Composition and Structure of Aquatic Macroinvertebrate Communities in Mining Areas Streams (Côte d’Ivoire, West Africa)

The existence of mining areas for decades could potentially affect aquatic environments and human health. This study allowed assessing the effects of mining activities on streams in three mining areas (Hiré, Lauzoua and Tortiya). Macroinvertebrates were collected on eight campaigns at eight stations using a hand-net, an Ekman grab and surber net. Environmental variables were recorded also. In this study the conductivity values were higher in the locality of Hiré. It was between 154.2 μS / cm (Tchindégri station) and 1753 μS / cm (Tributary Gbloh station). The highest temperature values were recorded in the stations of Tortiya locality (24 o C: Bou 2; 33.1 o C: Bou 1). A total of 184 taxa distributed among nine classes, 20 orders, 80 families and comprising 14 401 individuals were recorded. Insecta were the most diversified class in macroinvertebrate community (84% of taxa collected, eight orders and 59 families). Ecological indices (Shannon Weiner, rarefied richness, Pielou’s evenness) were significantly low at Lauzoua (N'Téko station) and Hiré (Tributary Gbloh station) (MannWhitney test, P <0.05). Ecological indices showed no significant variation between the stations of the locality of Tortiya. The PCA had grouped the stations into four clusters. Conductivity, ammonium, phosphate, nitrites, and nitrates were significantly higher (MannWhitney test, P < 0.05) in cluster I (Tributary Gbloh station) Compared to other clusters. Eight taxa (Limnius sp., Liberonautes chaperi, Gordius sp., Phyllogomphus sp., Orectogyrus sp., Bezzia sp., Adenophlebiodes, Parasedodes sp.) were specific to Tchindégri station (Hiré) and six taxa (Naucoris sp., Amphiops sp., Hydrobius sp., Pseudobagous longulus, Culicinae and Gomphus sp.) were associated to the Bou 1 and Bou 2 stations (Tortiya).

In many parts of the country, authorized and non-authorized mining are located, especially in Hiré , Lauzoua and Tortiya localities. Wastes from these mining activities can cover the surface of stream sediments, destroying habitat and reducing niches and nursery areas of aquatic macroinvertebrates (Jennings et al., 2008). In Côte d'Ivoire several studies explored the impact of mining activities based on the chemical quality of water (Coulibaly et al., 2009;Keumean et al., 2013;Yapi et al., 2014). However, there is a lack of study in Côte d'Ivoire using a biological community, especially macroinvertebrates, in order to assess the impact of mining activities on the rivers near the mining sites. This study will fill informations gap that will constitute a baseline for further studies.
This study aimed to inventory aquatic macroinvertebrates, to determine their composition, their structure and impacts of mining activities on communities of macroinvertebrates collected in streams in mining areas in order to assess their ecological quality.

Environment Variables
At each sampling station and each campaign, nine physicochemical parameters were measured using a variety of instruments. Conductivity was determined using a multi-parameter HANNA HI98703. A turbidimeter HANNA (HI 98703) was used to measure the turbidity and Dissolved oxygen was measured with an oximeter type HANNA HI9146. The pH and temperature were determined with a pH meter type HANNA HI991001. Water was also collected with bottle of 1 L at each sampling station for nutrients (ammonium, nitrate, nitrite, and phosphate) analysis in the laboratory.

Study Areas and Sampling Stations
The study was conducted in three mining areas in Côte d'Ivoire: Tortiya in the north (diamond mining), Hiré (gold mining) and Lauzoua (manganese mining) in the south of the country. Three streams were sampled in Hiré (Tributary Gbloh, Gbloh, Tchindé gri) and lauzoua (Tributary Dougodou, Dougodou, N'Té ko). In the town of Tortiya, the Bou stream was sampled. These localities were chosen according mining activities. Sampling stations ( Figure 1) were selected according their accessibility, the permanence of water at any time and the potential presence of the impacts of mining activities. In each stream one station has been defined, except the Bou stream where two stations (Bou 1 and Bou 2) have been defined. The characteristics of the stations are listed in

Macroinvertebrate Sampling and Identification
Macroinvertebrates were collected during eight campaigns (from November 2017 to January 2019) at each sampling station. Three gears were used to sample aquatic macroinvertebrates: Hand net (250 μm mesh, 50 cm length), Surber net (25 x 20 cm) and an Ekman grab (0.115 m 2 ). The samples were sieved in the field through a 1mm mesh sieve, and the material retained on the mesh was immediately fixed in 70% alcohol. In the laboratory, macroinvertebrates were identified to the lowest possible taxonomic level using a stereomicroscope Olympus SZ (40× magnification) and a series of identification keys (Monod, 1966;Dé joux et al., 1981;Day et al., 2001;Day et al., 2003;De Moor et al., 2003a;De Moor et al., 2003b;Stals & De Moor, 2007& Tachet et al., 2010.

Data Analysis
Aquatic macroinvertebrates structure was described through taxonomic composition, rarefied richness, Shannon-Weiner index, Pielou's Evenness index, frequency of occurrence and Trichoptera, Plecoptera and Ephemeroptera index (EPT). Taxa richness was rarefied to eliminate any bias related to differences in abundances between samples (Heck et al., 1975;Edia et al., 2016). Calculations were performed using the lowest abundance (11 individuals for this study) found in all stations as the target number of individuals (Oksanen et al., 2013). Trichoptera, Plecoptera and Ephemeroptera index (EPT) was also determined at all stations with the aim of determining the impact of mining activities on the water quality of the studied stations.This index represents the relative abundance of these three groups among macroinvertebrates collected at all the stations.
The Frequency of occurrence (FO) was calculated at all sampling station. FO is the percentage of samples in which each taxon occurred. It was calculated according to Dajoz (2000) to gives some information on the number of taxa frequently met in each station without any indication on their quantitative importance (Lauzanne, 1976;Hyslop, 1980).
In order to assess the structure of macroinvertebrate communities, in each mining area, between-stations variations of abovementionned indices were determined using Kruskal-Wallis and Mann-Whitney tests.
Principal Component Analysis (PCA) using the euclidean distance was performed to ordinates sampling stations according to environmental variables. A Hierarchical Classification Analysis (CAH) was performed on PCA axes in order to cluster sampling stations with similar environmental conditions. This analysis was carried out using the Factomine R and factoextra packages.
Variations of physico-chemical parameters and diversity indices (rarefied richness and Shannon-Weiner index) between clusters were assessed using Kruskal-Wallis and Mann-Whitney tests in order to characterize each group of stations. Before performing the comparison test, the normality of data was checked by Shapiro test.
Characteristic taxa of each group were determined through Indicator Value Method (Indval) (Dufrê ne & Legendre, 1997). This method matches information on species abundance and frequency of occurrence among groups. A Monte Carlo permutation test was employed to test significant associations of taxa and group of sites (p < 0.05).
The indicspecies package was used to perform this analysis. Data analyses were performed using R software version 3.6.3.

Environmental Variables
Variations of temperature, conductivity, turbidity, pH, dissolved oxygen, nitrate, nitrite, phosphate and ammonium between stations are summarized in Table 2. Temperature, pH and dissolved oxygen did not vary significantly between stations in the three localities (Kuskal-Wallis and Mann-Whitney tests, p > 0.05). The highest and lowest values of temperature were observed in Tortiya locality respectively in Bou 2 station (24 °C) and Bou 1 station (33.1). The pH varied from 6.15 (N'Téko station, Lauzoua) to 8.69 (Tributary Gbloh station, Hiré ). Dissolveld oxygen values oscillated between 1.76 mg / L (Tributary Dougodou station, Lauzoua) and 12.71 mg / L (Tributary Gbloh station, Hiré ). Conductivity and turbidity did not vary significantly between the stations of Tortiya (Mann Whitney test p > 0.05). However, these parameters were statistically higher (Mann-Whitney test, p < 0.05) respectively in Tributary Gbloh and Gbloh stations, in the locality of Hiré . Turbidity was significantly higher in Tributary Dougodou station, in Lauzoua locality. Concerning nutrients, excepted in Hiré locality, nitrate, nitrite, phosphate and ammonium did not varied significantly between stations in studied localities (Kruskal-Wallis and Mann-Whitney tests, p > 0.05).          Figure 2 shows variations of diversities indices between stations. The rarefied richness in Lauzoua was ranged between 2.96 (N'Téko station) and 8.34 (Dougodou station). This index was significantly higher in Dougodou station (Mann-Whitney test p < 0. 05). In Hiré , rarified richness was between 2.34 (Gbloh station) and 7.23 (Tchindé gri station). This index was significantly lower at Gbloh station (Mann-Whitney test p < 0. 05).

Macroinvertebrate Structure
In Lauzoua, EPT percentage varied between 4.50% (N'Té ko station) and 18.92% (Tributary Dougodou station). In the locality of Tortiya, the value of ETP index was 13% at the two stations (Bou 1 and Bou 2). Relative abundance of EPT was less than 6% in Hiré at all stations.

Abiotic and Taxonomic Differentiations of Sampling Stations
Principal component analysis (PCA) has established the abiotic typology of the stations studied ( Figure 3). The first two axes expressed 52.4% of the total variance, 39.5% for axis 1 and 12.9% for axis 2 ( Figure 3A). The correlation circle ( Figure 3B) revealed that all of the physico-chemical parameters were negatively correlated to axis 1 excepted the temperature which was positively correlated to axis 2. The factor map ( Figure 3C) distinguishes four groups of stations ( Figure 3D).  Indval method revealed that eight taxa (Limnius sp., Liberonautes chaperi, Gordius sp., Phyllogomphus sp., Orectogyrus sp., Bezzia sp., Adenophlebiodes, Parasedodes sp.) were specific to the station Tchindé gri (Hiré ) and six taxa (Naucoris sp., Amphiops sp., Hydrobius sp., Pseudobagous longulus, Culicinae and Gomphus sp.) were associated at stations in the locality of Tortiya.

Discussion
During this study, the highest values of temperature were recorded in the stations of Tortiya locality. This could be due to the absence of canopy at these sampling stations. Indeed, any surface of water not covered is subject to a very important sunning thus favoring the increase of the temperature of the water.
The higher values of conductivity (1162 to 1753 µS / cm) registered in Tributary Gbloh station (Hiré ) could be linked to the fact that this station was in an area subject to a permanent supply of effluent favoring the dissolution of the metals present in the rocks and sediments. These results corroborate those of Yapi et al. (2014) in this same locality.
Turbidity was higher in Tributay Gbloh station (102.3 to 189.3 NTU: Hiré ). This situation could be explained by the location of this stream in downstream which receive particles from upstream areas where mining activities are practiced. These results corroborate those of Bamba et al. (2013) on the impact of mining on rivers which can destabilize the banks and lead to a massive sediment supply which can locally disturb the balance of the rivers and increase the turbidity of the water.
The abiotic typology of the stations by the PCA revealed that the Triblutary Gbloh station (cluster I) is distinguished from the other stations by high values of the mineralization parameters (conductivity, ammonium, nitrate, nitrite, phosphate). This high mineralization observed in this station could be explained by the drainage of agricultural products in this watercourse. Indeed, according to Brugneaux et al. (2004) and Troeh et al. (2004), by the action of rain that drains cultivated land, surface waters receive increased nutrient inputs.
The Indval method revealed that among the indicator taxa of Tchnidé gri station, there were two polluo-sensitive organisms: Adenophlebiodes (Ephemeroptera, tolerance level = 2), Parasetodes sp. (Trichoptera, tolerance level = 4) recognized as good bio-indicators of watercourses because of their sensitivity to oxygen depletion (Hynes, 1957). We can therefore deduce that the waters of this station have a acceptable ecological quality.
A total of 184 aquatic macroinvertebrate taxa were collected and insects accounted for 84% of the taxa collected. Insects were the most abundant group among the macroinvertebrates collected in this study. Insects abundance http://journal.julypress.com/index.php/jess Journal of Environmental Science Studies Vol. 3, No. 1, 2020 could be explained by their omnipresence, which is due to their capacity for resilience. In several studies insects were most abundant (Diomandé et al., 2009;Akindele & Liadi, 2014).
In Hiré , the lowest rarefied richness (2.34) was obtained at Gbloh station. This station had the lowest values of Shannon Weiner index (0.68) and Pielou's eveness (0.17) index. These results obtained could indicate that this station was the most impacted by gold mining, which would have an impact on aquatic macroinvertebrate assemblages. According to Rosenberg & Resh (1993), human perturbations change community structure in watercourses because species are adapted to certain environmental conditions. The low diversity in this station may reflect the response of benthic macroinvertebrates to the toxicity in this station. This could be attributed to the loss of habitat diversity due to the reduction of ecological niches. The highest values of the Shannon Weiner index were obtained in Lauzoua and Tortiya respectively at Dougodou (3.02) and Bou 1 (3.2) stations. These results show that the aquatic macroinvertebrates of these stations were most diversified which could reflect good water quality at these stations.
EPT taxa are sensitive macroinvertebrates. They are met usually in water of good quality. This group were present at all stations with different proportions. These organisms were mainly composed of Baetidae (Baetis sp, Ephemeroptera) and Hydropsychidae (Hydropsyche sp., Trichoptera). These organisms are sensitive to metal pollution (Malmquist & Hoffsten, 1999) but can recolonized rapidly disturbed stations (Kiffney & Clements, 1994). The Tributary Gbloh station was the most affected by gold mining in Hiré with an EPT proportion of 4.91%. In Lauzoua, N'Té ko station recorded the lowest proportion (4.50%) of EPT. However, there was no mining activity at this station. The low proportion of EPT at this station may be due to other human activities. In Tortiya locality, EPT proportions were relatively higher: 13.62% in Bou 1 and 12.30% in Bou 2. Diamond mining has little impact on station water quality. These proportions obtained could be explained by the fact that no metal is used in diamond mining. One of the limitations of this study is the number of sampling points chosen which should have been several on each station instead of choosing only one sampling point. However, the results show plausible consequences of mining on the quality of rivers, therefore alert the authorities to provide guidance on the consumption of drinking water by populations near these sites.

Conclusion
This study allowed to collect macroinvertebrates from three mining areas. Gold and manganiferous mining had an impact on macroinvertebrate communities. However, this impact was more significant in the streams near mining operation. Comparing to the latter two mining sites, diamond mining causes least disturbances to macroinvertebrate communities. Therefore, gold and manganiferous mining have most impact on the ecological quality of the studied rivers. Based on findings, actions must be conducted in these mining areas in order to stop the impact of the tailings on the pollution of waters surrounding residential areas.