By using a digital camera for shooting, and a Garmin ETREX 30 GPS receiver for determining the geographical coordinates of artisanal gold mining points, several data were collected to map. The cartographic data consists mainly of maps in digital format of the Côte d’Ivoire geology, and also for the realization of the hydrographic network map of the zones, a file of type ‘‘vector” representing the Ivorian hydrographic network has been used. Then, the 1/2,000,000 geological map and the mining cadaster map of the South-East, East and North-East Côte d’Ivoire areas at 1/1,000,000, established by Tagini et al. (1972) and published by Mining Development Company, served to highlight the geological units (i.e. nature of the geological formation) and the map of gold mining sites encountered within the different zones of the study area, respectively.

Concerning topography, the 30 m resolution Shuttle Radar Topography Mission (SRTM) image was uploaded via https://earthexplorer.usgs.gov/ covering the study areas, as well as the 2011 data of Côte d’Ivoire administrative limit map provided by the National Committee on Teledetection and Geographic Information (NCTGI) were also used.

During the study areas’ investigations, an inventory basis on observations has been performed. It was: 1) to observe the different phases of gold extraction: these phases range from sinking to mercury washing of the soil and during these stages, there is first the digging of the soil which consists of making circular or rectangular holes with an average depth of about 25 m. Then, the stages of crushing and grinding, which make it possible to reduce the size of the particles and make them into powder in order to facilitate the recovery of the gold by washing. The second stage: 2) the general layout of the gold extraction site: it involved observing the layout of the sites and the third stage was: 3) the list of the various environmental risks of the gold extraction sites: it is the aim to list the various risks, namely contamination of the soil and sediments, surface water, deforestation, excavation of the soil.

In order to exploit and analyze the datas obtained from the study areas’ investigations, all data were processed using QGIS 2.12 and ArcGIS 10 softwares, from which all the established maps have been georeferenced in the WGS 84 UTM coordinate system, zone 30, northern hemisphere. With a spatial multi-criteria analysis of the sensitivity of the aquatic environments, all the layers of the simple indicator entities were rasterized in GRID format by defining a common resolution cell size of 30 m using the functions “Entities towards Rasters”. The “Reclassification” allows having the classes and codes envisaged. Surface water sensitivity ratings range from 1 (i.e., low) to 4 (i.e., very high) (see Table 1). Combinations were done in “Raster” mode using the “Raster Calculator” tool (Module weighted sum and multiplication) of the “Spatial Analyst” module of ArcGIS according to Equation (1).

S m a = [ α + d h + P e d o ] × [ G c × P l u ] (1)

with:

S_ma: Sensitivity of the aquatic environment;

α: Slope;

d_h: Proximity distance to the hydrographic network;

P_édo: Pedology;

Gc: Ground cover;

P_lu: Pluviometry.

The indicator parameters in Equation (1) are described as follows:

• Slope indicator (α)

An ASTGTM 2 Digital Image Elevation Model (DEM) with a 30-meter resolution covering the study area was used for the determination of altitudes and slope thresholds. Then, a code from 1 to 4 was assigned to each grade of the slope. According to Farr & Kobrick (2000), and Slater et al. (2006), slope types can vary from very low to high depending on thresholds (see Table 1).

• Distance indicator (dh)

The distance indicator map (dh) to the watercourse was done basis on the river system of the previously determined study area. This distance, which reflects the proximity of artisanal gold mining zones to the river system, was calculated using the “Euclidean Distance” function of the Spatial Analyst extension of the ArcGIS software. To further understand, a code was assigned to each distance obtained between the artisanal gold mining sites and hydrographic network, ranging from 1 (low) to 4 (very high) (Macary et al., 2007), thereby showing the exposure risk versus the distances (Table 2).

• Pedology indicator (Pedo)

This indicator was determined from the pedological sketch of Côte d’Ivoire at 1/200,000 of Perraud (1979), by digitizing and georeferencing it in order to obtain the pedological map of the study area. Then, soil types were codified according to their ability to transfer pollutants into surface water (see Table 3). As shown in Table 3, weakly desaturated ferrallitic soils with induration (i.e., code 4) are the most vulnerable to pollutant transfer, followed by indurated moderately desaturated ferrallitic soils (i.e., code 3) and moderately desaturated ferrallitic soils remained weakly indurated and impoverished (i.e., code 2), finally, ferrallitic soils moderately desaturated reworked and depleted designated by code 1.

• Ground cover (Gc)

The ground cover (Cv) representing the factor map was developed using the vegetation vector layers of 1979 in Côte d’Ivoire. This layer has been updated by the zonation footprint of the sites’ gold mining of the different study areas. The factor is defined as the ratio of bare soil loss under specific conditions to soil loss corresponding to soils under an agricultural system (El Garouani et al., 2008). Indeed, the plant cover protects the soil, ensures the damping of raindrops, the slowing down of runoff and infiltration. Thus, for low vegetation, soil loss decreases with increasing vegetation cover (Soutter et al., 2007). Here, Table 4 exhibits different codifications for the influence of soil occupancy on the runoff.

• Pluviometry indicator (Plu)

The study zones’ investigations allowed us to identify the use periods (i.e., March to November) of chemicals in gold ores. In fact, these periods concurring to the rainy season, denote workout risks of these products in the surface waters. Four (4) average rainfall classes were defined by the Jenks method, 1977 in Antoni et al., 2012, according to the calculated rainfall intensity intervals. Finally, a code from 1 to 4 has been assigned to these classes in order to indicate the pollutant entrainment risk level (see Table 5).

The analysis of metallic trace element concentrations (i.e., Hg, Cu, Zn, As) was performed at the National Laboratory for Agricultural Development and then spatialized. The spatialization consisted in mapping the average concentration of metallic trace elements obtained according to the different localities of Angovia, Kokumbo, Hiré et Agbaou. Indeed, four (4) concentration intervals were defined for each metal, ranging from low to very high, to show their level risk of contamination (see Table 6).

The activity of artisanal gold mine has destroyed of the vegetal layer and the soil through ditches more or less gaping left abandoned (see Figure 2(a)), siltation of rivers and vein wells (see Figure 2(b)), and dust emission, discharge of waste oil and washing of crushed rock (see Figure 2(c)). Hence, the soil is irreversibly degraded because it cannot be restored after the abandonment of the sites, disrupting the ecological balance.

Otherwise, the storage of barren ores causes near-rivers siltation following heavy runoff during the rainy season and the risk of soil erosion, as well as the depths of vein sinks increase (greater than or equal to 40 m) that pose risks of groundwater pollution (Figure 2(b)). The same observations have also been observed on artisanal gold mining sites in different neighboring countries such as Mali and Burkina Faso (Diallo et al., 2007; Hien, 2012; Roamba, 2014). Crushing and grinding of rocks containing gold by motorized machinery have also been observed, which could lead to noise pollution and health problems such as lung diseases due to dust, as well as soil pollution by spilling hydrocarbons from the rudimentary machines used (Figure 2(c)). These impacts result in the modification of stream ecology and the loss of aquatic fauna sensitive to these disturbances.

It worth noted that they use rudimentary equipment and did not have the personal protective equipment that could protect them from inhaling dust and the injuries. Kéita (2001) was able to identify among the miners’ pathologies such as musculoskeletal disorders, sleep and respiratory disorders.

Figure 3 represents the mapping of the sensitivity of surface water to metallic trace elements. Indeed, the analysis of this map showed a low sensitivity for all of the metallic trace elements (i.e., Hg, As, Cu and Zn) measured in the waters of the Hire and Agbaou localities, a moderate sensitivity in Kokumbo surface waters and a high sensitivity for the only locality of Angovia. This high Angovia water sensitivity is due to the strong gold mining activities, the topographic and geological, as well as the climatic conditions above-explained in this area.

Indeed, this locality has a steep slope (slope > 5%) and a relatively high rainfall (between 1000 and 1600 mm) which would favor the transport of the loose fractions of mounds of sterile land stored near the wells towards the watercourse causing their siltation, leading to the increased risk of transfer of pollutants to these rivers (Yapi et al., 2014; Ouedraogo, 2006). It worth noted that these sludges contain heavy metals resulting from gold recovery processes.