Introduction
Anaerobic wastewater treatment is widely accepted among various technological options due to its advantages such as low energy consumption, low sludge production, tolerance to high organic loads, energy generation from produced biogas, and low space requirements (
de Lemos Chernicharo 2007). However, despite these advantages, there are certain drawbacks, including a long start-up process in the absence of adapted seed sludge, the need for post-treatment processes, requirements for odor control, unsatisfactory removal of nitrogen, phosphorus, and pathogens (
de Lemos Chernicharo 2007), and low performance at moderate to low temperatures, an aspect that still needs to be clarified (
Gomec 2010).
An average total chemical oxygen demand (TCOD) removal of 70% is expected in anaerobic treatments at temperatures above 20°C (
Gomec 2010). Temperature decreases affect the performance of anaerobic reactors since biological processes are slowed (
Nachaiyasit and Stuckey 1997); nevertheless, the development of several configurations has led to significant performance improvements at low temperatures. Some examples include the expanded granular sludge bed reactor, in which granular sludge develops at high up-flow velocities. When provided with a well-adapted, open form of granular sludge, the system allows high treatment levels for low-strength wastewater even at temperatures below 10°C (
Kato et al. 1994;
Rebac et al. 1995;
Van Lier et al. 1996). Another important configuration is the staged multiphase anaerobic (SMPA) reactor system, introduced by Lettinga et al. (
1997). The reactors arranged in series allow the growth of consortia of suitable anaerobic microorganisms at each stage depending on the available substrate and specific environmental conditions such as pH and partial pressure of
. Thus, the SMPA can prevent sludge in different compartments from mixing and may encourage the production of biogas, resulting in higher treatment efficiencies. Moreover, this system can treat industrial and domestic wastewater within a wide temperature range, i.e., from very low (
) to very high (55°C).
An anaerobic baffled reactor (ABR) follows the SMPA concept as it can separate acetogenesis and methanogenesis along the reactor (
Wang et al. 2004;
Liu et al. 2020). The ABR was first described by Bachmann et al. (
1985) as a series of up-flow anaerobic sludge blanket reactors (UASB) confined in several compartments using vertical baffles to direct wastewater through them. Accordingly, sludge bacteria inside the reactor rise and settle with biogas production, resulting in a high solid retention time (SRT) at relatively short hydraulic retention times (HRT). For municipal/domestic wastewater treatment, HRT typically varies from 6 to 48 h (
Bodkhe 2009;
Feng et al. 2008;
Hassan and Dahlan 2013;
Nasr et al. 2009). Within this HRT range, SRTs of up to 100 days are achieved (
Grobicki and Stuckey 1992). Furthermore, this fluid flow type reduces bacteria washout and enables the ABR to retain biological mass without using any fixed media.
Since the original design of ABR, numerous modifications have been made to enhance the efficiency and reliability of the reactor in treating industrial and domestic wastewater (
Hassan and Dahlan 2013;
Liu et al. 2010;
Zhu et al. 2014). Several studies have focused on the following factors that affect the performance of ABR: number of compartments (
Khalekuzzaman et al. 2018), hydrodynamics (
Khalekuzzaman et al. 2018;
Xu et al. 2014), and organic and hydraulic loading (
Yenji et al. 2021;
Xi-quan and Zhao-hua 2008), among others. Among these studies, the effect of temperature dominates. Temperature significantly influences anaerobic processes such as ABR. A decrease in temperature causes the slowdown of the degradation rates of volatile organic fatty acids (VFA) and causes the accumulation of soluble microbial metabolic products (SMP) that increases the COD values at the exit of the ABR (
Nachaiyasit and Stuckey 1997). Barber and Stuckey (
1998) and Zhu et al. (
2014) showed that the ABR performs well in treating medium- and high-strength soluble industrial and agricultural wastewater at mesophilic temperatures, with COD removal efficiencies exceeding 95%. Regarding the treatment of domestic or municipal wastewater, which is generally low-strength, COD removal efficiencies above 80% have been reported in the
interval (
Feng et al. 2008;
Nasr et al. 2009). Other studies have been conducted at a wider range of temperatures varying from psychrophilic to mesophilic (
Ayaz et al. 2015;
Feng et al. 2008;
Gomec 2010;
Hahn and Figueroa 2015;
Nasr et al. 2009;
Schalk et al. 2019), reporting variable COD removal efficiencies from 43% to 84%. At low temperatures, the efficiency of ABRs is mainly attributed to the retention of particulate organic material within the ABR (
Schalk et al. 2019).
Most studies were carried out at laboratory- or pilot-scale under controlled conditions and do not necessarily reflect the operational reality of large-scale treatment plants. For example, in an evaluation carried out by Yulistyorini et al. (
2019) on 89 treatment plants based on ABRs in Indonesia, only 14% showed acceptable performance, with biological oxygen demand (BOD) removal in the range of 25%–98% (mean of 74%), averaging
in the effluent, which does not meet Indonesian standards (
) for wastewater discharge. These results were related to deficient infrastructure and lack of maintenance; however, the deficient performance of other examples was attributed to design and operating parameters. In this context, it is important to investigate the performance of full-scale treatment systems under real operating conditions, i.e., wastewater and atmospheric temperatures and dynamic flow and load conditions.
The efficiencies achieved by ABR, especially in areas of moderate to low temperatures, are lower compared to other secondary treatment processes; in addition, they have poor performance in the removal of nitrogen, phosphorus, and pathogens, thus requiring a polishing treatment. Constructed wetlands (CW) integrated with other processes are considered promising complementary treatments due to their simplicity and low energy consumption (
Dornelas et al. 2009). Among these, the so-called hybrid CW, which are a combination of horizontal and vertical subsurface flow wetlands, have shown better performance than individual unit systems (
Otieno et al. 2017). These systems have been used in combination with ABRs, reaching up to 90%, 89%, and 80% in the removal of COD, total suspended solids (TSS), and TKN, respectively (
Ayaz et al. 2015;
Singh et al. 2009;
Ali et al. 2018;
Munavalli et al. 2022). Although the use of plants in wetlands brings advantages, especially in the elimination of organic nitrogen, gravel biofilters, which are essentially CW without vegetation, may be adequate to obtain effluents suitable for use in crop irrigation as they retain nutrients while also having lower maintenance needs. The use of gravel biofilters is uncommon in practice, and the performance of a system integrating ABR and hybrid CW with no vegetation remains poorly understood.
This study focuses on evaluating a WWTP composed of ABR and a no-vegetation hybrid CW comprising a sequence of horizontal and vertical gravel filters (HGF and VGF) to comprehensively understand the factors influencing its performance, such as temperature, HRT, organic loading rate (OLR), and on small-scale operation under dynamic conditions. The study area is located in Tolata, Bolivia, with a population of 2,705 inhabitants where the WWTP is operated using a management scheme. The results of this research have important implications for engineers and practitioners to make adjustments to future designs.
Treatment System
Tolata is located in a semiarid area of central Bolivia at 2,720 m above sea level. The annual average air temperature is 16.5°C, with an annual average minimum of 8.8°C and an average maximum of 27°C. The average annual rainfall is 457 mm (
SENAMHI 2018). The area’s economy is based on agriculture and farming, with maize and alfalfa as the main crops (
PDM 2007). The WWTP treats the wastewater received from the sewage system installed in the town of Tolata, which mainly has a domestic origin. It was built and is currently operated by the Aguatuya Foundation.
The WWTP is composed of two parallel treatment trains that combine sequentially an ABR, a HGF, and finally a vertical gravel filter. Both trains are preceded by a pumping station, a rotating screen, and a grease trap. The treated wastewater from both trains is collected in a chamber before entering a chlorination tank and then passes through a sand filter before being deposited in a storage tank for later use in irrigation (
AGUATUYA 2015). The whole process is designed to have an HRT of 31 h from influent to effluent. The treatment plant shown in Figs.
1 and
2 were designed to cope with a flow rate of
(
AGUATUYA 2017).
The pumping station is equipped with a grid chamber (placed before the pumping tank), where large solids are retained. This receives all the wastewater coming from the public sewers, which then rises it toward the rotating screen (RS). This station also functions as a flow rate equalizer; nonetheless, the flow entering the WWTP is less than the design flow, thus, pumping is intermittent. The RS retains solids
, which are later dried and transferred to a sanitary landfill. The wastewater then is conducted to the grease trap where fatty material is separated by natural flotation and removed manually. Anaerobic biological treatment is carried out within the ABRs, where organic matter is decomposed into simpler compounds, methane, and carbon dioxide under anoxic conditions. The process generates sludge deposition in the bottom of both ABRs, which is digested and periodically removed through relief valves. The ABRs were designed to have a retention time of 9 h according to recommended design criteria for decentralized wastewater treatment systems (
Gutterer et al. 2009). The walls and their baffles are built with fiberglass reinforced polyester (FRP) (
Echeverría et al. 2019).
Additional treatment is carried out in the HGF and VGF. At the outlet of each ABR, collecting chambers are arranged to direct the flow to the HGFs, with a total area of approximately
and a depth of 0.8 m. The walls and bottom of the filter are lined with high-density polyethylene geomembrane (
Echeverría et al. 2019). At the entrance of each HGF, the solid media is composed of coarse gravel, while the medium-sized gravel (
) in the treatment area has a mean porosity between 36% and 40% (
Delgadillo et al. 2010). The effluent of the HGF passes through an aeration chamber and is then sprinkled to the VGF through a perforated pipe installed at each entrance. Both VGFs also have an approximate area of
and are packed with medium-sized gravel (
Delgadillo et al. 2010).
The sludge accumulated at the bottom of the ABRs is pumped to a sludge drying area of
with a depth of 0.15 m. The sludge accumulated at the bottom of the ABRs is pumped to a
sludge dryer with a depth of 0.15 m. The sludge removal frequency is about 3–6 months. The sludge that dries in this area is transferred to a sanitary landfill located in the same place. Excess water from the drying area returns to the inlet of the plant (
AGUATUYA 2017).
Operational Conditions and Treatment Efficiency at the ABR and Biofilters
The operational conditions and efficiency of the ABR process obtained at each sampling campaign are summarized in Table
3. The OLR values of the ABRs analyzed in this investigation were found in the range of
. The
and HRT ranges were
and
, respectively. The average temperature range in this study was
. The TCOD removal corresponding to these operating conditions was
.
The treatment efficiency of an ABR depends upon several operational parameters, including inlet concentration, OLR, HRT, number of chambers, and temperature (
Reynaud and Buckley 2016), as it is discussed in the following sections. Furthermore, the up-flow velocity greatly influences solid retention within the ABR compartments containing sludge, thus influencing the efficiency (
Sasse 1998).
The effect of main operational parameters such as organic and hydraulic loading on the performance of ABR was evaluated at laboratory scale. Several studies, treating low to high-strength wastewater (
), reported efficiencies between 52% and 94% under a wide range of OLRs (
) and HRTs (
) within a mesophilic temperature range (
Bodkhe 2009;
Feng et al. 2008,
2015;
Koottatep et al. 2004;
Nasr et al. 2009). The studies carried out at this level generally do not take into account the effect of dynamic operating conditions and diurnal variations in the amount and concentration of wastewater. As an exception, Abbasi et al. (
2017) assessed the effect of fluctuations of OLR and HRT and seasonal climatic conditions on the performance of ABRs. They reported the highest COD removal efficiency as 72% in summer (
) and the lowest efficiency of 60% in winter (
), increasing the HRT from summer to winter from 72 to 120 h.
On the performance of ABRs at pilot and field scale (up to
), under dynamic operating conditions with natural flow fluctuation, Bugey et al. (
2011), Sibooli (
2013), Reynaud (
2015), and Yenji et al. (
2021) reported COD removal efficiencies in a range of 22%–90% for an OLR between 0.19 and
and HRT between 18 and 55 h. Particularly, Yenji et al. (
2021) reported a field-scale study for a wide range of OLR variation (
) corresponding to concentrations between 381 and
. The efficiencies ranged from 60% to 90%, further demonstrating that ABRs have a high capacity to withstand shock loads. Additionally, they found that the removal of COD and TSS increases with the strength of the wastewater. All these reports correspond to a mesophilic temperature range.
Our results were compared with those of Ayaz et al (
2015), Hahn and Figueroa (
2015), Schalk et al. (
2019), Saif et al. (
2021), and Pfluger et al. (
2018) who reported the efficiency of pilot-scale or full-scale ABR under moderate-low temperatures and dynamic operating conditions. These references were included in Table
3.
Effect of Organic Loading, Hydraulic Retention Time, and Up-Flow Velocity on ABRs Efficiency
The OLR of this research lies within the range reported for other full- and pilot-scale ABR (See Table
3) and below the maximum design COD load,
, recommended by Gutterer et al. (
2009). Concerning pilot-scale studies (ABR volumes between 0.7 and
) at low-moderate temperatures, Ayaz et al. (
2015) reported a COD removal efficiency of 50% under similar load conditions of
at a temperature of 12°C. Hahn and Figueroa (
2015) showed a lower efficiency of 43% in a temperature range of
for an OLR of
; Schalk et al. (
2019) reported a COD removal of 52% under an OLR of
at 15.4°C. Similarly, Pfluger et al. (
2018) showed a COD removal of 49% for an OLR of
in a temperature range of 14.8 to 20.5°C. All these COD removal efficiency results are comparable to those of our study (57%) under similar ranges of OLR and temperature.
The study of Saif et al. (
2021) corresponds to a larger reactor scale (
) comparable to the scale of our study (
). These authors report a COD removal efficiency of 40%–47% for an OLR range between 0.125 and
, treating a domestic wastewater with a COD inlet concentration between 104 and
under a temperature range between 13 and 34°C. This lower efficiency may be related to the low wastewater COD concentrations. In our study, a 57% COD removal efficiency was observed with higher wastewater concentration (
), as discussed by Yenji et al. (
2021).
The intervals of HRT and
reported here also satisfy the design limits recommended for ABRs: HRT
and
(
Gutterer et al. 2009). Additionally, our results are similar to the reference range of pilot and full-scale studies at moderate temperatures (
and
). In general, the operating conditions are similar to those of common practice in full-scale ABR.
In Fig.
6 the observed values of TCOD removal lie below the line simulated by the BORDA tool which relates TCOD removal with HRT and
at a mean ABR effluent temperature of 21.4°C. The observed values in the range of this study (
) were lower than those simulated at 21.4°C. Additionally, it is observed that the simulated efficiency reached its maximum value of 80% at an HRT greater than 20 h. This verifies that efficiency is more limited by temperature rather than HRT or
.
This comparatively low performance might be additionally explained by the strong influence of the hourly temperature variability and the occurrence of hydraulic shocks caused by intermittent pumping, preventing the and HRT from being uniform. These factors are described as follows.
Effect of Temperature Variability on ABRs Efficiency
In the study area, atmospheric temperatures were highly variable, ranging from 3 to 31°C during the monitoring period of October 2018 to July 2019 based on stations operated by SENAMHI (
2018). This variability shows the minimum, maximum, and mean atmospheric temperatures, together with the mean wastewater temperature at the ABR (Fig.
7). In addition, hourly variations in wastewater temperature at the ABRs are shown in Fig.
8.
The average wastewater temperature at the ABR does not show large variations (), actually, it fluctuates in a narrow range within the psychrophilic and mesophilic temperatures. However, large variability of temperature at the ABR during the day was found, 11 to 27°C from 8 a.m. to 3 p.m. Nocturnal monitoring would make this temperature range even wider. Under these conditions, it is expected to occur even wider variations of the microbial activities would affect COD removal.
In our study, the clear dependence of efficiency on temperature is considerable (Fig.
9) is noticed. Similar behavior was reported by Saavedra et al. (
2019) for a UASB-HGF configuration located in the same geographical zone and under similar climatic conditions.
A graph of efficiency as a function of temperature that confirms this correlation is shown in Fig.
10. The additional dotted line represents the expected efficiency in the same temperature range according to the design criteria of the BORDA guide.
The graphs show similar linear trends in the dependence between efficiency and temperature, as evidenced by the slopes. It can be noticed that the measured efficiency is lower than the simulated efficiency. Although this correlation is valid only for this particular WWTP, it provides a useful reference to estimate the efficiency of ABRs under similar operating conditions.
Effect of Flow Pattern on ABRs Efficiency
The flow rates presented in Table
3 represent the average of eight campaigns. In each campaign, the flow measurement was performed hourly for 8 h. In the actual operation of the WWTP, the input flow to the ABR is nonuniform since the wastewater is received in a lift station and then intermittently pumped to the RS and degreaser, from which it is distributed to both parallel ABRs. The average flow received by the lifting station is
and the pumping flow is 4 to 6 times higher; thus, the ABR receives an intermittent flow of wastewater and experiences hydraulic shocks. Intermittent flow and low temperatures can cause the occurrence of hydraulic dead spaces, thus decreasing the actual HRT as discussed by Haque and Hasan (
2018). Moreover, the number of compartments can be related to a decrease in the presence of dead spaces. The ABR of this study has three compartments that are less than the number recommended by Xu et al. (
2014) (four or five compartments) to achieve optimal and economic performance.
Additionally, to prevent wastewater flow from being uniform, this pumping regime causes the drag of particulate organic material from the lift station tank to subsequent units, overloading the pretreatment units and, consequently, the ABR. This solids overload can affect ABR performance in terms of COD removal. Some studies of full-scale treatment plants have concluded that an ABR’s operational performance, especially in combined collection systems (as presented here), depends on the efficiency of the pretreatment units. Inefficient pretreatment may affect the management of solids in the ABR, requiring a greater effort and frequency of maintenance for desludging (
Schalk et al. 2019).
The transfer of the substrate to the microorganisms, uniformity of environmental factors, and effective use of the reactor volume are guaranteed when the flow pattern is uniform (
Xu et al. 2014). In this study, uniform flow is not accomplished. This condition, together with the variability of temperature, are possibly the causes of the suboptimal COD removal. Under constant flow conditions, the granular biomass would not suffer disturbances, disintegrations, and washouts, so greater total COD removal is likely.
Operational Parameters and Efficiency of Horizontal and Vertical Gravel Filters
The superficial loading rate (SLR) that HGF and VGF received was in the range of
and
, respectively. An HRT of 86 h for HGF and VGF was registered as is shown in Table
4. Removals of 59%, 79%, and 16% of COD, TSS, and
in HGF and 44%, 46%, and 32% of COD, TSS, and
, respectively, in VGF corresponding to these operational conditions were reported in this study.
The use of combined horizontal subsurface flow and vertical flow CWs (called hybrid CWs) has proven to be very advantageous in combination with anaerobic processes. Anaerobic treatment prevents biofilters from clogging, whereas horizontal flow wetlands remove organic matter and vertical ones provide nitrification (
Ayaz et al. 2011,
2015). The observed efficiency values of this research were slightly lower than those reported by Singh et al. (
2009), who evaluated a treatment system configured by ABR and a hybrid CW, a similar configuration of Tolata´s WWTP, with the difference in the hybrid gravel biofilters that are unplanted CWs. In the horizontal subsurface flow wetland, Singh et al. (
2009) reported efficiencies of 51%, 69%, and 24% of COD, TSS, and
, respectively; the equivalent removal values in vertical flow wetland were 46%, 58%, and 70.9% of COD, TSS, and
, respectively. These efficiencies are comparable to our results except for the removal achieved in the vertical flow wetland, which is greater in planted wetlands than in biofilters. Also, removals in the range of 77%–94% COD, 81%–96% TSS, and 74%–99%
have been reported in hybrid CWs by Sayadi et al. (
2012). Planted CWs are more efficient at removing pollutants (
Dornelas et al. 2009), especially nitrogen, likely since plants regulate biochemical pathways by increasing oxygen supply (
Paranychianakis et al. 2016). Since these effluents are used for irrigating crops, it is desirable to conserve nutrients; in this regard, gravel biofilters are a more economical option concerning operation and maintenance as a polishing treatment. Additionally, 72-h HRT is recommended for the operation of hybrid systems for economic and technical reasons (
Cui et al. 2006). This difference may indicate that increasing the treatment capacity of these filters may be possible by increasing the flow rate.
Conclusions
This study focused on evaluating the efficiency of Tolata’s WWTP, located in the high valley of Cochabamba, which is a semiarid area with large daily temperature fluctuations, where water is a limited and valuable resource for crop irrigation. Thus, evaluating the feasibility of combined ABR-HGF-VGF treatment to achieve acceptable standards for reuse is of particular importance.
The global efficiency results obtained from October 2018 to July 2019 were as follows: 92% of TCOD, 82% of SCOD, 98% of PCOD, 98% of TSS, 49% of , and 31% of P. The effluent concentrations were , , , , and which, except for the nutrient content, comply with parameter thresholds specified in the Bolivian legislation for WWTP discharge. This WWTP showed highly efficient removal of , COD, and TSS. The ABR and horizontal-vertical gravel filter stages contributed the most to the global efficiency.
The simulated efficiency of the ABR under the studied temperature range and operational conditions (HRT, OLR, and ) were calculated using the BORDA tool. An efficiency of 57% was observed for a temperature range of , lower than the simulated 78% efficiency at a mean temperature of 21.4°C for optimal ABR performance, indicating that temperature limits the efficiency the most. Another factor that may affect the performance is the intermittent flow regime due to on-and-off pumping.
In order to improve the performance of the treatment, we recommend the following actions. First, providing a more continuous flow into the system to obtain a constant up-flow velocity within the ABR could improve residence time. Another factor that could contribute to ABR efficiency is the implementation of a primary settler, considering the use of a chemically enhanced solid separation process that could reduce the inflow of inorganic and organic particulate matter. The frequent maintenance of all treatment units is an important factor to reach the optimal performance of all components of the system. Due to the content of nutrients in wastewater, we recommend its application in crop irrigation limiting to tall-stem crops or those that are not consumed unprocessed in order to avoid the implicit microbiological risks.