The Feasibility of Maintaining Biological Phosphorus Removal in A-Stage via the Short Sludge Retention Time Approach: System Performance, Functional Genus Abundance, and Methanogenic Potential

The increasing concerns on resource and energy recovery call for the modification of the current wastewater treatment strategy. This study synthetically evaluates the feasibility of the short sludge retention time approach to improve the energy recovery potential, but keeping steady biological phosphorus removal and system stability simultaneously. SBRS-SRT and SBRcontrol that simulated the short sludge retention time and conventional biological phosphorus removal processes, respectively, were set up to treat real domestic sewage for 120 d. SBRS-SRT achieved an efficient COD (91.5 ± 3.5%), PO43−-P (95.4 ± 3.8%), and TP (93.5 ± 3.7%) removal and maintained the settling volume index around 50 mL/gSS when the sludge retention time was 3 d, indicating steady operational stability. The poor ammonia removal performance (15.7 ± 7.7%) and a few sequences detected in samples collected in SBRS-SRT indicated the washout of nitrifiers. The dominant phosphorus accumulating organisms Tetrasphaera and Hydrogenophaga, which were enriched with the shortened sludge retention time, was in line with the excellent phosphorus performance of SBRS-SRT. The calculated methanogenic efficiency of SBRS-SRT increased significantly, which was in line with the higher sludge yield. This study proved that the short sludge retention time is a promising and practical approach to integrate biological phosphorus removal in A-stage when re-engineering a biological nutrient removal process.


Introduction
The vast amounts of energy and material consumption caused by the huge scale and number of constructed wastewater treatment plants (WWTPs) has attracted society's concerns. Thus, technologies that focused on energy saving or recovery [1,2], resource recovery or utilization [3,4], and meeting the strict effluent discharge standards [5] were developed. The certification of "energy neutral" was achieved in the Strass Wastewater Treatment plant that operated an A-B stage process, in Austria, Leading to the reconsideration of the A-B process to modify current WWTPs.
Generally, the first stage is the extremely high load biosorption of the biological concentration of sewage with minimum organic oxidation degradation. The concentrated sludge is anaerobically digested for the recovery of energy content from sewage organics. Subsequently, a low load biological stage follows the first stage, operated to ensure the removal of dissolved organics and ammonia. This is the so-called A-B process due to the two-stage operation model [6]. Thus, biological phosphorus removal is limited in a typical A-B process. Additionally, low available organic carbon in the B stage limits nitrogen removal.
To meet the strict discharge standards, enhanced biological phosphorus removal (EBPR) processes, such as anaerobic/anoxic-oxic (AO), anaerobic-anoxic-oxic (AAO), or their modified operation mode, are widely applied in WWTPs [7]. Generally, sequencing anaerobic and aerobic conditions is necessary to achieve efficient biological phosphorus removal [8]. However, the competition for insufficient organic carbon sources between polyphosphate accumulating organisms (PAOs) and denitrifying bacteria inevitably occurs in the anaerobic phase of WWTPs [9]. This unavoidable drawback would be overcome if biological phosphorus removal was separated from nitrogen removal when re-engineering a typical EBPR process. Based on the multiple autotrophic nitrogen removal technologies development and investigationin recent decades [10,11], modifying the classical A-B stage process into an A-stage for C and P removal, with the B-stage for autotrophic nitrogen removal, is a promising approach.
The high-rate activated sludge technology (sludge retention time, SRT 0.5-2 d) was employed as an A-stage for partial organic carbon capture, mainly based on bio-sorption in the typical A-B stage process [6]. However, the high-rate activated sludge technology lost sight of phosphorus removal, and chemical phosphorus removal was required to compensate for this deficiency. Integrating biological phosphorus removal into the A-stage might be an applicable and promising exploration when reengineering one EBPR process into the A-B stage. Furthermore, it would be really unfortunate if biological phosphorus removal was abandoned when reengineering one EBPR process into the A-B stage. Thus, researchers investigated the feasibility of efficient biological removal under short or ultrashort SRT [12,13]. Chan et al. operated sequencing batch reactors (SBRs) to treat synthetic sewage, reporting that phosphorus removal and system stability deteriorated when SRT was shorter than 3 d [14]. Shao et al. [15] explored the dynamics of organic carbon and phosphorus removal in an ultra-short SRT SBR system treating real domestic sewage. However, the sludge volume index (SVI) of the SBR system with SRT at 3 d was above 150 mL/g, indicating the system was at the edge of instability. The studies were mainly concerned with the removal of COD and PO 3− 4 -P. However, TP removal and the ammonia preservation in A-stage determine the necessity of a chemical phosphorus removal section and the applicability of autotrophic nitrogen removal, respectively. The attentions should be paid to the removal of total phosphorus (TP) and the preservation of ammonia when performance in real domestic sewage treatment. Additionally, the study evaluated the integration of short SRT EBPR in the A-stage, with sludge yield rate (Y or Y obs ) or the ratio of the mixed liquor volatile suspended solids (MLVSS) and the mixed liquor suspended solids (MLSSs) [14,16]. Generally, waste activated sludge (WAS) anaerobic digestion (AD) is applied for energy recovery via biogas production in WWTPs [17], and the main objective of the A-B stage is energy and resource recovery. However, a few investigations evaluated the methanogenic efficiency of WAS from the S-SRT system by a long-term AD process.
The main objects of this study are to investigate the stability and the methane production potential of short-SRT EBPR systems acting as the A-stage. Two sequencing batch reactors (SBR control and SBR S-SRT ) are set up and operated for 120d. The oxic phase duration of SBR control is adjusted (from 1.5 h to 2.5 h) to achieve efficient phosphorus and ammonia removal, while the SRT of SBR S-SRT is adjusted from 5 d to 3 d. The overall COD, phosphorus removal efficiencies, and operational stabilities of the two SBR systems are studied; the microbial community and structures are measured to analyze the primary driving force corresponding to the reactors' operations. Two AnSBR are also set up to anaerobically digest WAS from the two SBR systems treating sewage. This study proves the feasibility of short-time EBPR as a practical tool for the modification of conventional biological phosphorus removal processes into the A-stage towards improving the energy recovery potential.

Domestic Sewage and Reactors
Domestic sewage was sampled from a sewer every day in the time period of 16:00-17:00. The sewage was collected from the second campus of the Harbin Institute of Technology and the residential quarters near the campus. A total of 40 L sampled sewage was passed through a 40 mesh sieve and stored in a 60 L influent tank. The main characteristics of the sewage sample are shown in Table 1. Two SBR reactors with a working volume of 9 L were set up in this study. The schematic diagram is shown in Figure 1. Two anaerobic SBR (AnSBR) were also set up to investigate the methanogen efficiency of WAS from the aforementioned SBRs. The working volume of the two AnSBRs were both 2 L.

Domestic Sewage and Reactors
Domestic sewage was sampled from a sewer every day in the time period of 16:00-17:00. The sewage was collected from the second campus of the Harbin Institute of Technology and the residential quarters near the campus. A total of 40 L sampled sewage was passed through a 40 mesh sieve and stored in a 60 L influent tank. The main characteristics of the sewage sample are shown in Table 1. Two SBR reactors with a working volume of 9 L were set up in this study. The schematic diagram is shown in Figure 1. Two anaerobic SBR (AnSBR) were also set up to investigate the methanogen efficiency of WAS from the aforementioned SBRs. The working volume of the two AnSBRs were both 2 L.

Reactor Operation
Two SBR reactors were operated for four cycles per day. In each cycle, the incoming influent and discharge of effluent were both 4.5 L (half of the working volume) to maintain the hydraulic retention time as 12 h. The SRT of the SBRcontrol was 10 d by discharging 0.9 L mixture sludge at the end of the oxidation phase each day, but the discharging volume of the SBRS-SRT reactors varied according to the pre-set SRT of each operation stage. Each cycle of SBRS-SRT consisted of 15 min inflow at the first 15 min of the anaerobic phase, 2.5 h anaerobic, 1.5 h oxidation, 30 min settling, 15 min discharging, and 75 min idle. The SBRcontrol operated with the same durations of inflow, settling, discharging, and idle phase, but the duration of the anaerobic and oxidation phases varied. The biomass inoculum for the two experimental SBRs was collected from the second settling tank of the Wenchang Wastewater treatment plant (Harbin, China). The collected sludge was passed through 20 mesh sieves to remove large particles and then, the MLSSs were measured. Certain volumes of sludge were inoculated to maintain the initial MLSS concentrations of the two experimental SBRs as 4.0 g/L. Mechanical stirring was employed to maintain the suspension of the mixed liquor. Two air diffusers were set at the bottom of each SBR and connected with a glass rotameter to control the aeration intensity at 0.8 L/min. The oxygen concentration was measured daily. Additionally, the dissolved oxygen concentration of the two experimental SBRs was maintained above 3.0 mg/L. Each SBR was equipped with

Reactor Operation
Two SBR reactors were operated for four cycles per day. In each cycle, the incoming influent and discharge of effluent were both 4.5 L (half of the working volume) to maintain the hydraulic retention time as 12 h. The SRT of the SBR control was 10 d by discharging 0.9 L mixture sludge at the end of the oxidation phase each day, but the discharging volume of the SBR S-SRT reactors varied according to the pre-set SRT of each operation stage. Each cycle of SBR S-SRT consisted of 15 min inflow at the first 15 min of the anaerobic phase, 2.5 h anaerobic, 1.5 h oxidation, 30 min settling, 15 min discharging, and 75 min idle. The SBR control operated with the same durations of inflow, settling, discharging, and idle phase, but the duration of the anaerobic and oxidation phases varied. The biomass inoculum for the two experimental SBRs was collected from the second settling tank of the Wenchang Wastewater treatment plant (Harbin, China). The collected sludge was passed through 20 mesh sieves to remove large particles and then, the MLSSs were measured. Certain volumes of sludge were inoculated to maintain the initial MLSS concentrations of the two experimental SBRs as 4.0 g/L. Mechanical stirring was employed to maintain the suspension of the mixed liquor. Two air diffusers were set at the bottom of each SBR and connected with a glass rotameter to control the aeration intensity at 0.8 L/min. The oxygen concentration was measured daily. Additionally, the dissolved oxygen concentration of the two experimental SBRs was maintained above 3.0 mg/L. Each SBR was equipped with an air pump. All phases of each SBR were managed by time controllers. The detailed operational strategies of the two SBRs are shown in Table 2. The operation SRT and temperature of the two AnSBRs were 20 d and 37.0 • C, respectively. The detailed operation parameters of the two AnSBR are shown in the Supplementary Material Text S1 ( Figure S1).

High-Throughput Sequencing Analysis
Sludge samples (100 mL) were collected from the two SBRs on the last day of each operation stage, and the inoculum WAS was sampled as a control. Metagenomic DNAs were extracted by using the 3S DNA isolation kit for Environmental Samples COD, NH + 4 -N, PO 3− 4 -P, total phosphorus (TP), MLSS, MLVSS, and SVI were measured daily, according to the standard methods (APHA 2012). The indexes of influent were measured without pre-filtration by 0.45 µm polyethersulfone filter. The generated biogas from the AnSBR was collected by a 2 L gas bag. The methane contents were measured by gas chromatography (Agilent 7890, Agilent Co., Ltd., City of Santa Clara, CA, USA) that was equipped with a flame ionization detector and a thermal conductivity detector.
As the energy or resource recovery is mainly based on the collected WAS, the accumulated WAS discharge per unit of COD (Y obs ) was calculated by Equation (1) in this study.
where V discharge (L) is the volume of discharged WAS; S t (g/L) is the concentration of MLSS or MLVSS in SBRs at time t; V inf (L) is the volume of influent sewage per day; COD inf (mg/L) and COD eff (mg/L) are the concentrations of COD in influents and effluents of SBRs. The methanogenic efficiency (MP) of WAS and total methane recovery potential (MP total ) of the two experimental SBRs were calculated according to Equations (2) and (3) as follows: where V biogas is the total volume of generated biogas at time t; R methane is the volume ratio of methane in generated biogas; S WAS is the concentration of inflow WAS corresponded to biogas production; and MP average is the average methanogenic efficiency of specified operation duration. Y obs of the same specified operation duration was calculated according to Equation (1), and R COD is the average COD removal efficiency.

Operational Stability of Two SBRs under Different Stages
The operational stability was assessed by the measurements of MLSS, MLVSS, SVI, and Y obs as shown in Figure 2. The average MLSS, MLVSS, SVI, MLVSS/MLSS, and the p-value (t-test) of the two SBRs during each stage are shown in Table S1. The average concentrations of MLSS in SBR S-SRT decreased with the decrease in SRT, as shown in Figure 2a. The average ratios of MLVSS/MLSS decreased from 60.2 ± 6.1% to 54.0 ± 8.0% with the SRT shortening from 4 d to 3 d in SBR S-SRT , which was not consistent with the study in [15]. As shown in Table S1, the average MLVSS/MLSS ratios of SBR S-SRT remain slightly higher than SBR control (54.0 ± 8.0% vs. 51.2 ± 6.3%, p = 0.10). This phenomenon might be caused by the difference and instability of sewage quality. However, a higher MLVSS/MLSS was beneficial to the sequential AD treatment of WAS [18]. biogas production; and MPaverage is the average methanogenic efficiency of specified operation duration. Yobs of the same specified operation duration was calculated according to Equation (1), and RCOD is the average COD removal efficiency.

Operational Stability of Two SBRs under Different Stages
The operational stability was assessed by the measurements of MLSS, MLVSS, SVI, and Yobs as shown in Figure 2. The average MLSS, MLVSS, SVI, MLVSS/MLSS, and the pvalue (t-test) of the two SBRs during each stage are shown in Table S1. The average concentrations of MLSS in SBRS-SRT decreased with the decrease in SRT, as shown in Figure 2a. The average ratios of MLVSS/MLSS decreased from 60.2 ± 6.1% to 54.0 ± 8.0% with the SRT shortening from 4 d to 3 d in SBRS-SRT, which was not consistent with the study in [15]. As shown in Table S1, the average MLVSS/MLSS ratios of SBRS-SRT remain slightly higher than SBRcontrol (54.0 ± 8.0% vs. 51.2 ± 6.3%, p = 0.10). This phenomenon might be caused by the difference and instability of sewage quality. However, a higher MLVSS/MLSS was beneficial to the sequential AD treatment of WAS [18]. The SV30 and SVI were measured to assess the sedimentation performance of the activated sludge in the experimental reactors and the results are shown in Figure 2c,d. The SV30 of SBR control significantly decreased when the duration of the oxic phase was extended from 2 h to 2.5 h, while the SV30 remained relatively stable with the shortening of SRT in SBR S-SRT . The SVI of both experimental SBRs remained around 50 mL/g, indicating the excellent sedimentation capacity of the activated sludge in this study. It was notable that the SVI of SBR S-SRT was higher than SBR control during operation stage III. It might be due to the high organic substance content and also indicated the higher metabolic activity of the activated sludge in SBR S-SRT .
The accumulated WAS discharged from SBR control and SBR S-SRT is demonstrated in Figure 2e,f. The corresponding Y obs of MLSS and MLVSS were calculated by Equation (1), and the effluent MLSS or MLVSS was not included to represent the total feedstock of the subsequent WAS disposal processes realistically in this study. The extension of the oxidation duration slightly affected the Y obs of the SBR control , which were 0.4257 g/gCOD (Y obs-MLSS-2.5h ) and 0.3808 g/gCOD (Y obs-MLSS-2h ) when the oxic phase durations were 2 h and 2.5 h, respectively. The shortening of SRT from 4 d to 3 d increased the Y obs from 0.4223 g/gCOD (Y obs-MLSS-4d ) to 0.4849 g/gCOD (Y obs-MLSS-3d ). The Y obs-MLSS-3d of SBR S-SRT was 27.3% higher than the Y obs-MLSS-2h of the SBR control and the Y obs-MLVSS-3d was 35.2% higher than Y obs-MLVSS-2.5h , indicating that more removed COD was converted into WAS. The high accumulation of WAS production of SBR S-SRT at SRT 3 d indicates the outstanding COD capture capacity of SBR S-SRT . It was beneficial to increase the energy recovery potential of the subsequent WAS AD process.

Sewage Nutrient Removal Efficiencies of the Two SBR Systems
The SRT of the SBR S-SRT and SBR control was 5 d and 10 d, respectively, during the initial 21 d to accumulate PAOs and to acclimate to the operation mode for the start-up. The nutrient removal efficiencies of SBR S-SRT and SBR control EBPR reactors were shown in Figure 3. The average nutrient removal efficiencies were shown in Table S2. The average COD removal efficiencies of SBR S-SRT and SBRcontrol under each stage were 90.0 ± 3.9% and 90.7 ± 5.4%, 91.1 ± 3.2% and 90.7 ± 3.5%, 92.0 ± 3.8% and 91.5 ± 3.5%, respectively (p > 0.05), as shown in Table S2. The shortening of the SRT affected the COD removal performance of SBR S-SRT negligibly, while the extension of the oxic phase duration slightly enhanced the COD removal efficiency of SBR control , as shown in Figure 3a. COD was mainly removed by the adsorption of sludge flocs and the metabolism of activated organisms. The previous studies confirmed that short SRT enhanced the adsorption of sludge flocs and accelerated the microorganism activities [15], which accounted for the steady and excellent COD removal efficiency of SBR S-SRT .
Ammonia removal was the premise step that supplied NO − x -N for denitrification to achieve nitrogen removal in conventional WWTPs. The oxic phase duration of the SBR control was adjusted and extended to achieve efficient ammonia removal. The ammonia removal performances of the two experimental SBRs are shown in Figure 3b. The ammonia removal efficiency increased from 35.2 ± 19.9% to 88.9 ± 5.4% with the increase in the oxic phase duration from 1.5 h to 2.5 h, as shown in Table S2. However, the ammonia removal efficiencies decreased from 25.9 ± 18.5% to 15.7 ± 7.7% (p < 0.05, Table S2) with the SRT shortening from 5 d to 3 d indicating that only a limited amount of NH + 4 -N was removed in SBR S-SRT . An SRT around 4.3 d and a low dissolved oxygen were beneficial for the nitrification and the accumulation of AOBs [19]. The short duration of the oxic phase in SBR S-SRT accounted for the inefficient ammonia removal. The activity of the AOBs was inhibited when the SRT of the SBR system was 3.5 d [13]. The nitrification was even undetectable in a continuous A/O process when the SRT was 4 d [17]. However, the SBR S-SRT offered a lower ammonia removal efficiency, which was mainly by microbial assimilation. Ammonia removal was the premise step that supplied NO − x -N for denitrification to achieve nitrogen removal in conventional WWTPs. The oxic phase duration of the SBRcontrol was adjusted and extended to achieve efficient ammonia removal. The ammonia removal performances of the two experimental SBRs are shown in Figure 3b. The ammonia removal efficiency increased from 35.2 ± 19.9% to 88.9 ± 5.4% with the increase in the oxic phase duration from 1.5 h to 2.5 h, as shown in Table S2. However, the ammonia removal efficiencies decreased from 25.9 ± 18.5% to 15.7 ± 7.7% (p < 0.05, Table S2) with the SRT shortening from 5 d to 3 d indicating that only a limited amount of NH -N was removed in SBRS-SRT. An SRT around 4.3 d and a low dissolved oxygen were beneficial for the nitrification and the accumulation of AOBs [19]. The short duration of the oxic phase in SBRS-SRT accounted for the inefficient ammonia removal. The activity of the AOBs was inhibited when the SRT of the SBR system was 3.5 d [13]. The nitrification was even undetectable in a continuous A/O process when the SRT was 4 d [17]. However, the SBRS-SRT offered a lower ammonia removal efficiency, which was mainly by microbial assimilation.
Both the SBRcontrol and the SBRS-SRT removed PO -P efficiently, as shown in Figure 3c. The average PO -P removal efficiency of SBRS-SRT was above 95% under both SRT 4 d and 3 d, as shown in Table S2. It was comparable with the S-SRT or traditional EBPR processes of previous studies [17,20]. However, the evident value difference between the PO -P and TP content in sewage indicated the necessity to pay attention to the TP removal performance of the SBRS-SRT and the SBRcontrol. Furthermore, TP was the control target specified in effluent discharge standards. The average TP removal efficiencies of the SBRS-SRT were 93.8 ± 3.9% and 93.5 ± 3.47% at SRT 4 d and 3 d, respectively, as shown in Figure 3d. The Both the SBR control and the SBR S-SRT removed PO 3− 4 -P efficiently, as shown in Figure 3c. The average PO 3− 4 -P removal efficiency of SBR S-SRT was above 95% under both SRT 4 d and 3 d, as shown in Table S2. It was comparable with the S-SRT or traditional EBPR processes of previous studies [17,20]. However, the evident value difference between the PO 3− 4 -P and TP content in sewage indicated the necessity to pay attention to the TP removal performance of the SBR S-SRT and the SBR control . Furthermore, TP was the control target specified in effluent discharge standards. The average TP removal efficiencies of the SBR S-SRT were 93.8 ± 3.9% and 93.5 ± 3.47% at SRT 4 d and 3 d, respectively, as shown in Figure 3d. The corresponding TP concentrations in the effluent were 0.32 ± 0.21 mg/L and 0.32 ± 0.16 mg/L, respectively, which met the first A class of Chinese standards (GB, 18918-2002, <0.5 mg/L). The TP in sewage consisted of both inorganic phosphorus (mainly PO 3− 4 -P) and organic phosphorus. Organic phosphorus was removed by biodegradation and biosorption. The different removal pathways of organic phosphorus and PO 3− 4 -P might account for the lower TP removal efficiency than that of PO 3− 4 -P. However, the chemically enhanced phosphorus removal section was non-essential in the subsequent B-stage. In conclusion, the low COD concentrations, high ammonia retained ratios, and excellent phosphorus removal confirmed that the effluent from the SBR S-SRT was adapted to the autotrophic nitrogen removal processes and the additional chemical phosphorus removal section was non-essential.

Abundance Variation of Functional Microbes Related to System Stability and Phosphorus Removal
Sludge was sampled on day 0 (inoculum), day 21 (SRT10d1 and SRT5d), day 84 (SRT10d2 and SRT4d), and day 120 (SRT10d3 and SRT3d) from two operated experimental SBRs for high-throughput sequencing analysis, and the average length of OTU was 416 with a coverage above 99% in this study. The analysis of microbial diversity and community structure was conducted after data extraction flat. The indexes that reflected microbial diversity are shown in Table S3. Microbial diversity reduced compared with the WAS, indicating the reconstruction of microbial communities.
The abundance of functional bacteria related to phosphorus removal and system stability is shown in Table 3 and the details of the bacteria with abundance higher than 1% are shown in Figure S2. The detected sequence numbers of the microbial related to nitrification are shown in Table S4. The key to maintaining the stability of an activated sludge process is to avoid sludge bulking. The enrichment of the filamentous bacteria was one of the most significant factors that caused sludge bulking. The excessive growth of Thiothrix, which is a typical filamentous bacterium, results in bulking the sludge in conventional activated sludge systems [21]. Thiothrix was detected in the short-sludge-age or ultra-short SRT EBPR systems reported by [13,15]. The abundance of Thiothrix was increased to 1.86% in SBR S-SRT after 120 d operation due to the shortening of SRT, but it remained stable and lower than 1% in the SBRcontrol. The disagreement between the enrichment of filamentous bacteria and the good settling ability of sludge in SBR S-SRT was mainly due to the relatively low abundance of Thiothrix (ranged as the twelfth most abundant bacteria in the sample SRT3d) and the quality of the influent. However, the results suggest that the SBR S-SRT had the risk of sludge bulking. Only a few sequences were detected and the sequence numbers of Nitrospira and norank_Nitrosomonadaceae were both zero in the samples SRT4d and SRT3d, as shown in Table S4. These dates suggest that the nitrifying bacteria were washed out from the SBR S-SRT completely, which is in agreement with the poor ammonia removal performance of SBR S-SRT . Tetrasphaera is an important PAO [22] and has a comparable contribution to phosphorus removal to Candidatus_Accumulibacter in the EBPR process [23]. Tetrasphaera was the dominant PAO in both the SBR S-SRT and SBR control , while the abundances of Candi-datus_Accumulibacter and Dechlorimonas [24] were lower than 0.5% in the collected samples. The abundance of Tetrasphaera in the sample collected from SBRS-SRT at SRT 3 d was almost seven times higher than the inoculum (6.29% vs. 0.81%). The shortening of SRT was beneficial for the enrichment of Tetrasphaera. The abundance of Tetrasphaera in the samples collected from the SBRcontrol decreased from 12.22% to 3.56% when the duration of the oxic phase extended from 2 h to 2.5 h. The extension of the oxic phase improved the ammonia removal that resulted in the enhanced competition of the carbon source between denitrification and phosphorus release in the anaerobic phase. This might cause the reduction in PAO Tetrasphaera. Hydrogenophaga [25,26] and Flavobacterium [27] were reported to have the function of phosphorus removal and were enriched in SBR S-SRT and SBRcontrol, respectively. The abundance of the main glycogen accumulating organisms (GAO) Candidatus_Competibacter [17] increased from 1.35% to 3.26% when the SRT was shortened from 4 d to 3 d. The wash-out of the nitrifying bacteria in the SBRS-SRT eliminated the competition between denitrification and phosphorus release processes in the anaerobic phase, which left a surplus carbon source that accelerated the enrichment of both GAOs and PAOs. However, the abundance of the PAOs was still significantly higher than the GAOs, which was in line with the excellent phosphorus removal performance of SBR S-SRT .

Methanogenic Efficiency of WAS from the Two SBRs
The methane production and methanogenic efficiency of two AnSBRs (AnSBR control and AnSBR S-SRT ) treated WAS from the corresponding two SBR is demonstrated in Figure 4. The methanogenic efficiency of the AnSBR S-SRT treated WAS was significantly higher than the AnSBR control . The WAS average methanogenic efficiency decreased from 166.7 ± 54.7 mL CH 4 /gSS to 136.0 ± 53.2 mL CH 4 /gSS when the SRT of SBR S-SRT shortened from 4 d to 3 d. This result was consistent with the decrease in MLVSS/MLSS ratios (from 60.2% to 54.0%), as previously discussed. However, the average methanogenic efficiency of the WAS collected from the SBR S-SRT (SRT 3 d) was 54.1% higher than the methanogenic efficiency of the WAS collected from the SBR control (the oxic phase duration as 2.5 h), as shown in Table  S5. As discussed above, the two SBRS achieved equivalent COD removal efficiency. Taking the increase in Y obs and comparable COD removal efficiency into an account, the methane recovery potential of SBR S-SRT at SRT 3 d was 95.1% higher than the SBR control with the oxic phase duration as 2.5 h, according to Equations (2) and (3). The S-SRT approach was an effective and promising selection for the enhancement of energy recovery.

Conclusions
Two EBPR SBR systems were set up to treat real domestic sewage and two AnSBRs were conducted to compare the methanogenic efficiencies of the generated (or collected) WAS from the two systems. A T-test was employed to evaluate the significant difference between the two SBRs at each stage. The SBRS-SRT achieved efficient and comparable COD, PO -P, and TP removal (p > 0.05) and the removal efficiencies were 91.5 ± 3.5%, 95.4 ± 3.8%, and 93.5 ± 3.7%, respectively, at a SRT of 3 d. The SBRS-SRT remained steady sludge settleability, which was comparable with the SBRcontrol (p > 0.05). The ammonia in sewage was retained with an ammonia removal efficiency of 15.7 ± 7.7% indicated the wash-out of nitrifiers in SBRS-SRT when the SRT was shorter than T4 d. The enrichment of Tetrasphaera and the stable abundance of filamentous bacteria were the source force that maintained the stability of the SBRS-SRT. The methanogenic efficiency increased by 95% when the SRT of SBRS-SRT was 3 d compared with SBRcontrol as its oxic phase duration was 2.5 h. This study proved that the S-SRT EBPR was a promising and practical A-stage selection that maximally utilized the biological phosphorus removal function when shifting a conventional WWTP into a "source and energy recovery factory" operated as an A-B Time Time

Conclusions
Two EBPR SBR systems were set up to treat real domestic sewage and two AnSBRs were conducted to compare the methanogenic efficiencies of the generated (or collected) WAS from the two systems. A T-test was employed to evaluate the significant difference between the two SBRs at each stage. The SBR S-SRT achieved efficient and comparable COD, PO 3− 4 -P, and TP removal (p > 0.05) and the removal efficiencies were 91.5 ± 3.5%, 95.4 ± 3.8%, and 93.5 ± 3.7%, respectively, at a SRT of 3 d. The SBR S-SRT remained steady sludge settleability, which was comparable with the SBR control (p > 0.05). The ammonia in sewage was retained with an ammonia removal efficiency of 15.7 ± 7.7% indicated the wash-out of nitrifiers in SBR S-SRT when the SRT was shorter than T4 d. The enrichment of Tetrasphaera and the stable abundance of filamentous bacteria were the source force that maintained the stability of the SBRS-SRT. The methanogenic efficiency increased by 95% when the SRT of SBR S-SRT was 3 d compared with SBR control as its oxic phase duration was 2.5 h. This study proved that the S-SRT EBPR was a promising and practical A-stage selection that maximally utilized the biological phosphorus removal function when shifting a conventional WWTP into a "source and energy recovery factory" operated as an A-B stage process.