Bacterial Biomass Effect on Performance Improving of Vertical Flow Constructed Wetlands for Domestic Wastewater Treatment

Bacteria are frequently studied in constructed wetlands (CWs) due to their effective involvement in pollutants purication processes. In this study, aerobic, anaerobic and total bacteria densities and their vertical distribution prole within pilot-scale vertical ow CWs planted with different plant species were investigated. Five beds were planted in monoculture with Andropogon gayanus, Chrysopogon zizanioides, Echinochloa pyramidalis, Pennisetum purpureum and Tripsacum laxum, and one unplanted bed was used as control. At the end of the treatment trial, bacteria were collected by taking cores of sediment samples at the corners and the center of each bed following six layers in the vertical prole. In fact, the presence of plants on CWs improved the bacterial density and removal eciencies in the system, with yields from 5.9 to 24.1% regardless the pollutant. However, few anaerobic bacteria were obtained in the different wetlands, and unable to reduce NO 3− , excluding for beds planted with T. laxum and P. purpureum. Besides, the number of aerobic bacteria determined decreased (i.e., 17.4 10 6 to 0.1 10 6 CFU.g − 1 ), while that of anaerobic bacteria increased (i.e., 0.1 10 6 to 2.1 10 6 CFU.g − 1 ) from the upper to the bottom layers in the planted beds. Otherwise, anaerobic bacteria were more abundant in the control than in planted beds. Then, total bacteria were dominated by aerobic bacteria, and decreased from surface toward the bottom. As P purpureum promotes the best performance, CWs with this type of plant could be a cost-effective alternative method of treating wastewater. .L BOD


Materials
All the chemicals were purchased from Sigma Aldrich (France), and having high purity analytical properties. The chemicals and their concentrations used in this study were denoted in Table 1.

Experimental Set-up
The pilot was built with cement in the form of rectangular beds (i.e., length = 1.45 m, wide = 1.00 m, depth = 0.60 m and area of 1.45 m 2 ) (Figure 1), according to the description of Coulibaly et al. (2008a, b). The beds were lled from bottom to the top with 0.1 m gravel (5/15 mm) covered with cloth and 0.6 m white lagoon sand (mean sand diameter = 572 µm, uniformity coefficient = 0.4, porosity = 37.5%), previously washed to remove clay, loam and organic matter. Each bed was equipped with irrigation devices consisted of 6 Polyvinyl Chloride (PVC) pipes (length: 1.45 m; diameter; 0.008 m) containing 60 lateral holes (See Figure 1) to allow homogenous distribution of e uent on its surface. The bed slope was 1% oriented via PVC of 0.032 m diameter to drain out the e uent of bed. Treated water was collected at the beds' exit using tanks (100 L) thus to quantify it. The experiment was performed during seven (7) months from February to August 2017 on a pilot-scale located at NANGUI ABROGOUA University (Abidjan, Côte d'Ivoire). This zone is characterized by a humid tropical climate with an average temperature around 25 °C. The analyses were carried out both within the Laboratory of Environment and Aquatic Biology of NANGUI ABROGOUA University and the National Laboratory for Quality Assurance Testing, Metrology and Analysis of Côte d'Ivoire.

Plant selection
Five (5) (Nash, 1909) due to the economic interest of their biomass. Indeed, these plants are highly appreciated by agro-pastoralists, and also propitious to the Côte d'Ivoire climatic conditions. In addition to having a positive impact on wastewater treatment, they could generate further revenues that can support maintenance costs.

Synthetic domestic wastewater
Synthetic domestic wastewater was used in this study to avoid CWs clogging problems reported by Coulibaly et al. (2008a, b) and Ouattara et al. (2008), and minimize the fluctuation of pollutant concentrations in real wastewater used during each experiment. The composition of synthetic domestic wastewater was carried according to Rodgers et al. (2006), Healy et al. (2010) and Metcalf and Eddy (1991) methods. However, some modi cations were done in order to respect the characteristics (i.e. nitrogen (N), phosphorus (P) and carbon (C) concentrations) of domestic wastewater encountered in developing countries. The composition of the synthetic domestic wastewater used, was Chemical Oxygen Demand (COD), 5 day-Biochemical Oxygen Demand (BOD 5 ), Total Nitrogen (TN), Total Suspended Solids (TSS), Total Phosphorus (TP), with COD = 628 mg O 2 .L -1 , BOD 5 = 380 mg O 2 .L -1 , TN = 45 mg.L -1 , TSS = 300 mg.L -1 , TP = 12 mg.L -1 , pH = 6.7 -8.

Editing and experimental tests
Five (5) beds were planted with the plant seedlings (i.e., 9 plants m -²) spaced of 40 cm x 40 cm in monoculture and the sixth remained an unplanted control. These young plants were collected from nurseries established near the experimental pilot, and previously cut to 20 cm above the roots before bed's planting. Then, the planted beds were fed with tap water for one month to allow acclimatize. After the acclimation period, each bed was intermittently fed (3 days/week) with 23.64 × 10 -3 m.d -1 hydraulic loading of synthetic domestic wastewater over 6 months. The synthetic domestic wastewater tank was cleaned before and after feeding of beds to remove all the impurities settled. During the experiments, water samples of bed exits were collected each week, stored within an ethylene bottle at 4 °C to analysis. Finally, 24 samples were taken in each bed during the duration of the experiment. The growth response of the tested plant species was determined from stumps diameter and above-ground biomass of plants according to Ouattara et al. (2008). In fact, the plant aboveground biomass produced was harvested and weighed at the end of each two-month growth cycle as well as the plants stumps diameters measurement during the experiment. For the bed microfauna study, sampling consisted of a collection of sediments at the end of the treatment trial. This was performed in six sediment layers in the vertical profile, ranging from upper surface to the bottom of the beds as following: [0; 10  Where C i and C e are the influent and e uent concentrations (mg.L -1 ), V i and V e are the influent and e uent volume (L) in the CWs.
2.2.6. Sediment sampling and microbial biomass analysis Sediment sampling for bacteria analysis was performed by coring with PVC pipe (Φ = 16 mm). According to Puigagut et al. (2007), the surface of the beds was divided into three (3) equal sections for a better taken account of the bacteria distribution within the beds. In each section, three sampling points (one at each extremity of the bed, and one at the center) were uniformly distributed over the width of the reactors from which a composite sample of the sediment layer under consideration was formed. Thus, the samples were stored in jars at 2°C to analysis.
The analysis of the bacteria was carried out according to the technique of germs inoculation in Plate Count Agar (PCA) (Sirianuntapiboon and Jitvimolnimit 2007). In fact, 5 g of the substrate sample were suspended in a sterile saline solution (0.85% NaCl) of 50 mL and inoculated in triplicate onto PCA after stirring and sedimentation at room temperature. The aerobic germs were grown in a single layer of agar, whereas the anaerobic germs were within a double layer of agar. These germs were incubated at 37 °C for 48 h, and then the number of colonies formed were counted according to the international standard ISO 6222 (1999). The total number of bacteria in each sample was determined by adding the numbers of aerobic and anaerobic bacteria. The mean values of pH, dissolver oxygen (DO) and water volume at the inlet and outlet of all the beds (planted and unplanted) are denoted in table 3. As for pH, values obtained in the bed outlets were higher than those of the raw water. Moreover, the average pH values of the planted bed exits (between 6.92 and 7.17) were slightly lower than those of the unplanted bed (7.33). However, the sequence of pH mean values were between raw water and unplanted bed was ranked as decreasing in this order: wastewater pH (6.81) < pH (A. gayanus) (6.92) < pH (E. pyramidalis) (6.93) < pH (C. zizanioides) (7.05) < pH (P. purpureum) (7.06) < pH (T. laxum) (7.17) < pH (unplanted) (7.32). Besides, signi cant differences were observed between wastewater pH and those of planted beds, as well as those of the different beds between them (Mann Whitney test: p < 0.05).
DO values measured outlet the beds (i.e., 5.41 ± 0.9 and 7.53 ± 1.6 mg.L -1 ) were high compared with that of raw water (inlet) (i.e., 2.13 ± 0.6 mg.L -1 ), whereas those of the planted beds were greater than those of the unplanted bed. However, some significant differences were noted among those of the planted beds (Mann Whitney test: P < 0.05).
As for the water volume collected at the outlet of the beds, this remain less than that of wastewater applied (80 liters). Indeed, the average water volume collected in the unplanted bed (72.4 ± 1.9 L) was the highest and was followed by those of the different beds planted with C. zizanioides (62.2 ± 3.6 L), A. gayanus (60.3 ± 3 L), E. pyramidalis (58.6 ± 5.8 L), T. laxum (55.6 ± 3.8 L) and P. purpureum (54.2 ± 4.3 L).
Using ANOVA (p < 0.05), we observed signi cant difference between the unplanted bed and the planted beds, while those of the beds planted with P purpureum and T. laxum were of the same order of magnitude and signi cantly lower than water volume collected at the outlet of the other planted beds. gayanus, C. zizanioides and E. pyramidalis beds. Thus, the CWs removal e ciencies were ranked following order of performance: (P. purpureum) > (T. laxum) > (E. pyramidalis) > (A. gayanus) > (C. zizanioides) > (unplanted). However, the planted beds were signi cantly e cient than the unplanted beds (Mann Whitney test: p < 0.05). Moreover, the beds with P. purpureum and T. laxum were more signi cantly e cient than those of E. pyramidalis, A. gayanus and C. zizanioides (Kruskal-Wallis test, p 0.05).
To better insights the biological activity in the beds, the total bacteria density was assessed during the treatment trial ( Figure 3). In fact, the total bacteria density oscillated between 2.9 10 6 and 12.3 10 6 CFU.g -1 whatever the bed, while the median density of the bed planted with P.
To better understand the bacteria evolution within all the bed sediments, the vertical distribution of aerobic, anaerobic and total bacteria densities in the different sediment layers was investigated. Indeed, the gures 4, 5 and 6 showed the aerobic, anaerobic and total bacteria densities distributions, respectively. Overall, from the upper layer [0; 10 cm] to the bottom layer [50; 60 cm] of the beds, the number of aerobic bacteria decreased (17.4 10 6 to 0.1 10 6 CFU.g -1 ) (Figure 4), while that of anaerobic bacteria increased (0.1 10 6 to 2.1 10 6 CFU.g -1 ) ( Figure 5). However, the total bacteria density decreased from upper surface towards the bottom of the beds (1.5 10 6 to 17.4 10 6 CFU.g -1 ) ( Figure 6).

CWs performance
The increase of pH and dissolved oxygen (DO) at the beds' outlet compared to raw wastewater could be due to the consequence of the biodegradation of organic matter and/or the metabolism of nutrient assimilation by plant in the CWs (Koné et al. 2011). In fact, the absorption of nitrate ions through the roots of the plant is countercurrent to a transport of hydroxide ions (HO -) from the plant to the outside or a co-transport of hydronium ions (H 3 O + or H + ) within plant cells (Wegner, 2017). Hence, the release of the OHions in the medium during this process would also raise the pH of the bed ltrates. However, the high pH of the uncultivated bed compared to the planted beds could be due to the fact that the CO 2 resulting from the biodegradation of organic matter acidi es the medium in the presence of water, which could put back into solution the calcium and magnesium hydrogen carbonate content in the synthetic wastewater, adsorbed partly in the substrates. Therefore, the mineralization of these compounds would be increased the pH of the medium within unplanted beds. Finally, the difference of pH observed between the planted beds is due to the physiological speci cities of the plants used. On the other hand, the increase of OD in the bed ltrates would result from the aeration of the raw water during its application to the beds of the vertical ow wetlands and the oxygen released at the apex of the rootlets of the plants (Pérez et al. 2014).
The volume of water collected outlet the CWs beds was lower than the volume of wastewater applied, regardless of the beds considered.
According to Kadlec and Wallace (2009), this result is related to evaporation phenomena on the bed and evapotranspiration of plants, as well as the retention of a fraction of water in the bed sediments. Moreover, the difference of water volume between the planted beds is due to actual needs of each plant and the absence of plant within the control would increase the rejection of greater water volume in this unplanted bed compared to those planted.
Concerning the pollutants (COD, BOD 5 , TN, NH 4 + , NO 3 and PO 4 3-), high removal e ciencies were obtained within the bed ltrates, excluding NO 3 -, probably due to the type of wetland used in this study. Indeed, the type of vertical ow wetland used would favor aerobic and sedimentation and/or ltration mechanisms, which are the main removal mechanisms of theses pollutants in presence of the plants, whereas the degradation of NO 3 would need an anaerobic environment by appropriate organisms such as anaerobic bacteria (Norton 2014).
In fact, this type of system requires an intermittent feed of the wastewater, thus leading a higher recharge of the bulk of oxygen used for the metabolism of the bacteria during the biodegradation of these pollutants. In addition, lagoon sand of uniform particle size used in this study as sediment of the beds favor sedimentation and/or ltration mechanisms of the pollutant particles of the wastewater during the crossing of the bed sediments (Kadlec and Wallace 2009).
The improvement of removal e ciencies within the planted beds could be explained by the stimulating effect of the plants on degradation processes of the organic (COD and BOD 5 ) and nutrient (TN, NH 4 + , and PO 4 3-) pollutants through the secretion of root exudates (Weber and Gagnon 2014). In addition, the other removal pathways as the assimilation by macrophytes, the precipitation or the dissolution, and the adsorption in the sediment, would also contribute to further reduction of these compounds in planted beds (Vymazal 2007;Norton, 2014).
Moreover, the beds with P. purpureum and T. laxum were more e cient than those of E. pyramidalis, A. gayanus and C. zizanioides due, on In the case of this study, the bacterial densities obtained (from 2.9 to 12.3 10 6 CFU.g -1 ) was higher than those reported by Hatano et al. The signi cant differences in aerobic bacteria density between the rst two layers and that of the bottom in planted beds, and between the rst three surface layer and that of the bottom in unplanted bed could be explained by the abundant roots and macroinvertebrates in the surface layers, whose their activities would further aerate the super cial layers of the wetland (Ouattara et al. 2009, 2011). Moreover, the difference of bacteria density observed between the planted beds is due to from the morphology of the different plants in the wetlands.
Indeed, we note that the sequence of the density of the aerobic bacteria of the wetlands (P. purpureum > T. laxum > E. pyramidalis > A. gayanus > C. zizanioides) is relatively like that of the biomasses and plant stumps. Thus, these plant biomasses would have, depending on their importance, promoted a signi cant oxygenation of wetlands. This would justify the opposite sequences of anaerobic bacteria in planted beds and the fewest aerobic bacteria in unplanted bed.

Conclusion
The bacterial community investigated in CWs beds was dominated by aerobic bacteria. This study con rmed the positive in uence of plants on constructed wetlands performance, through the improvement of the bacterial density and removal e ciencies in the system (i.e., yield from 5.9 to 24.1%). However, few anaerobic bacteria were counted in the different wetlands, but were unable to reduce NO 3 outlet of them, except for wetlands planted with T. laxum and P. purpureum. Besides, the number of aerobic bacteria obtained decreased (17.4 10 6 to 0.1 10 6 CFU.g -1 ) in the wetlands, while that of anaerobic bacteria increased (0.1 10 6 to 2.1 10 6 CFU.g -1 ), from the upper to the bottom layers. Since P purpureum promotes the best performance, CWs with this type of plant could be a cost-effective alternative method of wastewater treatment.

Declarations
Funding Not applicable