure 2d, e, and f) found in system A, in the three phases, were higher than those in B system. In both systems, the mean COD concentration was higher than the mean observed by Souza et al. (2018) of 334 mg L-1 in the same PFA reactors. However, the mean inflowing concentrations in both systems are within the range commonly found in domestic sewage (450 - 800 mg L-1) (von Sperling 2017a).
The mean BOD/COD ratio of the effluent inflowing into the system, calculated from the data presented in Figure 2d, e, f, g, h, and i, demonstrated that the effluent had low biodegradability, namely, 0.3 in Ph1 and 0.4 in Ph2 and Ph3, and thus a high inert fraction. This characteristic can be attributed to the presence of dog hair, oils and fats, cleaning and hygiene products, and medications used in the treatment of the dogs (Souza et al. 2018; von Sperling 2017a; Affandi et al. 2014; Kornillowicz-Kowalska and Bohacz 2011; Merrettig-Bruns and Jelen 2009).
Giustina et al. (2010) found COD concentrations in raw domestic sewage close to the mean in the effluent inflowing into the PFA-WWTP, but with a biodegradability of 0.6. The low biodegradability found in the PFA effluent does not prevent but can hinder biological treatment and, consequently, result in unsatisfactory organic matter removal efficiency (von Sperling 2017a).
The mean organic matter (BOD and COD) removal efficiencies in A and B systems were not satisfactory, as the removal expected for these treatment units, 80 to 85% for BOD and 70 to 80% for COD, was not achieved (von Sperling 2017a) and the minimum efficiency required by law for effluent discharge into waterbodies, 85% for BOD and 75% for COD (Minas Gerais 2008), was not met (Table 2). The concentrations at the end of the anaerobic treatment were also higher than the maximum concentrations allowed for discharge into the stream (60 mg L-1 BOD and 180 mg L-1 COD), with values of 226 and 172 mg L-1 for BOD and 751 and 616 mg L-1 for COD in A and B systems, respectively, without a significant difference (p>0.05) in the comparison of the inlets and outlets of the systems. These findings justify the need to introduce the subsequent aerobic phase as a post-treatment, as already observed by Souza et al. (2018).
Table 2
Mean efficiencies and standard deviation of removal of COD, BOD, TS, SS, TP, TKN and ThermC in anaerobic treatment systems A and B (ST-ANF), from PFA-WWTP (mean of all phases)
Anaerobic
treatment
|
------------------------------P1 - P2 (%)----------------------------
|
P1A e P1B - P3 (log)
|
COD
|
BOD
|
TS
|
SS
|
TP
|
TKN
|
ThermC
|
A System
|
20±28
|
30±31
|
25±25*
|
42±32*
|
12±19
|
23±38
|
1±3
|
B System
|
17±16
|
25±32
|
12±21
|
14±29
|
5±28
|
20±33
|
Efficiencies higher than those found for ST-ANF units were reported in studies on domestic sewage, such as 59 and 77% for BOD and 53 and 79% for COD (Oliveira and von Sperling 2005; Silva and Souza 2011; Moura et al. 2011); this difference may be related to the low biodegradability under anaerobic conditions of the effluent of the PFA-WWTP (Souza 2015).
The removal efficiencies of A system were higher than those of B system, not only for organic matter but also for solids and nutrients. As shown in Figure 2, A system has a primary clarifier before the ST-ANF unit, which helps to increase the efficiencies through the retention of solids, including organic compounds and nutrients (nitrogen and phosphorus) bound to the solids.
It is known the nutrient removal of anaerobic treatment units is limited (Chernicharo 2007). However, the phosphorus incorporated into the sludge in the clarifier or ST-ANF unit can be removed by heterotrophic bacteria that accumulate polyphosphates within the cell, depending on conditions favorable to their growth and metabolism (Henrique et al. 2010). In addition, with the pH above 8.0 in A system, physicochemical removal by phosphorus precipitation may have occurred; with the precipitate retained in the clarifier, nitrogen loss by ammonia volatilization may have been favored (von Sperling 2017b). These processes were reported by Vivan et al. (2010) at a pH of 8.5 in stabilization ponds.
In anaerobic systems, Moura et al. (2011) found removal efficiencies for TP ranging from 9.8 to 86% and TKN ranging from 15 to 77%, while Oliveira and von Sperling (2005) found mean values of 30% for TP and 24% for TKN. According to von Sperling (2017a), no more than 35% TP removal is expected in domestic sewage treatment by ST-ANF due to what was discussed by Lamego Neto and Costa (2011) and Zhang et al. (2017). According to these authors, for complete TKN removal, there must be variation in the redox potential in the environment. Thus, most TKN and TP are usually conserved during anaerobic treatment, with an increase only in the mineralized fraction.
Therefore, the efficiencies found in the anaerobic treatment of the PFA-WWTP were low, and there were no significant differences (p>0.05) between the nutrient concentrations of the inlet and outlet of either system.
Figure 2p, q, r, s, t, and u shows a frequent increase in the concentration of nutrients from the anaerobic units (P2A and P2B) due to the accumulation of the mineralized nutrient fraction in the sludge of the ST-ANF system, which undergoes dragging due to hydraulic load shocks or flows into the system, as demonstrated by Souza et al. (2018). In this sense, the TKN concentrations at the outlet of the anaerobic systems did not meet the legal requirement, with 46 and 29 mg L-1 in A and B systems, respectively, although the law establishes a limit only for ammonia-N, of 20 mg L-1.
The removal of SS (Figure 2m, n, and o) was satisfactory in A system (which has a primary clarifier), meeting the threshold concentration established by law (150 mg L-1), with a mean of 89 mg L-1 and a significant difference between the inlet and outlet (p<0.03). The mean SS concentration at the outlet of B system (171 mg L-1) (without primary clarifier) did not differ significantly from that at the inlet and did not comply with the threshold established by law. This shows that the primary clarifier used in the treatment of the PFA effluent is extremely important for reducing the SS input into the subsequent treatment stages of the PFA-WWTP. Moura et al. (2011) found SS removal efficiencies in an ST-ANF system ranging from 71 to 88%, while Oliveira and von Sperling (2005) found a mean of 66%, considering a total of 166 WWTPs with different configurations. According to von Sperling (2017a), the common SS removal range is 80-90% for domestic sewage; therefore, the efficiency of the ST-ANF units of PFA could be higher if better managed by periodically removing part of the sludge from the treatment units.
Regarding coliform removal, Oliveira and von Sperling (2005) found a mean efficiency of 79% (0.9 log units) in ST-ANF systems, while von Sperling (2017a) reported a common range of 1 to 2 log units removed. Therefore, the efficiency found in the present study was within what is considered satisfactory (Figure 2v, w, and x), although there were no significant differences between the inlet and outlet. In addition, the geometric mean at the outlet of the anaerobic treatments at P3 was still high, 2x106 MPN 100 mL-1.
3.2 Aerobic post-treatment
As it was necessary to increase the efficiency of the PFA-WWTP, the facultative lagoons were replaced by an SABF, a clarifier, and CWs, and the contribution of the reactors is discussed below. The mean pH values (Figure 2a, b, and c) at P3 to P6 remained within the range appropriate for the activity of nitrifying microorganisms, 6.0 to 8.0, and decreased throughout the treatment, indicating nitrification since H+ ions are released in the involved reactions (von Sperling 2012; Zoppas et al. 2016).
Nitrification was higher at P4 (aerobic treatment by SABF) than in the other post-treatment units (Figure 3a and b), as expected, with higher nitrite and nitrate concentrations due to the air blowing and aerobic conditions, which are favorable to the activity of nitrifying bacteria (nitrosation and nitration).
Using the Tukey test, it was observed that only for the variable BOD there was a significant difference in the influent concentration in the different phases. The BOD concentration in phase 3 was significantly lower than in phase 2. Then, indicating that differences in post-treatment performance can be explained to the aeration conditions.
The nitrosation (Figure 3a) and nitration (Figure 3b) in the SABF were significantly higher in Ph3 (intermittent aeration and higher O2 supply), with higher mean concentrations of nitrite (0.05 mg L-1) and nitrate (3 mg L-1) than Ph1 (p<0.02 for nitrite and nitrate) and Ph2 (p<0.001 for nitrite; p<0.02 for nitrate). The results are consistent with the nitrifying bacteria count performed by Souza et al. (2020), an evaluation that showed that the highest density of nitrifying bacteria was found in this phase.
The increase in the concentration of NO2- and NO3- resulted in a higher concentration at the SABF outlet in Ph3, reducing the removal efficiency, which could explain the fact that there was no significant difference in the removal of nitrites and nitrates in Phase 3, as reported in Table 3.
Table 3
Average removal efficiencies of COD, BOD, TS, SS, TP, TKN, TN; NO2-, NO3- and ThermC in the aerobic post-treatment (SABF, SC and CW), of the PFA-WWTP, in the different phases of aeration
Phase
|
Step
|
COD
|
BOD
|
TS
|
SS
|
TP
|
TKN
|
TN
|
NO2-
|
NO3-
|
ThermC (log)
|
----------------------------------------------(%)--------------------------------------------
|
Ph1
|
SABF
|
24±31B
|
44±29A
|
5±6A
|
1±5A
|
14±20A
|
29±30A
|
27±29A
|
7±12A
|
8±19A
|
0±0A
|
SC
|
3±9B
|
33±25A
|
22±14A
|
47±28A
|
11±14A
|
21±28A
|
21±28A
|
78±31A
|
28±30A
|
CW
|
19±25A
|
53±32AB
|
13±16A
|
24±27A
|
23±25A
|
23±35A
|
22±34A
|
17±28A
|
7±15A
|
0±1A
|
Ph2
|
SABF
|
40±24AB
|
38±32A
|
3±6A
|
6±17A
|
5±13A
|
31±30A
|
30±29A
|
2±9A
|
9±16A
|
0±1A
|
SC
|
24±25A
|
51±23A
|
24±8A
|
63±22A
|
18±17A
|
17±24A
|
17±24A
|
80±21A
|
19±21A
|
CW
|
34±31A
|
66±22A
|
12±9A
|
39±35A
|
12±14A
|
21±19A
|
20±19A
|
14±22A
|
13±20A
|
1±0A
|
Ph3
|
SABF
|
51±13A
|
19±19A
|
2±16A
|
10±30A
|
9±13A
|
14±26A
|
14±26A
|
0±0A
|
6±11A
|
1±0A
|
SC
|
23±19AB
|
38±32A
|
25±16A
|
47±23A
|
21±26A
|
35±31A
|
33±30A
|
11±15B
|
29±38A
|
CW
|
23±27A
|
33±26B
|
16±11A
|
23±28A
|
8±14A
|
7±12A
|
5±9A
|
27±24A
|
22±30A
|
0±0A
|
In SC (P5) Ph3, nitrosation remained constant, and there was no nitration, possibly due to the reduction in DO in this reactor, since at lower concentrations, from 0.3 to 0.7 mg L-1, there is a prevalence of ammonia oxidation because the bacteria responsible for nitration are sensitive to low oxygenation of the environment (Zeng et al. 2013; Luesken et al. 2011; Ge et al. 2015; Zoppas et al. 2016; Giongo et al. 2018; Ilyas and Masih 2017; Pelaz et al. 2018). Thus, in this anoxic environment, greater denitrification (nitrate removal) occurred than in the other units and phases, in terms of concentration (greater NO3- concentration), as there was no significant difference in efficiency (Table 3).
It can therefore be inferred that there was SND in the SC according to recent studies (Yao and Peng 2017; He et al. 2016; Kinh et al. 2017; Zheng et al. 2018). Therefore, in this phase, the highest density of denitrifiers was found in the SC, even though this phase had the highest oxygen supply (Souza et al. 2020). The explanation may lie in the favoring of heterotrophic aerobic nitrification by the intermittent aeration, enabling the proliferation of both nitrifying and denitrifying bacteria (He et al. 2016; Kinh et al. 2017; Zheng et al. 2018; Souza et al. 2020).
In Ph3, there was also greater accumulation of sludge in the SC reactor (Figure 2), making the environment more anaerobic, which favored greater adaptation of heterotrophic bacteria, reducing nitrate to nitrogen gas (N2). Under these conditions, denitrifying bacteria may become predominant due to a higher growth rate and cell yield (7.2 d-1; 0.43 g g-1 COD), given the higher affinity for the substrate present (Beltran 2008; Nocko 2008; Zoppas et al. 2016; Metcalf and Eddy 2003). Consequently, there was greater N removal efficiency in the SC, with concentrations of 20 mg L-1 for total nitrogen (TN) and 18 mg L-1 for TKN, within the limit for discharge into waterbodies (≤20 mg L-1 ammonia-N).
In phases Ph1 and Ph2, the sludge accumulation was lower because periodic cleaning of the SC was performed every 7 months, occurring once at the beginning of the monitoring, again during Ph2, and coinciding with the end of the study; thus, Ph3 did not include a cleaning intervention. Another factor that may have contributed to the sludge accumulation in the SC in Ph3 was the greater oxygen supply. Aerobic treatments are commonly characterized by large sludge generation (von Sperling 2017a; Jordão and Pessoa 2017), as observed in P4 (Figure 2m, n, and o); therefore, a increased O2 availability contributed to greater organic degradation and sludge generation.
In Ph3, the SS concentration in the SC was 119 mg L-1, higher values than in other phases, however, the efficiencies of TS (dissolved and suspended) and TP retention were only numerically greater than in phases 1 and 2 (Table 3). It is known that TP removal is related to TS retention, through the incorporation of the nutrient in the microbial biomass and, consequently, in the removed sludge as discussed by Henrique et al. (2010). Due the greater production of sludge and the alternating of anaerobic and aerobic phases (Henrique et al. 2010; von Sperling 2017b), the expectation that the phase 3 would be the most effective phase. One hypothesis is that it is related to the need to remove accumulated sludge more frequently. However, notably, the SC, in all phases, was able to reduce the SS concentration to values acceptable by law (≤150 mg L-1) (Minas Gerais 2008).
Giustina et al. (2010) highlighted the need for devices to retain the sludge from the SABF. Due to the high sludge generation and low retention capacity of the analyzed biofilters (with no. 4 gravel, PET (Polyethylene terephthalate), and a Pall ring), a solid retention chamber after aerobic treatment was essential. According to these authors, due to the presence of the sludge retention chamber, immediately after the biofilter, it was possible to achieve a mean SS removal efficiency of 94% and a COD of 72%. In addition to increasing the solids removal, the clarifier is important for increasing the service life of the CWs (Matos et al. 2018).
Higher organic matter degradation efficiency was observed, considering the COD removal in the SABF during Ph3 (statistically compared to phase 1), a period with greater applied aeration rate (Table 3), favoring the metabolism of heterotrophic organisms (von Sperling 2017b). Hasan et al. (2014) also detected a higher COD removal efficiency of 99% in the phase in which the aeration rate increased from 0 to 2 L min-1.
Souza et al. (2018) analyzed the PFA wastewater and identified greater COD removal capacity in the presence of oxygen, explained by the high difficult-to-biodegrade fraction, as previously described. At P3, the BOD/COD ratio was 0.4 in Ph1, 0.3 in Ph2, and 0.2 in Ph3, i.e., the effluent inflowing into the post-treatment showed even lower biodegradability since most of the biodegradable fraction was removed in the previous treatment by the ST-ANF unit (in the raw effluent, the average BOD/COD ratio is 0.4; 0.3; 0.3, respectively). Thus, there was a greater percent reduction in the BOD concentrations than the COD concentrations by biological processes. It is noteworthy that, although the SABF affluent was less biodegradable in Phase 3 compared to Phase 2, with BOD concentration significantly lower, without the same occurring with COD, there was no significant difference in organic matter removal efficiencies. This condition indicates the importance of oxygen supply to increase the ability to remove organic matter.
The organic matter (BOD) removal by the SABF was complemented by the SC, which showed good efficiency; therefore, the BOD concentration at the outlet of the SC was 59 mg L-1 in Ph2 and 53 mg L-1 in Ph3, complying with the legal limit (60 mg L-1), which was not the case for COD, except in Ph3. In the phase with the longest aeration time, the CW was able to reduce the COD concentration to acceptable values for discharge into the watercourse (180 mg L-1) (Minas Gerais 2008), with an outlet concentration of 166 mg L-1.
As described in other studies (Lizama et al. 2011; He et al. 2013; Li et al. 2013; Avelar et al. 2014; Priya and Philip 2015; Tao et al. 2016; Ramos et al. 2017), the removal of P and especially of N was observed in the CWs in Ph1 and Ph2. However, due to the higher inflowing organic matter concentrations, which resulted in higher sludge generation in these phases, the treatment did not achieve the necessary efficiency. In contrast, in Ph3, there was an increase in nutrient concentration, given the accumulation of sludge in the SC and dragging to the CWs, reducing the removal capacity in the reactor. The mean concentrations were 25 mg L-1 for TN and 24 mg L-1 for TKN, values slightly higher than the limit of 20 mg L-1 for ammonia-N established by law. During this period, nitration (nitrite removal) and denitrification (nitrate removal) occurred in the reactors.
According to the literature, plants in CWs favor N removal by absorption and by oxygen absorption and pumping in the root zone, contributing to nitrification in aerobic microzones and to denitrification in predominantly anoxic or anaerobic regions (Adradros et al. 2014; Taylor et al. 2011; Sarmento et al. 2012; Samsó and García 2013; Avelar et al. 2014; Toscano et al. 2015; Fan et al. 2016). Therefore, there were conditions for SND to occur in the CWs, as observed by the nitration and denitrification efficiency data (Table 3). The co-occurrence of aerobic, anoxic, and anaerobic conditions within a CW can increase the nitrification and denitrification activity (Pelissari et al. 2017; 2018), with satisfactory results for TN removal, ranging from 40 to 70% (Dong and Sun 2007; Huang et al. 2017; Pelissari et al. 2017; Saeed and Sun 2017).
The CW in Ph2 removed 1 log ThermC, as the SABF followed by the SC in Ph3, that showed better conditions for ThermC removal, which may be due to oxidation or competition with other microorganisms. However, the sludge dragged to the CW and, consequently, the accumulation of microorganisms may have reduced the inactivation conditions (as it could have happened on the CW in phase 3), considering that cultivated CWs operated with lower organic loads usually have a greater capacity for coliform removal due to the release of antimicrobial substances by plants (Avelar et al. 2014; Brix et al. 1989).
Regarding the detection performed by the PCR technique, the presence of the protozoan genera Giardia and Cryptosporidium was confirmed in the effluent samples in Ph1 and Ph2 at all analyzed stages (Table 4), confirming that the post-treatment was not able to remove pathogenic microorganisms.
Table 4
Results PCR's with specific primers for Giardia and Cryptosporidium at different sampling points of the kennel WWTP
Samples
|
Giardia
|
Cryptosporidium
|
SABF Ph1
|
+
|
+
|
SC Ph1
|
+
|
+
|
CW Ph1
|
+
|
+
|
SABF Ph2
|
+
|
+
|
SC Ph2
|
+
|
+
|
CW Ph2
|
+
|
+
|
SABF Ph3
|
+
|
+
|
SC Ph3
|
+
|
+
|
CW Ph3
|
+
|
+
|
Microorganisms of the genera Giardia and Cryptosporidium are commonly found in dog fecal samples; however, there are few studies on the role of dogs in zoonotic parasite transmission, although it is known that dog feces are a potential source of infection by these microorganisms (Julien et al. 2019).
3.3 Evaluation of the PFA-WWTP as a whole
Analysis of the PFA-WWTP treatment system as a whole (Table 5) shows that Ph3, in addition to presenting satisfactory organic matter concentrations at the WWTP outlet (P6) (Figure 3f and i), showed higher removal efficiency than that established by law for COD (70%) and BOD (75%) (Minas Gerais 2008) (Table 5). In the other phases, the removal efficiency reached the standards established by law only for BOD. Therefore, the addition of more aeration time, with intermittent supply into the biofilter, contributed to N removal and COD stabilization, which was high in the effluent studied, as in the study by Hasan et al. (2014).
Table 5
Mean removal efficiencies and standard deviation of COD, BOD, TS, SS, TP, TKN, ThermC and TN in the PFA-WWTP, in the different aeration phases
Phase
|
P1A+P1B a P6 (%)
|
P3 a P6 (%)
|
COD
|
BOD
|
TS
|
SS
|
TP
|
TKN
|
ThermC
|
TN
|
Ph1
|
30±26B
|
89±9AB
|
34±25A
|
48±38A
|
30±30A
|
50±30A
|
4±6 log
|
35±33A
|
Ph2
|
66±25A
|
90±8A
|
34±22A
|
70±29A
|
10±14A
|
23±31AB
|
5±1 log
|
34±30A
|
Ph3
|
71±20A
|
77±19B
|
43±27A
|
43±27A
|
14±21A
|
7±22B
|
1±0 log
|
14±17A
|
The removal of SS was satisfactory in all phases, despite the accumulation and dragging of the material in the SC and CW in Ph3, due to the greater sludge generation in the SABF with the greater oxygen supply. Possibly for this reason, the removal efficiency of TKN was low in this phase, while P removal did not differ statistically (Tables 3 and 5).
However, the concentration of N (TKN and NT) at the WWTP outlet (P6) in Ph3 (Figures 2r and 3c) was close to the legal limit. This shows that the treatment system has the capacity to provide efficient N removal, and this aeration strategy can be adopted, as observed by Pelissari et al. (2017; 2018). However, a shorter interval between SC cleanings should be adopted to take advantage of the potential of the operating condition.
The accumulation and dragging of sludge in Ph3 may also have been factors that reduced the efficiency of ThermC inactivation because the system showed good removal of microorganisms in Ph1, beginning after the SCs were cleaned, and in Ph2, in which cleaning was performed through monitoring (Table 3). Ph3 included the aggravating factor of higher sludge generation and lack of cleaning in the monitored period, as explained by Avelar et al. (2014).