In this field study, the impacts of influent loadings and drying-wetting cycles on N2O emission in a subsurface wastewater infiltration (SWI) system were investigated. N2O emitted under different operation conditions were quantified using static chamber and gas chromatograph techniques. N2O conversion rate decreased from 6.6 ± 0.1% to 2.7 ± 0.1% with an increase in hydraulic loading (HL) from 0.08 to 0.24 m3/m2·d. By contrast, N2O conversion rate increased with increasing pollutant loading (PL) up to 8.2 ± 0.5% (PL 4.2 g N/m2·d) above which conversion rate decreased, confirming that N2O production was under the interaction of nitrification and denitrification. Taking into consideration the pollutants (chemical oxygen demand (COD), NH4+-N, NO3-N and total nitrogen (TN)) removal ratio and N2O emission, optimal loading ranges and drying-wetting modes were suggested as HL 0.08–0.12 m3/m2·d, PL 3.2–3.7 g N/m2·d and 12 h:12 h, respectively. The results revealed that in SWI systems, conversion ratio of influent nitrogen to N2O could be between 4.5% and a maximum of 7.0%.

INTRODUCTION

The discharge of effluent from domestic sources has detrimental effects on the aquatic ecosystem as this can deposit a large amount of organic matter, nutrients and pollutants leading to eutrophication, temporary oxygen deficits and accumulation of pollutants into receiving waterways (Kong et al. 2002). As a result, some advanced treatment technologies are generally deemed necessary to decrease these constituents prior to reuse. Among the many treatment technologies developed to improve effluent water quality, subsurface wastewater infiltration (SWI) approach is considered to be a promising alternative due to lower investment and operational costs, simple management requirement, prolonged reliability and durability.

As a kind of natural wastewater purification system, SWI is an effective way to treat wastewater according to integrated mechanisms of chemical, physical and biological reactions as it passes through the unsaturated soil in infiltration system (Pan et al. 2015). SWI is different from other well-known natural treatment systems, such as subsurface wetland and surface wetland, especially in nitrogen removal mechanisms. Constructed wetlands have higher NH4+-N removal efficiency because of aerobic conditions (by the oxygenation of rhizosphere and the aeration of top layer) developed within the system (Zapater-Pereura et al. 2016). In addition, plant uptake plays an important role in nitrogen removal (Zapater-Pereura et al. 2016).

In SWI treatment, organic nitrogen is easily converted to ammonia nitrogen (NH4+-N) by ammonifying bacteria and then the NH4+-N can be adsorbed onto the soil because zeta potential of the soil particles are negative, and adsorption capacity of the soil can be quickly recovered to the initial state by nitrifying NH4+-N to NOx-N, which is subsequently denitrified to N2 or N2O by the denitrifying biomass under the anoxic condition (Li et al. 2014). When one of the following conditions occurs, such as low dissolved oxygen level or nitrite nitrogen (NO2-N) accumulation or limited nitric oxide reductase (Nos) activity, N2O will be generated via nitrification or denitrification process (Sun et al. 2013). Mechanisms for N2O emission are shown in Figure 1.
Figure 1

N2O production mechanisms in SWI system.

Figure 1

N2O production mechanisms in SWI system.

The production of N2O during nitrification process mainly occurs in the oxidation of NH4+-N and hydroxylamine (NH2OH). When dissolved oxygen level is limited, NO2-N will accumulate and extend its toxicity to microbial cells. Thus, heterogeneous nitrite reductase will be produced to reduce NO2-N to N2O. On the other hand, denitrification process can promote N2O emission in two ways: Nos activity is restrained and NO2-N cannot be reduced completely, resulting in the NO2-N accumulation and N2O production. In addition, under acid condition, accumulated NO2-N reacts with soil humic acid and generates N2O. Therefore, there is a hypothesis that N2O generation is the interaction of nitrification and denitrification.

Kong et al. (2002) compared the N2O emission of SWI in Japan with that in China by fact-finding survey, and found a positive correlation between nitrogen removal, N2O emission and oxidation reduction potential. Studies conducted by Ahn et al. (2011), Rassamee et al. (2011) and Li et al. (2016) confirmed that loading rate and drying-wetting regimes could affect nitrogen circulation in the following ways: (1) increasing hydraulic loading (HL) means shortening hydraulic retention time (HRT), so organic nitrogen is not fully degraded before discharged from the SWI system; (2) increasing loading leads to stronger shock for media surfaces, which is also responsible for NO2-N accumulation and N2O production, as shown in Figure 1. Therefore, it is clear that loading and drying-wetting cycles can have a significant effect on the N2O emission. This field study was carried out to (1) explore the influence of loading rate and drying-wetting cycle on N2O emission and (2) suggest optimal loading and cycle ranges taking into consideration of treatment efficiency and N2O emission abatement.

MATERIALS AND METHODS

Site description

The experiment was carried out at the campus wastewater treatment plant in Shenyang, China. The SWI system has a design capacity of 3,000 people equivalent and covers 3,000 m2, dividing equally into eight independent cells. The inflow rate was 300 m3/day. The primary effluent flows by gravity into the primary and secondary distribution tanks. Then, the wastewater flows through the SWI systems, and the effluent is finally collected into the collection tank (Figure 2). The SWI system is planted with turf type ryegrass (Lolium perenne L.).
Figure 2

Sketch showing of the SWI system and static chambers. (a) Box. (b) Foundation support.

Figure 2

Sketch showing of the SWI system and static chambers. (a) Box. (b) Foundation support.

Static chambers were used to monitor N2O production (Tsushima et al. 2014). Four gas chambers (Figure 2, stainless steel, length 50 cm, width 50 cm, height 40 cm, sealed with a water-filled U trench, painted white to avoid heating during the application) were installed at four cells.

Matrix

The filling materials were meadow brown soil, dry activated sludge and slag mixed evenly in volume ratio 7:2:1. The meadow brown soil was sampled from the top 20 cm from Shenyang Ecological Station. The activated sludge was obtained from the aeration tanks in Shenyang Northern Municipal Sewage Treatment Plant, China, air dried after being centrifuged for 15 min at 1,500 rpm. Other materials (gravel and coal slag) were purchased from a local market (particle size: gravel 10–25 mm and coal slag 4–8 mm). The infiltration rate, porosity and surface area of the matrix were 0.37 m3/m2·d, 59% and 5.21 m2/g, respectively. The maximum adsorption capacity for NH4+-N was 0.724 mg/g.

Sampling and analytical methods

Water quality

Water samples were taken between 9:00–10:00 a.m. once a week from the secondary distribution tank and collection well, respectively. Influent and effluent samples were analyzed immediately for pH, chemical oxygen demand (COD), biochemical oxygen demand (BOD), NH4+-N, NO3-N and total nitrogen (TN) using Chinese EPA standard methods (Chinese EPA 2002). From May 5th to September 15th, 2015, the variations of influent were 7.0–7.4 for pH, 210–361 mg/L for COD, 125–270 mg/L for BOD5, 17–25 mg/L for NH4+-N, 27–43 mg/L for TN and 2–8 mg/L for NO3-N.

Gas collection and analytical method

Gas sampling was carried out once a week simultaneously with that of wastewater. The first air sample was collected immediately after the chamber was sealed. According to Kong et al. (2002), saturation time for field scale static chamber would not exceed 60 min. Thus, five air samples were taken from the roof of the chamber at 15 min intervals (0, 15, 30, 45 and 60 min) using a 50-mL plastic syringe with a three-way stopcock and a Teflon tube. The final accumulation value minus the initial one served as the basis for the calculation of N2O production rates.

Gas samples were analyzed using a gas chromatograph (GC, 7890A, Agilent Technologies, USA) equipped with micro-electron capture detector running at 350 °C. High-purity dinitrogen (N2 99.999%) was used as the carrier gas for N2O analysis. The system was calibrated during each measurement cycle using known concentrations of gas (0.390 ppm N2O in pure N2). The calibration line used to analyze N2O air samples by GC had a precision of 96.8%. Therefore, an error of ±3.2% may be assumed. Air samples were analyzed three times to quantify the human error that can be introduced by manual injection of the samples into the GC. Correlation analysis was carried out using SPSS 18.0.

Calculation methods

The production rate of N2O was estimated as the production of N2O per unit area and unit interval. It was calculated by formula (1) (Sun et al. 2013), where F was N2O production rate (mg/m2·h), ρ was the density of N2O at 0 °C and 760 mm Hg (kg/m3), ΔC/Δt was the rate of N2O concentration increased in the chamber (ppmv N2O/h), H was the height of the chamber headspace (m) and T was the ambient air temperature (°C).
formula
1
Conversion ratio was the percentage of N2O removal in influent TN. Its calculation was based on formula (2), in which P was the conversion ratio of N2O emission (%), m was the quantity of nitrogen in N2O (g) and M was the quantity of total nitrogen in influent (g).
formula
2

Statistical analyses were carried out with MicroCal Origin 8.0 (OriginLab) and SPSS 18.0.

Experimental operation

During the whole experimental period, SWI systems were operated in alternation of wetting and drying cycles. Each cycle of the intermittent operation included a continuous flow period of 12 h and a drying period of 0, 4 h, 6 h, 12 h, 24 h and 36 h, indicating that the drying-wetting cycles were 0, 1:3, 1:2, 1:1, 2:1 and 3:1, respectively.

RESULTS AND DISCUSSION

N2O emission and conversion under variable HLs

During the first study period, HL (wastewater load per unit area per day) was set at 0.08, 0.12, 0.16, 0.20 and 0.24 m3/m2·d. Effluent quality, N2O production and conversion ratio were analyzed, as shown in Figures 3 and 4(a), respectively.
Figure 3

The variation of effluent quality under different HLs.

Figure 3

The variation of effluent quality under different HLs.

Figure 4

The variation of N2O production rate and conversion ratio under HL (a) and PL (b).

Figure 4

The variation of N2O production rate and conversion ratio under HL (a) and PL (b).

With the increase of HL, the removal rates of BOD5 and NH4+-N decreased gradually, while that of NO3-N presented an increasing trend. When the HL was lower than 0.12 m3/m2·d, removal rates for COD and TN increased with HL increase. However, when HL was higher than 0.12 m3/m2·d, the removal rates decreased dramatically. Correlation analysis for Figure 3 demonstrated that BOD5 and NH4+-N removal rates had negative correlations with HL (P < 0.05). A high loading rate of carbon (BOD5) leads to increased denitrification rates both by increasing the oxygen consumption making conditions more suitable for denitrification and providing organic carbon in the denitrification process (Li et al. 2011). In general, the COD requirement for denitrification is 3.0–4.5 mg COD/mg TN (Desloover et al. 2011). Thus, COD and TN removal presented similar trends. With respect to NO3-N removal rate, HL had a positive effect on it. The N2O production rate demonstrated an increasing trend firstly and then fell. The production rate, as high as 4.1 ± 0.2 mg/m2·d, was achieved when the hydraulic was 0.16 m3/m2·d. On the other hand, N2O conversion ratio decreased from 6.6 ± 0.1% to 2.7 ± 0.1% with an increase of HL from 0.08 to 0.24 m3/m2·d.

These observations confirmed that HL influenced organics removal in SWI in the following ways: first, the increase of HL shortened HRT. The organics were not fully degraded before discharged from SWI (Tsushima et al. 2014). Second, the increase of HL led to stronger shock for the matrix, which was also responsible for the decrease of removal rates (Bednarek et al. 2014). For NH4+-N and NO3-N removal rates, the results were a consequence of competition between heterotrophic and autotrophic bacteria (Li et al. 2011). As mentioned before, the increase of HL resulted in higher organic loadings, which could shift in favor for heterotrophic bacteria rather than autotrophic bacteria contribution. Taking pollutants removal and N2O emission into consideration, the recommended range for HL was 0.08–0.12 m3/m2·d.

N2O emission and conversion under variable pollutant loadings

During the second experimental period, HL was fixed at 0.10 m3/m2·d. pollutant loading (PL) gradually increased from 3.2 to 3.7, 4.2, 4.7 and 5.2 g N/m2·d. Water quality, N2O production and conversion ratio were analyzed and shown in Figures 4(b) and 5.
Figure 5

The variation of removal efficiency under different PLs.

Figure 5

The variation of removal efficiency under different PLs.

Correlation analysis suggested that PL exerted significant negative influence on the removal of COD (r = −0.976, P = 0.04) and NH4+-N (r = −0.994, P = 0.01). Average NO3-N removal rates increased from 86.4% to 99.1% with the increase in PL. Correlation analysis revealed that PL positively correlated with NO3-N removal (r= 0.990 and P = 0.01).

Increased denitrification with PL probably resulted from the decreased oxygen diffusion into microsites and decreased oxygen concentration due to increased O2 consumption (Pan et al. 2015). Zhu et al. (2015) studied the effect of nitrogen application on N2O emission and suggested that external nitrogen source can increase N2O emission and the high peaks of N2O emission were observed during external nitrogen source application. By contrast, the results from this study showed that when PL grew as high as 4.2 g N/m2·d, N2O emission reached its maximum followed by a dramatic decrease. In addition, correlation analysis showed that under loading rates of low to medium level (HL 0.08–0.16 m3/m2·d, PL 3.2–4.2 g N/m2·d), N2O production rate showed a positive correlation with loadings. Instead, loading rates had significant negative effects on N2O production when the loadings were in high ranges (P < 0.05). Variations of N2O emission along with loading rates verified that N2O production was the interaction of nitrification and denitrification process (Rajagopal & Béline 2011; Ak & Gunduz 2013; Hernandez-Paniagua et al. 2014).

Considering the pollutant (COD, NH4+-N, NO3-N and TN) removal rates and N2O generation rate, influent PL with ranges of 3.2–3.7 g N/m2·d was recommended. Under the optimal conditions (HL 0.08–0.12 m3/m2·d and PL 3.2–3.7 g N/m2·d), effluent quality could meet the Water Quality Standard for Scenic Environmental Use in China (GB/T 18921–2002, BOD5 ≤ 10 mg/L, NH4+-N ≤ 5 mg/L, TN ≤ 15 mg/L and TP ≤ 0.5 mg/L). Conversion ratio for nitrogen to N2O was between 4.5% and 7.0%.

During the studied period, clogging did not occur owing to the proper construction and operation of the system. The results of infiltration rate indicated that HL and PL had negative effect on infiltration capacity. After some period of operation, SWI experienced a sufficient decline in the infiltration capacity under high loading rates. Especially in the 0.2–0.4 m depth interval, infiltration rates decreased gradually to 9.0 × 10−4 cm/s (60–80% of the initial value). This decline was attributed to the accumulation of biodegradable organic matter (measured by total solids, total suspended solids and BOD) at high loading rates. If the infiltration rate continues to decrease, the environmental conditions of upper layer would immediately begin to shift from aerobic to anaerobic, corresponding to the pollutant (COD, NH4+-N, NO3-N and TN) removal rates decreasing and matrix clogging. Therefore, in order to extend the lifespan and keep a reliable removal performance, field scale SWI should be operated under the optimized loadings. Moreover, water pH in the upper layers (0.2–0.4 m) declined from 7.2 (mean value of influent) to 6.9 (on average), indicating that, in this section, nitrification process played a dominant role in nitrogen cycle. Then, pH showed an increase to 7.1 (mean value of effluent) caused probably by denitrification process. As a result, pH and N2O could be seen as indicative indexes for the nitrification and denitrification process.

N2O emission under variable drying-wetting cycles

Table 1 describes pollutant removal efficiencies and effluent concentration under continuous and intermittent feeding modes. It is clear that under continuous operation (drying-wetting cycle 0), pollutant removals were relatively low, especially for COD and NH4+-N. When the feeding days was extended, removal efficiency of NH4+-N continuously declined to 88.4 ± 0.9% under drying-wetting cycle of 1:3. In contrast, TN removal efficiency increased with the declining of the drying-wetting cycle. The possible reasons may be: prolonged feeding days led to increased denitrification rates both by increasing the oxygen consumption making conditions more suitable for denitrification and providing organic carbon needed in the denitrification process.

Table 1

Effluent quality under different drying-wetting cycles

Drying-wetting cycleCOD
BOD
NH4+-N
NO2-N
Effluent mg/LRemoval %Effluent mg/LRemoval %Effluent mg/LRemoval %Effluent mg/LRemoval %
3:1 20.0 ± 8.0 93.3 ± 2.6 2.0 ± 1.0 99.1 ± 0.4 1.0 ± 0.2 95.3 ± 0.9 0.26 ± 0.05 13.3 ± 16.7 
2:1 14.0 ± 4.0 95.3 ± 1.3 5.0 ± 1.0 97.8 ± 0.4 1.0 ± 0.15 95.3 ± 0.7 0.27 ± 0.07 10.0 ± 23.3 
1:1 26.0 ± 5.0 91.3 ± 1.6 5.0 ± 1.0 97.8 ± 0.4 1.8 ± 0.2 91.6 ± 0.9 0.48 ± 0.03 60.0 ± 10.0 
1:2 32.0 ± 6.0 89.3 ± 2.0 9.0 ± 2.0 96.1 ± 0.9 2.0 ± 0.3 90.7 ± 1.4 0.42 ± 0.04 40.0 ± 13.3 
1:3 43.0 ± 3.0 85.7 ± 1.0 12.0 ± 1.0 94.8 ± 0.4 2.5 ± 0.2 88.4 ± 0.9 0.00 ± 0.00 100.0 ± 0.0 
60.0 ± 2.0 80.0 ± 0.7 15.0 ± 2.0 93.5 ± 0.9 5.0 ± 0.15 76.7 ± 0.7 0.05 ± 0.05 83.3 ± 16.7 
NO3-N
TN
TP
Drying-wetting cycleEffluent mg/LRemoval %Effluent mg/LRemoval %Effluent mg/LRemoval %
3:1 1.5 ± 0.10 50.0 ± 3.3 4.8 ± 0.3 80.8 ± 1.2 0.0 ± 0.0 100.0 ± 0.0   
2:1 1.2 ± 0.08 60.0 ± 2.7 4.3 ± 0.3 83.2 ± 1.2 0.0 ± 0.0 100.0 ± 0.0   
1:1 1.0 ± 0.05 66.7 ± 1.7 2.8 ± 0.2 88.8 ± 0.8 0.2 ± 0.1 95.0 ± 2.5   
1:2 0.8 ± 0.05 73.3 ± 1.0 2.6 ± 0.1 88.6 ± 0.4 0.2 ± 0.2 95.0 ± 5.0   
1:3 1.0 ± 0.05 66.7 ± 1.4 2.5 ± 0.4 89.6 ± 1.6 0.2 ± 0.1 95.0 ± 2.5   
0.5 ± 0.05 83.3 ± 1.2 6.4 ± 0.3 74.4 ± 1.2 0.2 ± 0.1 95.0 ± 2.5   
Drying-wetting cycleCOD
BOD
NH4+-N
NO2-N
Effluent mg/LRemoval %Effluent mg/LRemoval %Effluent mg/LRemoval %Effluent mg/LRemoval %
3:1 20.0 ± 8.0 93.3 ± 2.6 2.0 ± 1.0 99.1 ± 0.4 1.0 ± 0.2 95.3 ± 0.9 0.26 ± 0.05 13.3 ± 16.7 
2:1 14.0 ± 4.0 95.3 ± 1.3 5.0 ± 1.0 97.8 ± 0.4 1.0 ± 0.15 95.3 ± 0.7 0.27 ± 0.07 10.0 ± 23.3 
1:1 26.0 ± 5.0 91.3 ± 1.6 5.0 ± 1.0 97.8 ± 0.4 1.8 ± 0.2 91.6 ± 0.9 0.48 ± 0.03 60.0 ± 10.0 
1:2 32.0 ± 6.0 89.3 ± 2.0 9.0 ± 2.0 96.1 ± 0.9 2.0 ± 0.3 90.7 ± 1.4 0.42 ± 0.04 40.0 ± 13.3 
1:3 43.0 ± 3.0 85.7 ± 1.0 12.0 ± 1.0 94.8 ± 0.4 2.5 ± 0.2 88.4 ± 0.9 0.00 ± 0.00 100.0 ± 0.0 
60.0 ± 2.0 80.0 ± 0.7 15.0 ± 2.0 93.5 ± 0.9 5.0 ± 0.15 76.7 ± 0.7 0.05 ± 0.05 83.3 ± 16.7 
NO3-N
TN
TP
Drying-wetting cycleEffluent mg/LRemoval %Effluent mg/LRemoval %Effluent mg/LRemoval %
3:1 1.5 ± 0.10 50.0 ± 3.3 4.8 ± 0.3 80.8 ± 1.2 0.0 ± 0.0 100.0 ± 0.0   
2:1 1.2 ± 0.08 60.0 ± 2.7 4.3 ± 0.3 83.2 ± 1.2 0.0 ± 0.0 100.0 ± 0.0   
1:1 1.0 ± 0.05 66.7 ± 1.7 2.8 ± 0.2 88.8 ± 0.8 0.2 ± 0.1 95.0 ± 2.5   
1:2 0.8 ± 0.05 73.3 ± 1.0 2.6 ± 0.1 88.6 ± 0.4 0.2 ± 0.2 95.0 ± 5.0   
1:3 1.0 ± 0.05 66.7 ± 1.4 2.5 ± 0.4 89.6 ± 1.6 0.2 ± 0.1 95.0 ± 2.5   
0.5 ± 0.05 83.3 ± 1.2 6.4 ± 0.3 74.4 ± 1.2 0.2 ± 0.1 95.0 ± 2.5   

When the running time was extended, N2O emission showed periodic variations (Figure 6). During the wetting period (saturated conditions in lower layers), pollutants were removed due to the suspended solids deposition and bacterial growth in soil spaces. For one thing, prolonging the wetting period prevents the penetration of oxygen into the soil matrix, promoting denitrification smooth processes (as shown in Table 1, NO2-N, NO3-N and TN removal rates increased by prolonging the wetting days). In addition, a longer wetting period ensured longer residence times, which, in essence, were observed to be a crucial factor in the removal of more complex organic matter present in water (Rassamee et al. 2011). Hence, lower drying-wetting cycle (0, 1:3 and 1:2) showed relatively stable N2O emission rate.
Figure 6

N2O production rate under variable drying-wetting cycles.

Figure 6

N2O production rate under variable drying-wetting cycles.

By contrast, intermittent operation rather than continuous feeding is an encouraging method to ensure the dissolved oxygen availability for the growth of the nitrifying bacteria, especially in upper layers. Furthermore, periodic resting is a passive method for removing the microbial metabolites and restoring the hydraulic capacity and oxygen reduction potential (Desloover et al. 2011; Tsushima et al. 2014). Thus, intermittent operation was favorable for nitrifying bacteria growth and accumulation of NO3-N (Table 1), which was ready to be denitrified to N2O. However, when the drying time was extended, a more easily biodegradable portion of organic matter was quickly consumed by microorganisms in the first 10 cm of the layer where oxygen levels peaked (Li et al. 2011). Consequently, denitrification process would be suspended resulting from the absence of a carbon source.

The optimal drying-wetting cycle from the point of view of maximum infiltration along with the desired objective of quality improvement in the wastewater was found to be 12 h of flooding followed by 12 h of drying, i.e. drying-wetting cycle 1:1.

CONCLUSIONS

Influent loadings and drying-wetting regimes impacted N2O generation and conversion by affecting HRT and nitrogen circulation indirectly. N2O conversion ratio decreased with an increase in HL (decreased from 6.6 ± 0.1% to 2.7 ± 0.1% when HL increased from 0.08 to 0.24 m3/m2·d). N2O conversion rate increased with an increase in PL up to 8.2 ± 0.5% followed by a decrease to 3.8 ± 0.2%. Optimal drying-wetting cycle from the view point of maximum infiltration along with the desired objective of quality improvement in the wastewater was found to be 12 h of flooding followed by 12 h of drying. In order to abate N2O emission as well as improve effluent quality, optimal HL and PL were suggested as ranges of 0.08–0.12 m3/m2·d and 3.2–3.7 g N/m2·d, respectively. Under these conditions, conversion rate for influent nitrogen to N2O would be between 4.5% and 7.0%.

ACKNOWLEDGEMENTS

This work has been financially supported by the National Natural Science Foundation of China (no. 41571455 and 51578115) and Basic Science Research Fund in Northeastern University (N160104004).

REFERENCES

Chinese Environmental Protection Agency
2002
Methods for Water and Wastewater Analysis
.
Environmental Science Publishing House of China
,
Beijing
.
Desloover
J.
De Clippeleir
H.
Boeckx
P.
Du Laing
G.
Colsen
J.
Verstraete
W.
Vlaeminck
S. E.
2011
Floc-based sequential partial nitritation and anammox at full scale with contrasting N2O emissions
.
Water Research
45
(
9
),
2811
2821
.
Hernandez-Paniagua
I. Y.
Ramirez-Vargas
R.
Ramos-Gomez
M. S.
Dendooven
L.
Avelar-Gonzalez
F. J.
Thalasso
F.
2014
Greenhouse gas emissions from stabilization ponds in subtropical climate
.
Environmental Technology
35
(
6
),
727
734
.
Li
H. B.
Li
Y. H.
Xu
X. Y.
Wang
P. F.
Zhou
Y. C.
2014
Effects of intermittent operation mode on nitrogen removal by an overland flow system: a pilot study
.
Ecological Engineering
69
,
192
195
.
Rassamee
V.
Sattayatewa
C.
Pagilla
K.
Chandran
K.
2011
Effect of oxic and anoxic conditions on nitrous oxide emissions from nitrification and denitrification
.
Biotechnology and Bioengineering
108
(
9
),
2036
2045
.
Sun
S. C.
Cheng
X.
Sun
D. Z.
2013
Emission of N2O from a full-scale sequencing batch reactor WWTP characteristic and influencing factor
.
International Biodeterioration & Biodegradation
85
(
7
),
545
549
.
Tsushima
I.
Michinaka
A.
Matsuhashi
M.
Yamashita
H.
Okamoto
S.
2014
Nitrous oxide emitted from actual wastewater treatment plants with different treatment methods
.
Journal of Water & Environment Technology
12
(
2
),
191
199
.
Zapater-Pereura
M.
Lavrnić
S.
van Dien
F.
van Bruggen
J. J. A.
Lens
P. N. L.
2016
Constructed wetroofs: a novel approach for the treatment and reuse of domestic wastewater
.
Ecological Engineering
94
,
545
554
.