Integrated Methods for Household Greywater Treatment: Modified Biofiltration and Phytoremediation

Most countries around the world have experienced water scarcity in recent decades as fresh water consumption has increased. However, untreated wastewater is routinely discharged into the environment, particularly in developing countries, where it causes widespread environmental and public health problems. The majority of wastewater treatment method publications are heavily focused on high-income country applications and, in most cases, cannot be transferred to low and middle-income countries. An experimental study was conducted to evaluate the performance efficiency of pilot-scale physicochemical and biological treatment methods for the treatment of household greywater in Jimma, Ethiopia. During the experiment, grab samples of greywater were taken from the combined treatment system's influent and effluent every 7 days for 5 weeks and analyzed within 24–48 hours. Temperature, DO, EC, turbidity, TDS, and pH were measured on-site, while BOD, COD, TSS, TP, TN PO4−3-P, NO3-N, NH4-N, Cl−, and FC were determined in the laboratory. During the five-week pilot-scale combined treatment system monitoring period, the combined experimental and control system's mean percentage reduction efficiencies were as follows: turbidity (97.2%, 92%), TSS (99.2%, 97.2%), BOD5 (94%, 57.4%), COD (91.6%, 54.7%), chloride (61%, 35%), TN (68.24, 42.7%), TP (71.6%, 38.7%), and FC (90%, 71.1%), respectively. Similarly, the combined experimental and control systems reduced PO4−3-P (12.5 ± 3 mg/L), NO3-N (4.5 ± 3 mg/L), and NH4-N (10.19 ± 2.6 mg/L) to PO4−3-P (3.5 ± 2.6 mg/L, 7.5 ± 1.6 mg/L), NO3-N (0.8 ± 0.5, 3.6 ± 2.3 mg/L), and NH4-N (7 ± 2.9 mg/L, 15.9 ± 3.9 mg/L), respectively. From the biofiltration and horizontal subsurface flow constructed wetland combined systems, the experimental combined technology emerged as the best performing greywater treatment system, exhibiting remarkably higher pollutant removal efficiencies. In conclusion, the combined biofiltration and horizontal subsurface flow constructed wetland treatment system can be the technology of choice in low-income countries, particularly those with tropical climates.


Introduction
Increasing population growth and industrialization have led to an exponential rise in global water pollution in recent decades [1]. It is expected that by 2050, the world population will have risen to approximately 9 billion people [2]. Both developed and developing nations experience this global population growth, though the latter is more signifcant. In turn, this leads to a rise in fresh water demand and a decline in fresh water resources due to population growth and improvements in people's daily life. Due to this, there is a rise in the amount of wastewater produced, a need for more sanitation methods, and an overall rise in environmental and public health issues [2,3].
Tus, currently, two-thirds of the world's population lives in water-stressed areas for at least one month out of the year and by 2025, half of the world's population will be living in water-stressed areas, increasing the demand for direct and indirect wastewater reuse for survival [2,4]. As a result, water scarcity, poor water quality, and water-related disasters in general have emerged as the most pressing global issues pertaining to current and future water resources [2,3].
Used water from a household, municipality, or industry that contains 99% water and 1% suspended and dissolved solids is referred to as wastewater [2]. Household wastewater is wastewater produced at the household level by human activity [5], which is categorised into greywater and blackwater [6]. Greywater is the wastewater produced in bathtubs, showers, hand basins, kitchen sinks, dishwashers, and laundry machines which accounts for the majority of the fow (50-80%) [2,7,8].
In developing countries, particularly in urban and periurban areas, greywater is frequently discharged untreated into street drains or onto open ground, primarily ending up in rivers, resulting in oxygen depletion, eutrophication, and microbial and chemical contamination of soil and aquatic systems [2,9,10]. In the absence of treatment, this greywater contributes an excessive and frequently unwanted amount of chemicals and pathogenic microorganisms, endangering the environment, food safety, and human health [11][12][13]. Each year, at least 1.8 million children under the age of fve die due to water-related diseases, accounting for roughly 17% of all deaths in this age group [2]. Tis suggests that the situation in developing countries is getting worse [2,14]. Hence, these pollutants should be removed before greywater is reused or released into the environment in order to protect the environment and public health.
Over the last two decades, numerous studies have been conducted to assess the energy, economic, and environmental benefts of recycling greywater through nature-based solutions (NBSs) such as constructed wetlands (CWs) and biofltration (BF) [15][16][17][18][19]. Tese technological advancements imitate how greywater is treated in nature. Terefore, one clever way to reclaim and lessen the detrimental efects of untreated greywater on the environment and public health is to encourage nature-based greywater treatment solutions.
Constructed wetland is the other recently proved and efcient greywater treatment technology that employs macrophytes such as vetiver grass [1,32]. In comparison to conventional treatment systems, constructed wetland is a nature-based greywater treatment system that has been efective, afordable, and environmentally sound for repurposing greywater [17,32]. Te vetiver system, which is based on vetiver grass (Chrysopogon zizanioides L. Roberty), has been successfully used as a phytoremediation tool to remediate greywater due to its demonstrated extraordinary and unique morphological and physiological characteristics [1,33]. When compared to other plants, vetiver plant was a candidate for removing a variety of pollutants, particularly N and P from greywater [1,10,17,34]. Te interaction of vetiver, microorganisms, and substrates results in natural processes (biological, chemical, and physical) that can remove pollutants from greywater [30].
Tus, the purpose of this study was to evaluate the treatment efciency of physicochemical and biological methods for removing pollutants from household greywater by combining biofltration with a constructed wetland planted with vetiver grass and to justify the potential efciency of such technology as an alternative treatment method in the climatic conditions of Ethiopia.

Study Area.
Te study was conducted in a single purposefully chosen household in Jimma town, Southwest Ethiopia ( Figure 1). Jimma town, the capital of the Jimma Zone, is located in Oromia Regional Administration, Southwest Ethiopia, at a latitude of 7°40′49″°N and a longitude of 36°49′18″°E; it is 354 kilometers from Addis Ababa and has an elevation of 1718 m. Te study area receives a mean annual rainfall of 1529 mm, which is caused by a long and short rainy season. Te town's average monthly temperature ranges from 11.7°C to 27.6°C [35].

Study Design and Period.
Experimental study design was conducted in Jimma University at Laboratory of Environmental Health Sciences and Technology from June to July 2021.

Experimental Set Up and Operating Conditions of the
Greywater Treatment System. Te combined pilot-scale modifed biofltration and horizontal subsurface fow constructed wetland system was installed in a single purposefully chosen household with seven inhabitants (Figure 2). Te specifc area was located at a longitude of 036°50′48.129″°E and a latitude of 07°40′58.759″°N, with an altitude of 1734 m ASL. Te variables considered for the combined system's design and construction were the required water quality, hydraulic loading rate (HLR), hydraulic retention time (HRT), substrate, and vegetation [36,37].

Layout and Confguration of the Pilot-Scale Treatment
System. Figures 2 and 3 depict the layout, confguration, and other characteristics of the combined biofltration and constructed wetland system, which are summarized in Table 1. A 200 L polyethylene fltration tank with an area of 1.8 m 2 (r = 0.23 m, h = 1 m) was installed as secondary treatment and flled with diferent layers of flter media [10]. Te sand and gravel flter media used in the fltration system were obtained from a quarry or gravel pit. Te other flter materials were collected from the environment because they are easily   accessible. Before flling the tank, the collected materials were thoroughly washed with tap water and dried in the sun.
At the bottom of the barrel, a cast iron pipe with 8 mm diameter holes spaced 20 mm apart was inserted. Ten, the bottom 20 cm layer was flled with drainage gravel aggregates of size 10 cm. Te second layer, 10 cm thick, was flled with 5 cm corn cob, which is an important biosorbent for removing toxic organic and inorganic pollutants from water. Te third layer was 10 cm thick avocado peel, with a size of 4 cm. Charcoal 3 cm in size was layered in the fourth layer, which was about 10 cm thick. Te top most layer of about 20 cm was 0.8-1 cm grain size sand. Te storage capacity above the sand was set to 30 cm in order to provide the required head to produce the design fow through the bed shown in fgure 2 [38][39][40].
Te fow rate of water through the fltration system was determined to estimate the capacity or amount of water that the flter could produce in liters per hour or liters per day. Te fow rate was determined using a 1 L measuring cylinder and a stopwatch (discharge rate). Based on the following equation, the average time taken by the water through the flter bed to fll a 1 liter measuring cylinder for three replications was determined [41].
where Q � fow rate, V � volume of the efuent, and t � discharge time through the flter media. Based on previous literature, a surface area of 0.65 m 2 ( Figure 3) was determined for the pilot-scale constructed wetland [42,43]. It was constructed from a board measuring 1.3 m in length, 0.5 m in width, and 0.40 m in depth and was aligned to 1% slope to maintain the hydraulic gradient [9,[43][44][45][46]. Te interior of the wetland was lined with highdensity polyethylene (0.1 mm) plastic sheet [46]. As a substrate, it was layered with sand and gravel [44]. Te bottom 20 cm was flled with 2-3 cm gravel aggregates. A 10 cm thick layer of 0.8-1 cm sand was flled in the middle, and the topmost 10 cm was flled with fne sand (0.1-0.4 mm). To avoid creating insect breeding sites, the water surface was kept 5 cm below the upper part of the substrate. In order to achieve a uniformly distributed fow, 10 cm gravel was placed in the inlet and outlet zones of each cell [44,[46][47][48].

Installation of Inlet and Outlet Pipes.
Polypropylene random (PPR) pipe with 3/4 inch diameter and 0.75 m length was installed to the inlet and outlet of each system to transport greywater from the treatment system's collection tank to the fltration and CW system ( Figure 3). Te greywater was then gravity-fed from the feeding tank to the fltration and wetland system. Te pipeline at the inlet of each biofltration and wetland system was ftted with a gate valve to control the fow of greywater and to prevent raw greywater from entering each CW during the hydraulic retention time (HRT). As a result, the amount of greywater entering the biofltration cell was adjusted using a stopwatch and a graduated glass container; i.e., the HLR at the biofltration's inlet was set to 0.4 m 3 /day [37].
At the CW inlet point, a 30 cm long PPR pipe with 7 mm diameter holes drilled every 5 cm was inserted. It was designed to ensure equal greywater fow distribution at the inlet of the horizontal CW, so that efuents from the fltration system run slowly through the porous fll media beneath the bed's surface in a more or less horizontal path until they reach the outlet zone. As an outlet pipe, a 1/2 inch PPR faucet was inserted 10 cm from the bottom of the CW. All of the pipes, valves, and the wetland system in general were checked twice a day to ensure proper greywater fow and system operation.

Experimental Plant Selection and Establishment.
In order to remediate greywater, the vetiver system, which is based on vetiver grass (Chrysopogon zizanioides L. Roberty), has been successfully used as a phytoremediation tool. Tis is because vetiver grass has been shown to have extraordinary and unique morphological and physiological characteristics when compared to other plants. It is the most efective species among the top three, including Phragmites australis and Cyperus alternifolius. Vetiver can withstand a wide pH range of 3.5 to 11.5, as well as extreme heat and cold (−14°C to 55°C) and salinity. Additionally, it was capable of absorbing signifcant amounts of potassium, phosphorus, and nitrogen. It has a high level of disease and pest resistance. Since it is sterile, it has no chance of growing into an aquatic weed [1,33]. Along with these qualities, this plant was chosen due to its high biomass, root density and depth, ease of propagation, capacity to absorb or transform pollutants, tolerance of eutrophic conditions, ease of harvesting, and potential for using harvested material [1,42].
We obtained 240 green, young vetiver grass samples from the Jimma University College of Agriculture and Veterinary Medicine nursery. Each grass had at least 2-3 well-developed leaves and roots. Five vetiver tillers were then planted in the substrate of the horizontal subsurface fow system (HSSFCW) with their roots and leaves pruned to 20 cm and 30 cm, respectively. Te second CW was left vegetated-free as a control to measure the efectiveness of the planted CW [46,49,50]. Te planting took place in April. Te wetland beds were flled with tap water immediately after the plantation was completed, and watering of the wetland with rain water and tap water continued for two months, until the wetland plants adapted to the environment and grew well [51]. Following that, after two months, the biofltration and CW were supplied with 100 liters of greywater from the receiving tank on a regular basis to monitor the combined treatment system's performance efciency [46]. Moreover, since it was a rainy season, covers were designed for HSSFCW to eliminate the efect of the rainfall throughout the experiment period.

Operation and Maintenance of the Biofltration System.
Regular maintenance is required for a proper functioning of the greywater flters. Clogging is a frequent issue with biofltration and can be caused by components of greywater, as well as any material that falls on the flter. In this case, the use of pretreatment is recommended. If clogging occurred, it could be remedied by carefully removing the cover layer(s) and mixing the flter media. Applying sand or gravel that has not been properly washed or that is of the wrong grain size may be the cause of persistent or frequent flter clogging. To ensure a good fow, the pipes should be regularly inspected [36,51]. For organic flters, service life is an especially crucial factor to take into account. Because of the variety of flters and raw greywater material, there is no exact or specifc rule for when flter media should be cleaned or reflled. However, greywater fltration by various types of flter media is checked and cleaned on average every 1-6 months and is maintained when signifcant head loss occurs, causing the flter to be out of service for 1-2 days [41,52,53]. UNEP/ WECF [51] recommended that the sand and gravel media have to be replaced every 5-10 years.

Operation and Maintenance of the Constructed Wet-
land System. Constructed wetland should be inspected on a regular basis to ensure their proper and ongoing operation. Pollutant removal in subsurface fow constructed wetland frequently relies on a diverse range of coexisting physical, chemical, and biological routes, which are critically dependent on numerous environmental and operational parameters [30,31].
Experience has shown that early failure is likely to happen on many sites without the implementation of a maintenance regime. Te problems that frequently occur include blockages of inlets and outlets, fow-regulating devices, siltation of storage areas, algal growth, and plant dieback. Depending on the pollutant loading, maintenance is expected to include cleaning or removal of sections of contaminated substrate and associated vegetation. If the vegetation has been destroyed, plant replacement may be required [42].
Once the plants are established, quick visual inspections are considered an important part of routine maintenance. Te primary goals of these inspections are to maintain the desired plant communities within the wetland and to identify any problems before they become major problems. In addition to this, weeds and any patches of permanently dead plants should be removed, and bare patches should be replanted with transplants from healthy areas of the wetland [54].
Tere is debate over whether it is appropriate to harvest plants from artifcial wetlands. Te study conducted by Ellis et al. [42] indicates that there is no enough information available to decide whether or not harvesting is preferable [42]. However, Paul Truong [55] recommended harvesting of plants planted for phytoremediation two or three times a year for biomass utilization purposes or to export nutrients. Plant harvesting can improve removal efciencies in treatment wetlands [56]. On the other hand, plant harvesting may also create an adverse impact on the biomass population attached to the plant roots, thereby decreasing the performance of constructed wetlands [57,58]. Terefore, a general answer cannot be given, but plants need to be harvested for biomass utilization purposes or if they interfere with the operation or the maintenance activities of the treatment system [10, 58, 59].

Productivity, Utilization Options, and Economic
Potential of Vetiver Grass. Vetiver is a highly productive plant species. Te dried, partially dried, or even fresh material that is obtained from the vetiver plant's harvested leaves, stems, and roots can be processed completely or in part. Vetiver's unprocessed products are used for animal feed (for dairy cows, cattle, sheep, horses, or rabbits), compost, thatching, agricultural use (as mulch), and biofuel. Te plant's semiprocessed by-products are used as a raw material for furniture, pressed-fber pots, ethanol production, botanical pesticides, hats, mats, and other handicrafts. Te fully processed products are used for industrial products as well as fberboard, herbal medicine, and essential oils [1, 10, 55, 60, 62].

Collection of Greywater.
Greywater for this study was collected from a nearby kitchen, laundry, shower, and hand basin, primarily early in the morning when cleaning in most households was at its peak. Te residents of the household were asked to manually dispose of greywater into the receiving tank through a 1 mm mesh sieve as primary treatment to prevent vegetable peels and food residue, plastics, papers, and hair from entering the treatment system [63]. Ten, equal volumes of 25 L of real greywater from the kitchen, laundry, and shower/hand basin were collected and mixed in the feeding tank. Te mixed 100 L greywater was then allowed to pass through a 3/4 inch PPR drain pipe into a 200 L polyethylene barrel flled with flter media [9,51] and fnally to the two mini-constructed wetland ( Figure 3). Te fltration valve was opened, and either the CW with vegetation or the control constructed wetland was opened to allow water to fow into the basin. When the CW with vegetation system pipe network was open, the control constructed wetland pipe network was kept closed, and vice versa. Initially, clear water from the overhead tank was fowed into the basin for two days prior to beginning testing for the horizontal system with greywater to ensure the overall system's functionality and to wash out the organic contents/debris within the flter media.

Sample Collection and Analytical Methods.
Water samples were taken by the grab sampling technique every week from the inlet and outlet of the compartments of the treatment setup to determine the water quality [64]. Te following sampling points were chosen based on the treatment system's confguration: sampling point 1 (raw greywater), sampling point 2 (after biofltration), sampling point 3 (after HSSFCW), and sampling point 4 (after HSSFCW, after control CW). A total of 20 samples were collected throughout the monitoring period after the hydraulic retention time of 3 days [65,66]. Temperature, dissolved oxygen (DO), pH, electrical conductivity (EC), total dissolved solids (TDS), and turbidity were measured on the spot with an HI 98290 multiparameter meter and an HI 7629829/10 probe (Hanna Instruments, Woonsocket, RI, USA). For at least 40 seconds, the probe was gently stirred in a 5 L bucket of water. Total suspended solids (TSS), biological oxygen demand (BOD 5 ), chemical oxygen demand (COD), nitrate nitrogen (NO 3 -N), ammonia nitrogen (NH 4 -N), total phosphorous (TP), orthophosphate, and chloride (Cl) concentrations were all determined at Jimma University Environmental Health and Science Technology Laboratory using American Public Health Association standard methods (APHA) for water and wastewater examination. Total nitrogen (TN), however, was determined at the Jimma Agricultural Research Center ( Table 2). As a result, with a thoroughly cleaned high-density polyethylene bottle (HDPE), 1 L of water for physicochemical and 300 ml of water in a sterile glass bottle for bacteriological analysis were sampled from each sampling point. In the case of physicochemical samples, the bottles were flled until they overfowed and then stoppered. Tese samples were sealed and labeled before being placed in an ice-flled cooler box and transported cool and dark to the lab, where they were stored at 4°C [67].
For treatment system efciency, the percentage removal efciency (%) of the pilot-scale treatment system was calculated for each of the wastewater parameters based on the mass fow diference between the efuent and infuent relative to the infuent [68].
where Ci � concentration of waste material in infuent and Ce � concentration of waste material in efuent.

Statistical Analysis.
Data entry and analysis were conducted using SPSS version 20 software. Signifcant diferences in infuent and efuent concentrations were determined using one-way ANOVA (analysis of variance) at p < 0.05 with α � 0.05. Mean and standard deviation determinations, as well as the analysis of variance method, were used to determine statistically signifcant diferences between the data. Te hypothesis of normality was verifed via the Shapiro-Wilk test. As the condition of normality was not met for at least one sampling point (

Modifed Biofltration.
In this study, a fltration system comprised of charcoal, avocado peels, and corncob was used in addition to sand and gravel.

Data Quality Management.
For each analytical experiment, a standard reference material was used for calibration and quality assurance. All of the chemicals and reagents used in the laboratory tests were of analytical grade and of standard approved make. To achieve the best possible results, standard solutions and necessary reagents were prepared on a regular basis. Reagent blanks and analytical duplicates for the wastewater samples were also included to evaluate the analytical reproducibility, accuracy, and precision. Laboratory instruments were calibrated in the same way, using freshly prepared standard solutions. All experimental development, calibrations, standard preparations, experimental methods, data generation, and activity documentation were carried out in accordance with the APHA standard guideline for wastewater analysis [67].

Ethical Consideration.
Ethical clearance was obtained from the Institutional Review Board, Institute of Health. Moreover, informed verbal consent was also received from the household where pilot setup was set up.

Results and Discussion
During the monitoring period, the pilot-scale biofltration treatment system was operated at HLR of 0.4 m 3 /day, resulting in a HRT of 0.5 day or 12 hours for a flter barrel volume of 0.17 m 3 (r � 0.23 m, h � 1 m). Simultaneously, the biofltration efuent rate was measured to be 0.2 L/min, resulting in 0.288 m 3 /day. Moreover, the performance of combined systems of biofltration and constructed wetland planted with and without vetiver grass in terms of contaminant removal was evaluated, based on physicochemical and biological parameters between infow and outfow. Te combined treatment system efciency was also evaluated over the course of 3 months of system operation in terms of the feasibility of reclaimed greywater for non-potable reuse (Table 3). It was also discovered that the vetiver grass displayed some common stress symptoms during the acclimatization period, such as leaf tip drying. Plant growth in the constructed wetland system, on the other hand, increased signifcantly when the system was flled with greywater effuent from the biofltration system and changed color from pale green to dark green. Tis is because the nutrient supply provided by the greywater efuent allowed the plant to grow taller and survive the experiment. Te plant can also withstand harsh climatic conditions as well as a wide pH and salinity range [1]. Table 3 and Figure  4, the mean temperature values for infuent and efuent recorded during the monitoring period were range-bound between 21°C and 23°C, which is the optimum temperature for bacterial activity in the treatment system [30,48]. Tis fnding is in agreement with the results reported by Haddis et al. [70] and Selemani and Njau [71]. On the other hand, the lowest mean temperature (21.87 ± 0.4°C) recorded at the HSSFCW efuent could be attributed to the long HRT in the HSSFCW as well as shading efect of the vetiver grass surrounding the wetland [72,73]. Tis low temperature result at the planted CW efuent is consistent with the results reported by Ling et al. [74]. Despite the fact that there was little diference in mean temperature across the treatment systems, the ANOVA test revealed that there was no signifcant diference in mean temperature (p > 0.05) across the combined treatment system categories. Tis trend might be attributed to the atmospheric weather conditions at the experimental site. Furthermore, the mean temperature score at the outlet of the HSSFCW treatment system was within the maximum allowable limit for efuent discharge into the environment. As a result, this could be regarded as a promising treatment option for environmental protection.

Physical Parameters. As presented in
Turbidity in water denotes the presence of organic, inorganic, and suspended matter which afects water transparency [75]. According to our fndings ( Figure 5), mean water turbidity in raw GW of 951.6 ± 66.3 NTU was reduced to 26.7 ± 4.9 NTU in the combined biofltration and HSSFCW system (p � 0.001). Te removal of this water turbidity was attributed to fltration and sedimentation,  which were facilitated by the flter media in the biofltration and the macrophytes roots, which reduce interspaces between gravel by forming dense flter media capable of removing suspended particles in the CW [30]. Tis may also be explained by the biofltration treatment system's flter media depth [76]. Te wetland was not fully utilized in removing these contaminants from the wastewater because the majority of the organic, suspended, and colloidal matter was fltered and left in the biofltration. In this study, high mean turbidity value (951.6 ± 66.3 NTU) was recorded during the monitoring period when compared to the results (444 NTU) reported by Oteng-Peprah and Nanne [77]. Te diference in the fndings might be due to the raw GW, which has the majority of its sources coming from the kitchen and laundry, as more turbid greywater is expected from these sources [77]. Tis fnding is in agreement with the results reported by Edwin et al. [78]. Te fnding also supports the conclusion that the source, lifestyle, and daily activities of the household have a signifcant impact on the characteristics of greywater [79,80]. Based on the fndings obtained, the turbidity concentration at the HSSFCW's outlet was within the US Environmental Protection Agency's permissible domestic wastewater discharge limit.
Te total suspended solids (TSS) in the infuent and effuent of each compartment of the combined treatment system were used to calculate the amount of accumulated solids expected to be on the substrate as well as on the root of the vegetation for each compartment of the combined treatment system [81]. Te TSS value (1073.5 ± 386.19 mg/L) recorded in the raw GW was found higher when compared to the results (537 mg/L) reported in [77]. Te washing of clothes, shoes, vegetables, fruits, tubers, and a variety of other items that may contain sand, clay, or other materials could explain the diference in results [77]. Biofltration and HSSFCW combination, on the other hand, resulted in TSS efuent of 10 mg/L over 3-day HRT, which is a value accepted by the US EPA standard for non-potable reuse of reclaimed domestic greywater. Analysis of variance revealed that, similar to water turbidity, the compartments of the treatment methods achieved a signifcant reduction in TSS (p � 0.002).
Moreover, TSS removal was very efcient and consistent within the bioflter (96%), as shown in Tables 3 and 4, demonstrating the importance of the biofltration as a secondary flter for TSS removal. Te combination of the biofltration with HSSFCW system in reducing TSS by 99.2% has proven to be very efective compared to the results of TSS (49%) reported in [82]. Tis value is also greater than that reported by Rahmadyanti et al. [83] who reported TSS removal of 90% in HRT of 3 days. Tis efcient removal of TSS from raw GW is critical for preventing clogging of the bed media in the HSSFCW system. Similar observation of mean TSS reduction was also reported by Ling and Apun [84] who studied bioflters combined with HSSFCW in the treatment of household greywater using ornamental plants in Kuching City, Malaysia. Te greater performance of HSSFCW over control CW in TSS removal ( Figure 5) could be attributed to the dense network of plant roots, particle sedimentation, and biodegradation of suspended particulates [30,75,85]. Table 3 and Figure 6. According to Oteng-Peprah and Nanne [77], pH is heavily infuenced by the constituents of greywater and is typically in the range of 5-9. In this study, the raw GW had an acidic pH (4.86 ± 0.2), when compared to the results (6.05 ± 0.26) reported by Pakanati Chandra Sekhar Reddy and Arun [86]. Te diference in results could be attributed to microorganisms decomposing the organic fraction of greywater. Tis acidic pH might have also resulted from greywater generated from the kitchen [87]. Bakare et al. [87] stated that kitchen GW had the lowest pH values when compared to other GW sources, which might be attributed to the rapid degradation of food particles and oils under anaerobic conditions.

Chemical Parameters. Variation of pH in diferent parts of the treatment methods is presented in
In the HSSFCW, however, there was a slight increase in mean pH of 6.68 ± 0.4 ( Figure 6). Similar observation of increase in mean pH was also reported by Ling and Apun [84] who studied bioflters combined with HSSFCW in the treatment of household wastewater using ornamental plants in Malaysia. Tis increasing trend of pH across the treatment system might be attributed to denitrifcation processes producing alkalinity [31]. As the mean concentration of DO was depleted (Table 3), the pH rose because denitrifcation took over and produced alkalinity. Tis pattern could also be attributed to the anaerobic decomposition of organic matter and the dissolution of bed granules in water [75]. Despite the fact that the mean pH score in the raw GW was 4.86 ± 0.2, the mean pH score at the HSSFCW outlet meets the US Environmental Protection Agency domestic wastewater discharge limit (p < 0.05).
Another important environmental parameter considered in this study was DO, which is important in wastewater treatment because it facilitates the degradation of organic matter by microorganisms [30]. As illustrated in Table 3 and Figure 4, the mean DO concentration in raw GW was low, since the raw GW had high organic content. However, the mean DO concentration increased slightly after passing of the raw GW through the biofltration, to 0.47 ± 0.2 mg/L, and rose continuously as it passed through the HSSFCW to 0.51 ± 0.2, where it gained more oxidizing potential, whereas, in the control CW, the mean DO concentration was lower than that of the HSSFCW (0.42 ± 0.1). Tese fndings are in agreement with those of Amiri and Tarik Hartani [67] who reported mean DO value of 0.26 mg in raw GW and 0.27 ± 0.86 mg/L in biofltration and HSSFCW combined system.
Moreover, because DO was consumed for the degradation of organic matter in the biofltration, the removal of these organic matter in the biofltration resulted in less demand for DO in the CW, and thus an increase in DO was observed (Figure 4). Similar observation of a slight increase in DO mean concentration was also reported by Amiri and Tarik Hartani [66] who studied bioflter combined with HSSFCW in the treatment of household wastewater using ornamental plants in Malaysia. Te observed decrease in mean DO from the outlet of the biofltration (0.47 ± 0.1) to the outlet of the control CW system (0.42 ± 0.1) could be attributed to anaerobic microbial degradation of organic matter in the control CW [30,66]. Te increase in mean DO concentration after passing through the planted CW (0.51 ± 0.2), on the other hand, could be attributed to oxygen supply from vegetation roots [30,88]. Moreover, intermittent feeding of the wastewater into the bed of the treatment system might have resulted in the reaeration of the system [89]. Despite the fact that slight diferences in mean DO concentration were noted among the compartments of the combined methods, the mean DO diference between these treatment methods was not statistically signifcant (p > 0.05).
Total dissolved solids describe all solids (including mineral salts) dissolved in water. As illustrated in Figure 7, a slight increasing trend of TDS was observed from the infuent of biofltration (655.2 ± 33.4 mg/L) through the HSSFCW (904 ± 1.1.9 mg/L) and control CW (863 ± 62.67 mg/L) systems. Tis increasing trend of TDS through the combined treatment system could be attributed in part to the mineralization of organics or ion dissolution during the degradation of water pollutants in the compartments of the treatment system [30,75,78]. Signifcant mean diference (p < 0.05) of TDS between the compartments of the treatment system was also noted. Te current study also revealed that the HSSFCW signifcantly reduced the mean chloride concentration of 170.55 ± 54.14 mg/L to 66.4 ± 11.21 mg/L, which could be attributed to fltration and sedimentation in the flter medium bed and HSSFCW's vegetation root network (p � 0.004).
In this study, the mean concentration of EC value in the raw GW was 1259.2 ± 107.3 (Figure 7). Tis result is in agreement with the results reported in [90]. Furthermore, the mean concentration of EC value increased through the compartments of the combined treatment system from the biofltration inlet to the HSSFCW and control CW outlets,  with the HSSFCW having the highest mean EC value of 1601 ± 305.8 S/cm (Table 3). However, no signifcant difference in EC was observed (p > 0.05) across the compartments of the treatment method. Te presence of high EC values throughout the feld test indicates that the household greywater was possibly saline, which could be harmful to the environment if discharged untreated. Tis can be attributed in part to the mineralization of organic matter and dissolution of bed granules [30]. Despite an increasing trend in EC mean concentration across the combined treatment system, the values remained within the USEPA's allowable discharge limit. Moreover, EC is a surrogate measure of TDS [91]. Te type and nature of the dissolved cations and anions in the water infuence the relationship between TDS and EC [92]. Te average TDS/EC ratio for most wastewater ranges from 0.5 to 0.9 [93]. In this study, the TDS/EC ratio in the raw GW was determined to be 0.52, which is in agreement with the results reported in [94]. Tis relationship is not directly linear because the conductive mobility of ionic species varies depending on the nature of soluble ionic components, their concentration, and the temperature of water [95]. In general, the TDS : EC relationship is given by the following equation: TDS � k * EC [91]. Te value of k will increase along with the increase of ions in water.
Te chemical parameters in greywater treatment systems are primarily made up of dissolved organic matter, which are expressed as BOD and COD. Tese variables are critical indicators of organic load in domestic greywater [95]. According to the fndings of this study, the BOD 5 and COD mean concentrations in the efuent of the combined biofltration and HSSFCW were 71.3 ± 26.5 mg/L and 109.5 ± 57.9 mg/L, respectively, after 5 weeks of combined treatment system operation (Figure 8). Similarly, the combined biofltration and HSSFCW's recorded mean percent reduction for BOD 5 and COD was 94% and 91.6%, respectively, at a DO mean value of 0.51 ± 0.2 mg/L (Table 4). Tese values are lower than the results of BOD 5 (99%) and COD (95%) at mean DO concentration of 3.4-4.6 mg/L reported by Ling et al. [74] who evaluated the performance of pilot-scale bioflters and constructed wetland with ornamental plants in greywater treatment. Te diference in the fndings could be explained by the flter medium used, organic matter content of the raw wastewater, and mean DO concentration available in the HSSFCW treatment system [96].
Furthermore, throughout the treatment system, mean COD concentration was found to be greater than mean BOD 5 concentration (Figure 8). Te presence of synthetic organic compounds found in domestic wastewater such as bleaches, surfactants, and beauty products, which can also be transformed into other by-products during the chemical and biological treatment of greywater, could explain the dominance of COD over BOD 5 [96].
Te average BOD 5 /COD ratio for most GW ranges from 0.31 to 0.71 [95,97]. In this study, the BOD 5 /COD ratio in the raw GW was 0.9, indicating that more than 90% of organic matter in the wastewater could be biodegradable. Tis result is in agreement with the fndings of Nyoman et al. [83]. In contrast to our fndings, Albalawneh and Chang [95] reported BOD 5 /COD ratio of 0.58 which was much less than our fndings. Tese disparities in results could be attributed to the characteristics and composition of the raw GW generated [96].
On the other hand, the COD : N : P ratio in raw GW in this study was found to be 100 : 1.1 : 1.6, indicating severe nitrogen defciency in comparison to the optimal values of 100 : 20 : 1 as reported in [97]. As a result, the raw greywater used in this study had a balanced carbon and phosphorus composition in terms of what was required for bacterial growth, but with limited nitrogen. Tis fnding is in agreement with the results reported in [78,98]. In general, the combined biofltration and HSSFCW efuent BOD 5 and COD levels met the discharge standard and reuse limit for restricted irrigation.
Nitrogen transformation in HSSFCW and biofltration is a multifaceted process that involves plant uptake, sediment adsorption, and microbial metabolism. According to Table 3, the mean concentration of TN in raw RG was 14.45 ± 13.3 mg/L, originating primarily from kitchen wastewater (Figure 9). Tis fnding is consistent with that of Boyjoo et al. [90]. Moreover, the mean TN reduction in biofltration, HSSFCW, and control CW efuent was 7.84 ± 2.8 mg/L, 4.59 ± 2.6 mg/L, and 8.28 ± 1.4 mg/L, respectively. In the biofltration, the mean percent reduction of TN was 45.7%. Tis result was higher than the results Similarly, the biofltration + HSSFCW combined treatment system reduced TN by 68.24% at a mean DO concentration of 0.51 ± 0.2 mg/L and at a mean temperature of 21.7 ± 0.56°C (Table 4). Tis is lower than the result (87.6%) reported by Amiri and Tarik Hartani [66] who assessed an integrated bioflter + CW system for household wastewater treatment in Algeria's arid regions. Te reason for this could be that nitrogen removal is frequently infuenced by nitrogen species transformation, vegetation, media, wastewater type, and hydraulic retention times [99]. Despite the fact that diferences in mean TN concentration were noted among the compartments of the combined methods, the mean TN diference between these treatment methods was not statistically signifcant (p � 0.187).
During the monitoring period, NO 3 -N was found to increase from 4.5 ± 3.3 mg/L in raw GW to 5.3 ± 2.5 mg/L in the efuent of biofltration, but it was found to decrease signifcantly in the efuent of HSSFCW (0.84 ± 0.5) (Figure 9). Tis pattern could be found in many studies on nitrogen transformation [100]. Te increasing trend of NO 3 -N at the biofltration flter outlet indicates that the nitrifying bacteria oxidized NH 4 -N to NO 3 -N [30]. On the other hand, according to Mtavangu et al. [75], nitrate removal in the HSSFCW is due to nitrifcation and denitrifcation, sedimentation and volatilization, and plant uptake; however, it is attributed to the denitrifcation and sedimentation process in the biofltration medium column. Similar observation was also reported by Ling et al. [74] who studied bioflters combined with HSSFCW in the treatment of household wastewater using ornamental plants.
Moreover, the reduction of NH 4 -N in the combined treatment system was attributed to microbial metabolism, plant uptake, and its sorption to the sediment of the flter bed [30,67]. Furthermore, unlike oxidized forms of N, ammonium nitrogen is removed from greywater via adsorption onto the wetland matrix's active cation exchange site [30].
Ye and Li [101] reported that DO concentration of ≥1.5 mg/L is essential for nitrifcation; however, denitrifcation occurs at ≤0.5 mg/L. Since the DO concentration in this study was 0.5 mg/L, denitrifcation was the dominant nitrogen removal process. As a result, it can be concluded that the main removal mechanism of organic nitrogen under anaerobic conditions (denitrifcation) in the treatment system was the reduction of NO 3 -N to molecular nitrogen (N 2 ) [102]. On the other hand, Yin et al. [100] stated that, at pH < 8.5, denitrifcation of gaseous products was the main route of gas loss which was similar to the present study. Te reason could also be attributed to the process of converting NH 4 -N to NO 2 -N, followed by the conversion of NO 2 -N to N 2 gas via the denitrifcation process [31]. Furthermore, NH 4 -N could be removed through volatilization from wastewater when the pH of a system is >9. However, because the pH level in HSSFCW was 6.68 ± 0.36, volatilization of NH 4 -N might not have been the removal mechanism in this study [102].
As illustrated in Table 4, the mean concentration of TP in the raw GW was found to be 21.12 ± 2 mg/L, owing primarily to the washing detergents. Tis is consistent with the fndings of Boyjoo et al. [90]. Furthermore, TP removal efciency in biofltration (39.1%) and HSSFCW (53.4%) was low when compared to the results reported by Amiri and Tarik Hartani [66] in the combined bioflter systems for the treatment of wastewater in Algeria. Tis low phosphorus removal in biofltration and HSFCW could be attributed to the fact that the media used in these treatment systems (gravel, crushed stones, charcoal, avocado peels, and corn cob) do not typically contain large amounts of iron, aluminum, or calcium to facilitate phosphorus precipitation and sorption [103].
When compared to HSSFCW, phosphorus removal in the wetland with no vegetation was also low ( Figure 10). As a result, the HSSCW removed the majority of the TP via sedimentation, vegetation uptake, precipitation, and flter material interception [30,103]. Kadlec and Wallace [30] stated that orthophosphate is removed through bacteria and plant uptake, while precipitation and adsorption are responsible for the removal of all forms of phosphorus. Barker [104] also stated that the increased polyphosphate- accumulating organism (PAO) population in the wastewater treatment system as well as the excess phosphorus uptake within each PAO reduces the phosphorus content of the water.
Furthermore, the combined biofltration and HSSFCW removed 71.6% of TP at mean temperature and pH value of 21.7 ± 0.6°C and 6.68 ± 0.36, respectively. Tis was less than the mean percent value of TP (87.72%) reported by Amiri and Tarik Hartani [66] in the household wastewater treatment by biofltration combined with HSSFCW. Te differences in the fndings could be attributed to diferences in the phosphorus concentration in the infuent to the beds, as well as diferences in the HRT, and material used to construct the beds [105].

Biological Parameters.
In terms of FC, the mean FC counts in raw GW were signifcantly higher than expected ( Figure 11). Te raw GW had a mean FC count of 2.44 × 10 6 CFU/100 ml, which was reduced by biofltration to 1 × 10 6 CFU/100 ml, and the HSSFCW further reduced FC count to 2.46 × 10 5 CFU/100 ml. Te presence of these bacteria in raw greywater could be caused by improper food handling in the kitchen, as well as direct contact with contaminated food, which has been identifed as a source of enteric pathogenic bacteria [77,106]. However, a signifcant reduction of FC (90%) was recorded at the outlet of the combined biofltration and HSSFCW (p � 0.001). Tis result is in agreement with the results of Ling et al. [74] and higher than the results reported by Nyoman et al. [82] who studied bioflters integrated with HSSFCW in the treatment of household wastewater.
Te high removal of FC (90%) by the combined treatment system depicts the importance in safeguarding the public health from pathogenic microorganisms ( Figure 11). According to Kadlec and Wallace [30], the mechanisms involved in the removal of FC are natural die-of, resource competition, sedimentation, predation, and photolysis. Because the compartments of the treatment systems were covered, photolysis was not the mechanism of removal of pathogenic microorganism in this study. However, adsorption and agglomeration in the bioflters could have resulted in signifcant fltration and sedimentation.
Moreover, because many enteric bacteria cannot survive long outside of host organisms, the efcient removal of coliforms by the HSSFCW might also be attributed to the long duration HRT [107].

Conclusions
In Ethiopia, where the practice of discharging untreated greywater onto open ground or into water bodies is widespread, the application of efcient, afordable, less energyintensive, and simple to operate greywater treatment methods is a major concern. In addition to being essential for environmental and public health protection, reusing treated greywater for non-domestic uses like irrigation, building projects, and ground water recharge can also open up additional opportunities.
Te results of this study provide an overview of how well the combined treatment system performs in tropical regions like Ethiopia. It is anticipated to advance knowledge of how the combined treatment system with vetiver plants and fll media operated under the local climate in Jimma. Te efect of pollutant loading on the treatment of domestic greywater is also important to consider. However, the treatment system's practical applicability is not as straightforward as claimed in various studies. Numerous issues could arise Journal of Environmental and Public Health 13 from failing to take into account even the most elementary technical tasks. According to the raw greywater characterized in this study, the high pollutant load values, especially when compared to US EPA guidelines, were found to be an important indicator of the pollutant's potential impact on the environment and public health. However, the combined greywater treatment system showed an outstanding performance in remediating turbidity, TSS, chloride, BOD, COD, TP, TN, and FC. Te efuent concentration of all greywater pollutants at the combined biofltration and HSSFCW complies with the unrestricted discharge limit standard; however, BOD, COD, TDS, and FC comply with the restricted discharge limit standard set by US Environmental Protection Agency.
When the performance of the combined biofltration and HSSFCW was compared to the performance of the combined biofltration and control CW pilot-scale systems, the pollutant removal efciencies of the combined biofltration and HSSFCW system were found to be more effcient than the biofltration and control CW combined system for all parameters during the monitoring period.
Because of the presence of aerobic and anaerobic phases, such a combination frequently optimized nitrogen and organics removal. Particularly, related to the biofltration treatment system, the wider implementation of organic flter media might have shifted the existing research from external carbon addition to internal carbon supplies, thereby improving denitrifcation performances. For the greywater composition studied here, the process was run for any strength of greywater generated during the monitoring period, showing that the combined treatment system can be a load-tolerant greywater treatment technology, making it suitable for decentralized greywater treatment and greywater reuse in developing countries.
In general, it can be concluded that the treatment performance of the pilot-scale combined treatment system applied at Jimma town was very promising for the promotion and application of treatment system as an alternative greywater treatment system to protect the environment and public health. Ethiopia has favorable climatic conditions for the implementation of treatment system, and hence the technology can continue as competent solution to alleviate the inherent environmental problems associated with discharging of untreated greywater. Terefore, it is recommended to use the pilot-scale combined treatment system for wider application in diferent towns for the treatment of greywater in small communities/institutions such as universities, colleges, military camps, and farms.
Te following suggestions and future research directions are provided based on the results obtained and the difculties encountered during the experiment period: (i) Managing household water pollution is a major concern for the government, policymakers, and the general public. Although there are a number of established and new methods for treating greywater pollution, ecologically based methods are largely preferred because of the broad acceptance among key stakeholders. (ii) Greywater must frst undergo pretreatment before entering the combined treatment system at the plot level. Screening systems should be implemented to prevent the clogging of the treatment system by vegetable peels and other coarse materials, which are frequently found in many low-income communities that do not segregate waste and have primitive disposal methods. (iii) Principles for choosing vetiver plants and media were very efective. If pretreatment systems to lessen the strength of the greywater and a fow management system are in place, vetiver's performance in contaminant removal in a tropical climate can be trusted. (iv) Despite promising eforts thus far, there are still limitations in certain areas to demonstrate the efcacy of the pilot-scale combined treatment system in the treatment of emerging chemicals that defy conventional remediation methods in order to establish acceptable remediation strategies and ecological benchmarks for greywater treatment optimization.
(v) More research should be conducted on the microorganism profle, the mechanisms that could explain plant growth (plant biomass, role of nutrients, nutrient and heavy metal status in the roots and shoots of vetiver, and water use efciency, among other things), and how toxic contaminants can be converted into less harmful substances to avoid pollution transfer from one source to another. (vi) It would be wise to conduct more research on the treatment system's long-term performance as well as the frequency of replacement and harvesting of vetiver plants. (vii) A brief three-month observation period was used for the study. It might not be all-inclusive. Terefore, more research should be done in different seasons while considering other factors, such as heavy metals and greenhouse gas (GHG) emission. (viii) Future research should also concentrate on examining diferent organic media, the proper ratio, and the flter life. (ix) Finally, it would be interesting to analyze the sustainability of the pilot-scale combined treatment system process and optimize it while taking the possibility of treatment system clogging into consideration.

Data Availability
Te data used to support the fndings of this study are available from the corresponding author upon request.

Disclosure
Te funder had no role in this study.

Conflicts of Interest
Te authors declare that they have no conficts of interest.

Authors' Contributions
Tolossa Waqkene was responsible for the study proposal, laboratory experiment, sample collection, data entry, data analysis, thesis writing, and manuscript preparation. Seid Tiku Mereta and Wuhib Zeine Ousman provided advice on the study proposal, assisted in laboratory experiment, and provided advice and assistance in manuscript writing. Amare Terfe was responsible for laboratory experiment, sample collection, and manuscript preparation.