VEGETATION EFFECT UPON MACROINVERTEBRATE COMMUNITIES IN A VERTICAL-FLOW CONSTRUCTED WETLAND TREATING DOMESTIC WASTEWATER.

Franck Michaël Zahui, Jean-Marie Pétémanagnan Ouattara and Lacina Coulibaly. Laboratory of Environment and Aquatic Biology, Department of Sciences and Environment Management, Nangui Abrogoua University, 02 BP 801 Abidjan 02, Côte d’Ivoire. ...................................................................................................................... Manuscript Info Abstract ......................... ........................................................................ Manuscript History

Constructed wetlands play an important role in wastewater pollutants removal. Thus, several investigations have been carried out about the different components of the system in order to acquire scientific knowledge on the pollutants removal processes that take place there. However, there is little knowledge about the biodiversity that can be found in these artificial environments, especially concerning the invertebrate community level. Our study investigate the effect of different plants species (i.e. Andropogon gayanus, Chrysopogon zizanioides, Echinochloa pyramidalis, Pennisetum purpureum, and Tripsacum laxum) upon macroinvertebrates communities and their abundance in the sediments of a vertical flow constructed wetland treating synthetic domestic wastewater. The experiment was performed in a pilot-scale composed of six beds constructed with bricks. Each bed was filled from the bottom to the top with 0.1 m of gravel covered with cloth and 0.6 m of white lagoon sand. Five beds were transplanted wile one unplanted was used as control. 80 L of domestic synthetic wastewater was then applied on the beds intermittently during six months. All the five plant species have well developed, improving the removal of total nitrogen (TN), total phosphorus (TP) and total suspended solid (TSS) from domestic wastewater in the constructed wetlands. However, P. purpureum provided highest above ground biomass and favored the best TSS, COD, TN and TP removal efficiencies. As for macroinvertebrates, 12 taxa belonging to 12 families, 6 classes and 6 orders were recorded. The community structure was influenced by plant species. Macroinvertebrates were significantly more diversified in the planted beds than in the control. Insects dominated all the planted beds, while Oligochaeta were the most abundant in unplanted bed. The bed planted with P. purpureum housed the largest number of organisms and was follow by those planted with T. laxum, E. pyramidalis, A. gayanus and C. zizanioides, respectively. Following the vertical profile, macroinvertebrates abundance decreased from the upper layer to the bottom layers and only Oligochaeta (Lumbricidae) were recorded in all beds sediments layers of the CW. Natural wetlands have historically been used as receiving bodies for effluent discharges. However, their use was more like a practical spill inducing the degradation of wetlands than an optimized treatment system. Constructed wetlands (CWs) for wastewater treatment were born from the improved knowledge of the functioning of natural wetlands, and Dr. Käthe Seidel was the first to experiment with this type of field in 1960, in Germany (Molle, 2012). This process, inexpensive, efficient and easy to perform compared to conventional sewage treatments systems (e. g., activated-sludge systems), has been proven itself in various parts of the world in the protection of environment and populations health (Molle et al., 2004, Lachat et al., 2005, Vymazal et al., 2006. Depending on the type of design of the CWs, the applied wastewater flows essentially, vertically or horizontally into the beds during the treatment process. However, in conventional horizontal flow wetlands, the oxygen concentrations are limited (Nivala et al., 2013) while unsaturated vertical-flow wetlands operate under a regime mainly aerobic. This condition should favor the oxygen transference rates (Schwager & Boller, 1997;Nivala et al., 2013) and then can enhance the nitrification rates as well as the oxidation capacity in order to remove organic matter (Cooper, 1999;Zhang et al., 2015). In addition, the use of vertical-flow CWs reduces environmental risks such as the release of odors and methane and the proliferation of mosquitoes, which are attributed to horizontal-flow CWs (Mander et al., 2005).
Unlike most treatment processes that focus on a single treatment mechanism or a type of pollutant, wetlands use numerous interdependent symbiotic processes for concurrent removal of several different types of pollutants. Indeed, these process are summarized in a set of physical, chemical and biological mechanisms, more or less intertwined, via the main components of the system namely, plants, sediments and fauna (micro and macro fauna) living in the sediment (Poulet et al., 2004). However, if sediment and plant can be manipulated during the construction of the wetland, the fauna develops itself naturally in the sediment of the beds (Stottmeister et al., 2003, Vymazal, 2008. Thus, in the quest for scientific knowledge on the pollutants (i. e., organic matter, nitrogen and phosphorus, pathogenic germs, heavy metals, etc.) removal processes in CWs, several investigations have been carried out about the different components of the system (Wu et al., 2015). However, if more attention has been paid to plants (Brix, 1994, Stottmeister et al., 2003, Gagnon et al., 2007, Coulibaly et al., 2008a, 2008b, Vymazal, 2011, Wu et al., 2011, 2013a, 2013b, Shelef et al., 2013, to sediments (Prochaska & Zouboulis, 2006, Xu et al., 2006, Calheiros et al., 2008, Bruch et al., 2011Saeed & Sun, 2013), and microorganisms (Münch et al., 2005, Tao et al., 2006, Gagnon et al., 2007, it has not been the same for macroinvertebrates. However, symbiotic relationships have been established between macroinvertebrates and microorganisms in the transformation of organic matter and nutrients in soils (Lavelle & Gilot 1994, Lavelle 1997).
Organic pollutants of wastewater and plant residues in the CWs bed sediments are composed mainly of biopolymers that assimilation by microorganisms proves difficult (Park et al., 2009). Therefore, these are first ingested and decomposed by macroinvertebrates into simple elements, making them easily assimilated by microorganisms. In addition, macroinvertebrates can create tunnels in the beds substrate that could increase its aeration and the water infiltration rate (Ouattara et al., 2009;2011). In short, macroinvertebrates could create specific micro-habitats in the beds of constructed wetlands and favored the proliferation of the microorganisms, which are mainly involved in the degradation of pollutants in these systems (Lavelle, 1997;Deprince, 2003;Stottmeister et al., 2003). Therefore, knowledge of the ecology of this fauna could help to evaluate the wetlands performance and better understand the purification mechanisms of the pollutants that take place there. However, studies in this area, such as those of Ouattara et al. (2009 ;2011) are limited to the assessment of the effect of hydraulic load and that of a single plant species (i.e., Panicum maximum), on the structure of the macrofauna in the constructed wetlands.
Objectives of this study were to (i) determine the performance of the vertical flow CWs using different plants species (i.e. Andropogon gayanus, Chrysopogon zizanioides, Echinochloa pyramidalis, Pennisetum purpureum, and Tripsacum laxum), to (ii) evaluate macroinvertebrates biodiversity, and to (iii) investigate the effect of the different plants species upon macroinvertebrates communities and their vertical distribution in the sediments of CWs.  (Nash, 1909) were experimented. These are perennial forage plants highly appreciated by agro-pastoralists for their palatability, adaptation to local climatic conditions and presence in Côte d'Ivoire (Boudet, 1991).
In order to minimize the variability in the experiment and address CWs clogging problems reported by Coulibaly et al. (2008aCoulibaly et al. ( , 2008b and Ouattara et al. (2008), synthetic domestic wastewater was employed in the experiment as the influent of the wetlands. The composition of synthetic domestic wastewater was made according Rodgers et al. (2006) and Healy et al. (2010). However, some modifications were done in order to respect the characteristics (i.e. nitrogen, phosphorus and carbon concentrations) of domestic wastewater encountered in developing countries according to Metcalf & Eddy (1991). The characteristics of the synthetic domestic wastewater used were: COD = 628 mg O 2 /L, BOD 5 = 380 mg O 2 /L, TN = 45 mg/L, TSS = 300 mg/L, TP = 12 mg/L, pH = 6.7-8.00.

Experimental procedure:-
The effect of the five plants specie on the performance and macroinvertebrates biodiversity of vertical-flow CWs was examined over a seven-month period of experimentation. Five beds were transplanted with the plant seedlings (i.e. 9 plants/m²) spaced of 40 cm x 40 cm between the stems on 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 the bed planting.
The plants thus transplanted 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 hydraulic loading of synthetic domestic wastewater according to Ouattara et al. (2008Ouattara et al. ( , 2009, 2011) over 6 months. The synthetic domestic wastewater tank was cleaned before and after the feeding of beds to remove all the impurities settled.
Wastewater samples were taken once a week at inlet (influent) and outlet (effluent) of each bed, stored in an ethylene bottle at 4 °C until analysis. A total of 24 water samples were taken in each bed during the whole period of the experiment. At the end of each two-month growth cycle, plants were harvested according to Ouattara et al. (2008) and the plant aboveground biomass produced was determined by weighing. Three measures of plant aboveground biomass were carried out during the experiment in each bed. For the macroinvertebrates study, sampling consisted of a collection of beds sediments at the end of the treatment trial.

Physical and chemical analyses and removal efficiency:-
For each water sample collected, pH, and dissolved oxygen (DO) were determined according to ISO 10523 (2008) and ISO 5814 (2012), respectively, while total suspended solids (TSS) by ISO 11923 (1997). Then, chemical oxygen demand (COD), total nitrogen (TN), total phosphorus (TP) analyses were done according to ISO 6060/2 (1989), ISO 5663 (1984) and ISO 6878 (2004), respectively. Finally, removal efficiencies was calculated according to Abissy and Mandi (1999) for TSS, COD, TN and TP as follows: Five quadras (i.e., one at each corner and one in the center of the beds) were selected for sampling macroinvertebrates. Thus, five sediment (10 cm ×10 cm ×10 cm) cores were taken from the five quadras of each layer considered at the end of treatment trial. The five sediment cores of each layer were mixed to make the composite core of this one, stored in jars, preserved in 5% formaldehyde and analyzed in the laboratory.
Analysis consisted to pass under a water tap, the contents of each jar previously put in a sieve of 1 mm mesh vacuum, to leach out the sediments and eliminate formaldehyde. Sieve rejects were then recovered and organisms were sorted by hand magnifying glass and forceps under a light source and stored in pill containers containing 5% formaldehyde. Using a binocular loupe, these organisms were identified, counted and pooled by classes, orders and family, combining the identification keys of Bouché (1972) and Dejoux et al. (1981).
Data analysis:-Pollutant and above-ground biomass data were not normally distributed (Shapiro-Wilk normality test), so the statistical tests used for data set analysis included the Kruskal-Wallis and Mann-Mann tests. Whitney. In this case, the Mann-Whitney test was used to compare the physico-chemical parameter values and the biomass of the plants between the beds two by two. The average values of physico-chemical parameters were used for the presentation of the results.
As for macroinvertebrates, several indices were used to analyze their settlement. The taxonomic richness (R) was simply taken as a count of number of taxa present in each bed. The diversity indices provide more information about community composition, about scarcity and commonness of taxa in a given community; Shannon-Wiener index (H), Equitability index (E) are commonly used to characterize taxa diversity in a community (Magurran, 1988). Shannon-Wiener diversity index was expressed following the equation (2) below, where NS is the proportion of individuals found in the taxon; NT is the number of taxa in the samples.

( ) ( ) ( )
Species equitability or evenness (E) was determined by the equation (3), where H was the Shannon-Wiener diversity index, R was the number of taxa in samples.

( )
The density is an important tool to measure macroinvertebrates (Barbour et al., 1996). Thus, in this study, the relative abundance (Ra) estimated according to equation (4) below, was considered: Analysis of these indices were performed using Kruskal-Wallis and Mann-Whitney tests for taxa richness (R) and density (D) while ANOVA variances, and T-test were used for Shannon-Wiener Diversity (H) and equitability (E) to investigate differences between CWs bed. The statistical tests used were performed using R studio 3.3.2 software (Ihaka & Gentleman, 1996). In all tests, the differences were considered statistically significant for p < 0.05.

Results:-
Plant biomass production and wastewater treatment in CWs:-Plant biomass production:- Figure 2 illustrates the variation in plant biomass produced on the different beds of CWs during plants harvests during the treatment trial. The plant biomass ranged from 14. 55  Considering the planted beds, the overall order of performance is as follows: performance of B PP > performance of B TL > performance of B EP > performance of B AG > performance of B CZ . However, some significant differences were found between the beds performances (Kruskal-Wallis test, p ˂ 0.05). However, the beds planted with P. purpureum and T. laxum were significantly efficient than those planted with E. pyramidalis, A. gayanus and C. zizanioides (Whitney test: p < 0.05).
Of the 12 taxa collected in the beds sediments, all were found in the beds transplanted with T. Laxum and P. Purpureum, 11 taxa were obtained in the beds transplanted with E. Pyramidalis and A. Gayanus, 9 in the bed transplanted with C. Zizanioides and 7 in the unplanted bed. However, all the beds had in common 3 classes (Insecta, Malacostraceae and Oligochaeta) divided into 7 orders (Coleoptera, Diptera, Ephemeroptera, Hymenoptera, Lepidoptera, Isopoda and Lumbricinas) and 7 families (Passalidae, Chironomidae, Ephemerellidae, Formicidae, Crambidae, Aselidae and Lumbricidae).
Considering the planted beds, they had in common 9 taxa. Arachnidae were found only in the beds transplanted with T. Laxum and P. Purpureum while, Glossiphoniidae and Acrididae were absent in the bed transplanted with C.
Zizanioides. The number of taxa encountered in the beds sediments differed significantly between planted and unplanted beds (Mann Whitney test, p < 0.05). However, between the planted reactors, it remains similar (p > 0.05).

1032
Insert Table 2 Diversity of macroinvertebrates settlements:- Figure 3 illustrates the variations in the Shannon diversity index and the Equitability of macroinvertebrates living in the sediment of CWs beds. It appeared that macroinvertebrates settlements in planted beds had higher index values than those recorded in the control.
For the Shannon diversity index (Fig. 3A) As for the equitability index (See Fig. 3B Insert Table 3 Examination of the macroinvertebrates settlements in the beds sediments indicated on the whole that the families (i.e. Formicidae and Lumbricidae) were the most abundant. The number of these organisms was respectively 7 (0. Depending on the classes of macroinvertebrates collected, we noted that insecta and oligochaeta appeared the most abundant (Fig. 4). , number of individuals collected in the sediments of the planted beds was higher than that obtained in the control. Comparing the number of macroinvertebrates recorded in each of the layers of the planted beds, it appeared that, that of the bed planted with P. Purpureum remained higher. It was followed by those of the beds planted with T. Laxum, E. Pyramidalis, A. Gayanus and C. Zizanioides, respectively. In the two last layers of the bottom, it was noted that the beds planted with P. Purpureum and T. Laxum were the least populated.
Insert Table 4 Discussion:- Overall, these biomasses were of the same order of magnitude as estimated by Talineau  gayanus. However, P. purpureum appeared to be the best adapted plant species to the culture medium for having achieved the highest above ground biomass during the experience. In fact, this plant species has long been investigated as a species that provides a significantly age-related yield of dry matter that is adapted to the tropics and can be grown and harvested all year round for agro-industrial purposes (Ferraris & Sinclair, 1980). Pillai and Vijayan, (2013) even obtained higher biomasses of 20-25 kg/m 2 in a vertical-flow CW, but using a crop association of Pennisetum varieties (i.e. P. purpureum and P. typhoides).

CWs performance:-
Monitoring of physico-chemical parameters of raw water (influent) and those of the filtrates (effluents) of the beds during the treatment trial revealed higher values of pH and DO in the filtrates compared to raw water. In addition, unlike DO, the pH of the control effluent appeared greater than those of the planted beds. The increase of pH values in the effluents could be explained by the activity of denitrifying bacteria in the deep layers of the developed vertical flow CWs. The work of Finlayson and Chick (1983) and Koné et al. (2011) also report increasing pH in the planted beds effluents. In addition, the highest pH values observed in the control filtrate compared with those of planted beds could due to the action of plants (Shelef et al., 2013). The increase in DO in the effluents would result from the aeration of the untreated water during its application to the vertical flow CW beds used in this study. In addition, small amounts of oxygen from the aerial parts are rejected at the apex of rootlets of plants that could contribute to higher oxygen levels in the influents of planted beds (Poulet et al., 2004;Pérez et al., 2014). However, the different noted between beds effluent DO would be due to a likely variation in the amount of oxygen released into the plant's rhizosphere in the CW, depending on the plant's physiology (Stottmeister et al., 2003;Gagnon et al., 2007).
Concerning pollutants (e. g. TSS, COD, TN and TP), all were significantly removed in the beds effluents. However, except TSS, the other pollutants were significantly more reduced in the planted beds, than in the control. The significant reduction of pollutants in all the bed effluents could be due to CW removal process (i.e., macrophytes assimilation, precipitation, dissolution, adsorption or desorption), but especially, to sediment storage and microbial uptake (Yang et al., 2001, Vymazal, 2006 However, many studies like those of Molle et al (2006), Coulibaly et al. (2008aCoulibaly et al. ( , 2008b and Ouattara et al. (2008), reported that plants would developed channels via roots in the beds sediment, which should improve planted beds permeabilities, contributing in this way to water infiltration into them, than in the control. These channels would favored the transit of suspended solids in the effluent of the planted beds, unlike control's that remained devoid of plants, justifying the strong reduction of TSS in control filtrates compared to those of planted beds. However, the best TSS reduction noted among the effluents in the bed planted with P. purpureum could be attributed to the root system of this species. Indeed, unlike other species, it was found during the demolition of the beds at the end of the treatment trial that the roots of P. purpureum had formed a kind of mat much denser than the other plants at the bottom of the bed. This would have behaved as a second filter at the bottom of the bed which would have decreased more the TSS in this bed. However, if the results of this study were in disagreement with that of Pillai and Vijayan (2013), who obtained the lowest TSS removal rate (65%) in the control against 89% and 93%, respectively in the beds planted with P. maximum and a mosaic of (P. purpureum + P. typhoides), this should be justified by the use of gravel, sand and coconut fibers as filtration materials.
On the other hand, the greatest removing of DOC, TN and TP recorded in planted beds than, in the control would be related to the presence of the plants species. Indeed, according to several reviews of the ways in which plants can affect CW processes, such as those for Brix, (

Macroinvertebrates communities:-
The analysis of macroinvertebrates living in the CWs showed that the abundance and distribution of these organisms in the beds sediments were more or less influenced by the plant species studied. The presence of organisms in the CWs could be explained by various ecological factors, particularly the favorable physicochemical conditions and sufficient nutrient resources, to guarantee their survival (Gagnon et al., 2007, Ouattara et al., 20092011).
The inventory of the macroinvertebrates in the beds of the CWs has allowed identifying 12 families belonging to 6 classes, of which that of the Insects appeared the most diversified. This situation should due to the large taxocenosis of the Insect class (Tachet et al., 2003). However, in addition to the classes of macroinvertebrates recorded by Ouattara et al. (2009;2011), one new class (i.e. Acheta) was found in the bed sediment in this study. In fact, Acheta are pollution resistant taxa often parasitic of invertebrates, of the large faunistic group of Annelids, proliferating favorably in humid environments of high degree of organic pollution (Abbou & Fahde, 2014). Thus, these organisms would have been transported on the reactors by higher animals, attracted by the plants, and would have favorably swarmed the planted reactors, because of the strong vegetation cover of the latter.
The Shannon diversity index values (1.74-2.19 bits/ind.) and Equitability (0.64-0.90 bits/ind.) obtained in this study indicated that the macroinvertebrates settlements recorded in the CWs beds sediments were little diversified and poorly organized (Dajoz, 2000). This diversity of macroinvertebrates settlements in the CWs could be explained by the duration of this treatment trial. The present study, which lasted 6 months, would be relatively short to allow for significant bed colonization. However, analysis of Shannon diversity index values, Equitability and the abundance of taxa indicated that macroinvertebrates were more diversified and more abundant in planted beds than in the control. This situation confirmed the assumed role of plants in generating micro aerobic environment in the rhizosphere and a hygrometry favorable to the colonization of planted beds thanks to the shading they provide (Brix, 1035, Andrews & Harris 2000, Wießner et al., 2002. This result corroborated those of Gagnon et al. (2007) and Ouattara et al. (2009). However, the results indicated that between the planted beds, that planted with P. Purpureum was the most diversified and had more taxon. It was followed by beds planted with T. Laxum, E. Pyramidalis, A. Gayanus and C. Zizanioides, respectively. The settlement difference observed between the planted beds could likely be related to the difference between plant biomass abundance of the tested plants species, especially since the order of abundance of the taxa of the planted beds was similar to that of the abundance of plant biomass produced.
The quantitative analysis of the macroinvertebrates settlement showed that the macroinvertebrates abundance was significantly higher in the planted beds than in the control. This result could be justified by the presence of the different plants, which would have conferred more favorable ecological conditions for proliferation of the fauna. Ouattara et al. (2009;2011) recorded also greater number of macroinvertebrates settlement in beds planted than in the control. On the other hand, the proportion of insects ranging from 40.5 to 47.8% and that of Oligochaeta fluctuating between 22.7 and 51.4% dominated the macroinvertebrates settlement recorded in CW bed sediments. According to Mason (1991), the abundance of these two classes of macroinvertebrates could be explained by the fact that they contain opportunistic taxa, which are able to monopolize the available resources, allowing them to survive longer than the others. However, the result showed that insects dominated most of the planted reactors while Oligochaeta dominated the settlement in the control. The abundance of Oligochaeta in the sediment of the control could be explained by the fact that these organisms being scavengers, they would take advantage of the significant accumulation of the organic matter from the wastewater in this bed one compared to the planted beds, to pullulate (Berry & Jordan, 2001). As for the dominance of Insects in planted beds, this would be justified by their diet. Indeed, according to Gobat et al. (1998) and Lamarre et al. (2012), Coleoptera, Diptera, Ephemeroptera, Lepidoptera, abundantly found in planted beds, are omnivorous and scavenger. As a result, these organisms would have been attracted to plants, biofilm, organic debris and microscopic algae contained in the planted beds.
Results showed a decreasing of the abundance of macroinvertebrates with depth, following the vertical profile in all the CWs beds sediment. These results were similar to those obtained by Ouattara et al. (2009;2011), investigating the effect of P. maximum on macrofauna communities in a CW. These authors attributed this situation to the living conditions of this environment, where oxygen and available energy resources would decrease with depth. Indeed, several studies report the decrease in oxygen concentration and the amount of nutrients required for optimal growth of organisms (Münch et al., 2005, Truu et al., 2009). These would decrease significantly from the first ten centimeters and become rare beyond 20 cm deep.

Conclusion:-
This study confirmed the contribution of plants to the wastewater treatment process in constructed wetlands. All the five experimented plant species grew well and their presence improved significantly pollutants removal into the beds even if TSS removal in the control was slightly higher than those in the planted beds. Average concentrations of TSS, COD, TN and TP in the effluents of all planted beds were below the current standards of wastewater discharges in Cote d'Ivoire of 50 mg TSS/L, 300 mg O 2 COD/L, 50 mg TN/L and 15 mg TP/L. However, the plant species, P. purpureum is suggested in constructed wetlands for the treatment of domestic wastewater, for having contributed the most to the removal of pollutants from wastewater.
Overall macroinvertebrates settlement within the beds was little diversified. Twelve families belonging to 6 classes were recorded with one new class (i.e., Acheta) in addition to the classes of macroinvertebrates previously recorded. The presence of the different plant species had a significant impact on macroinvertebrates diversity and their abundance. Indeed macroinvertebrates were more diversified and more abundant in the planted beds than in the control. Insects dominated all the planted beds, while Oligochaeta were the most abundant in unplanted bed. However, the bed planted with P. purpureum housed the largest number of organisms and was follow by those planted with T. laxum, E. pyramidalis, A. gayanus and C. zizanioides, respectively. Following the vertical profile, macroinvertebrates settlement abundance decreased from the upper layer (0-10 cm) to the bottom layers (50-60 cm