The Tongon site lies in the M’Bengue department (Poro region), northern Cote d’Ivoire, just south of the Tongon gold mine, within the Bandama River basin. The Katie River borders the site; the mine has built a water reservoir on it, with a 48.1 km perimeter covering about 11 km², for its industrial needs. The relief is semi-accidented, made of rolling plateaus between 300 and 400 m elevation, interspersed with iron-capped hills (bowe) and granitic ridges. The climate is tropical of the Sudanese-Guinean type, with a rainy season from May to October and a dry season from November to April, marked by the harmattan from December to February. Over 2016-2021, annual rainfall at the Korhogo synoptic station, the most representative of the site, ranged from 658 mm to 1452 mm, with a mean of 1055 mm. Mean monthly temperature varies from 24.7 degrees C (August) to 29.5 degrees C (March-April). The calculated mean annual evaporation is 1653.4 mm, peaking in January (255 mm) under the harmattan. Mean relative humidity ranges from 31% (December-January) to 77% (June-September), and the dominant winds, driven by the monsoon, are southwesterly at 3.1 to 5 m/s.

The site geology consists of outcrops of medium-grained biotite granites, locally associated with green rocks (schists, quartzites, manganese-bearing rocks) and with concordant granitoids and basic plutonites. The weathering profile, revealed by hand-auger borings (down to 4 m, or refusal between 0.20 and 3.80 m) and by the pressuremeter borings SP1 to SP10 (12 m deep) carried out by LABOGEM, shows from top to bottom: a thin humiferous soil, a layer of gravelly and lateritic clays (sandy clay, lateritic gravel, lateritic clay in brown, grayish, yellowish, or reddish shades), 1.40 to 3.10 m thick, and then weathering clays over the whole tested depth.

The field surveys used: 1) a differential GNSS receiver to position the measurement points, site boundaries, and sampling points; 2) a current meter for discharge measurements in the Katie River; 3) a field multiparameter probe (electrical conductivity, pH, temperature) for in situ hydrochemical characterization; 4) a handheld GPS and a flood-stage chart to record observed water levels; and 5) a geographic information system (QGIS/ArcGIS) coupled with remote-sensing data (multi-date satellite images and a digital elevation model) for the spatial analyses. Statistical and cartographic processing was done in QGIS, Excel, and the HYFRAN software [10] for frequency analyses.

Four categories of data were used:

climate data (2016-2021) from the Korhogo synoptic station [11]: rainfall, maximum and minimum temperatures, evaporation, relative humidity, and the 2021 wind rose;hydrogeological data from the ONEP [12] on the abstraction works of the Tongon locality (total depths and depth to bedrock, static water levels, exploitable discharges, coordinates), completed by the LABOGEM geotechnical borings (hand auger and pressure meter tests);water-quality data: arsenic concentrations (2013-2016) in the surface and groundwaters of the mine and the riparian villages [9];field data from the hydrological and hydrogeological campaign of 10-14 May 2022: discharges, flood levels, resurgence points, conductivities, photographs, and testimonies of the riparian populations.

The flood-risk assessment followed the classical hazard x vulnerability framework [3][13]. First, thematic maps of the Katie sub-catchment were produced from the DEM, land cover, distance to the drainage network, pedology, and rainfall: slope, drainage density, contour lines, distance to watercourses, soil types, and land-cover classes. Each thematic map was classified and weighted, and their weighted arithmetic combination gave the hazard map, using an AHP-type multicriteria analysis already applied with success in West Africa [4][14].

Second, the vulnerability map was drawn from the existing and planned stakes (footprint of the site, access infrastructures and bridge, settlements, and agricultural activities). Multiplying the hazard map by the vulnerability map gave the flood-risk map, classified into four levels (low, moderate, high, very high). The map was validated against the field observations of the May 2022 campaign (areas actually flooded in August-September, water levels reached, condition of the downstream bridge of the mine) and the testimonies of riparian populations [15].

The discharge of the Katie River was measured at low flow (May 2022) at the Tongon bridge by the velocity-area method: the cross-section was divided into verticals, velocity was measured with the current meter, and the elementary discharges were integrated. The maximum flood level was recorded from flood marks and testimonies (16 cm at the site boundary). The annual volume was first approximated by

with t = 24 × 365 h, assuming a regime close to the measured one. This is a conservative assumption, given the strong seasonality of Sudanese-Guinean regimes [16][17]. The frequency analyses of regional floods followed the approaches of daily flood regional analysis [18][19] and the WMO manuals for low flows [20].

3.3.1. Meteoric Water (Rainwater Harvesting)

The volume recoverable from a corrugated roof was estimated by the classical collection formula:

where P is the rainfall (mm), S the collecting area (50 m²), and Cp the roof loss coefficient (0.8). The computation was repeated for the maximum (1452 mm), mean (1055 mm), and minimum (658 mm) rainfall, in line with stormwater management guidelines [21].

3.3.2. Surface Water

The potential of the Katie River was assessed from the measured discharge, extrapolated to an annual volume, and compared with the competing uses (an identified agricultural pumping of 60 m³/h). The feasibility of a hillside reservoir of the type built by the mine was examined against the needs of the plant (panel cleaning and related uses).

3.3.3. Groundwater

The hydrogeology of the area was characterized according to the conceptual model of hard-rock aquifers [5][6]: an unconfined aquifer in the regolith (10 - 45 m) overlying a deep fractured aquifer. The parameters came from the ONEP borehole records (static level, exploitable discharge, total depth, depth to bedrock) and from the LABOGEM borings (regolith thickness, static water level, test discharges). Discharges were converted into annual volumes (Q × 24 × 365) and compared with the project’s needs.

The arsenic contents measured by Affoun [9] over 2013-2016 in the surface waters (RW and SW points) and groundwaters (TGW and GW points) of the mine and the villages were compared with the WHO guideline for drinking water [22] (10 ug/L, i.e., 0.01 ppm) and with the IFC environmental guideline for effluents and surface waters [23] (0.1 ppm). The interpretation separated the anthropogenic origin (mining activities, management of waste rock and effluents) from the natural origin (arsenopyrite of the deposit), based on the mine/village comparison and the absence of a marked temporal trend [9].

The water requirements of the plant, estimated by the developer at 740 m³/year for the construction and operation phases, were compared with the sum of the annualized resources (recoverable meteoric water, annual volume of the Katie, exploitable volumes of the boreholes) to establish the availability balance of the resource.

The climate statistics of 2016-2021 reflect the Sudanese-Guinean regime of the site: the amplitude of the mean maximum temperatures (3.3 degrees C) exceeds that of the minima. Evaporation is highest in the dry season (>200 mm/month from January to March) and lowest in the rainy season (<80 mm/month from August to September) (Table 1 and Table 2).

Table 1.Mean monthly temperatures at the Korhogo synoptic station (2016-2021) [11].

Table 2.Mean monthly evaporation (mm) at Korhogo (2016-2021) [11].

The combination of the thematic maps (slope, drainage, pedology, land cover, rainfall) produced the hazard and vulnerability maps, whose crossing gives the flood-risk map of the Katie sub-catchment. The project site falls in a low to moderate flood-risk zone. The areas that can be flooded are the banks of the Katie River, which borders the site, in August and September: the river overflows into the plain downstream of the mine, and the bridge downstream of the Tongon gold mine is regularly submerged (field observation and testimonies of the riparian population). When the river leaves its bed, the water depth can reach 16 cm at the site boundary. This may hinder access during construction and operation, but it does not threaten the main layout of the plant. The discharge measurement of May 2022 gives a low flow of 1.912 L/s (6.88 m³/h) at the Tongon bridge, hence an annual volume of about 60,268,800 L (extrapolated over 24 × 365 h). A resurgence was recorded on the site, with conductivity measurements confirming the groundwater-river interaction. The mine reservoir (48.1 km perimeter; 11 km²) shows the storage potential of the watercourse (Figures 1-3).

Figure 1.Discharge measurement of the Katié River at Tongon (May 2022).

Figure 2.Resurgence zone and conductivity measurement at Tongon (May 2022).

Figure 3.View of the reservoir of the Tongon mine (impoundment on the Katié River) (May 2022).

4.3.1. Meteoric Water

For a 50 m² roof with a loss coefficient of 0.8, the recoverable volume varies from 26.3 m³ (dry year) to 58.1 m³ (wet year), i.e., 42.2 m³ for the mean year (Table 3).

Table 3.Recoverable rainwater volume (50 m² roof, Cp = 0.8).

4.3.2. Surface Water

Figure 4.Generalized drainage network of the Bandama basin.

The local hydrography is dominated by the Bandama and its tributaries, including the Katié. The extrapolated annual volume of the Katié (60.3 million liters) easily covers the industrial needs of the plant (panel cleaning and related uses); an agricultural pumping of 60 m³/h already exists on the watercourse. A hillside reservoir modeled on the mine’s appears technically feasible (Figure 4 and Figure 5).

Figure 5.Local hydrography of the Tongon site.

4.3.3. Groundwater

Two aquifers coexist: 1) the unconfined regolith aquifer, the most important, between 10 and 45 m depth, with a static water level encountered between 1.50 and 10.50 m during the LABOGEM borings (June 2022); and 2) a deep fractured aquifer, poorly developed but present in the mineralization zone, with levels from 2 to 26 m for discharges of 0.03 to 2.5 L/s. In the regolith, the mean depth is 38 m with a static level at 4.1 m and mean borehole discharges of about 1.94 m³/h; in migmatites, the regolith reaches 52 m depth with a static level at 12.3 m, and the boreholes do not dry out in the dry season. The two exploited boreholes of the locality [12] have static water levels of 28.6 m and 24.5 m and exploitable discharges of 0.60 m³/h (5,256,000 L/year) and 1.20 m³/h, for total depths of 61 m and 68.3 m (Table 4). The regolith boreholes tend to dry out in the dry season and recharge during the rainy season (Figure 6).

Figure 6.Conceptual model of a hard-rock aquifer: stratiform, multilayer aquifer linked to former lateritic weathering profiles [5].

Table 4.Characteristics of the exploited boreholes at Tongon [12].

4.3.4. Water Quality (Arsenic)

In the surface waters of the mine, mean annual arsenic concentrations frequently exceed the IFC guideline (0.1 ppm), notably at points RW01 (2.25 ppm in 2013; 1.00 ppm in 2016) and RW02 (0.158 ppm in 2016); in the surface waters of the villages, contents reach 0.335 ppm (SW13, 2016) (Table 5). In the groundwaters of the mine, several exceedances of the WHO recommendation (0.01 ppm) are recorded (TGW13: 0.438 ppm in 2014; TGW29: 0.795 ppm in 2014; TGW28: 0.226 ppm in 2015), while the village piezometers (GW02, GW10) show occasional exceedances in 2013 without a clear temporal trend (Table 6). The absence of a temporal gradient linked to the mining operation suggests a contamination of natural origin related to arsenopyrite, possibly aggravated at the surface by the management of waste rock and effluent treatment [9].

Table 5.Mean arsenic concentrations (ppm) in surface waters [9].

Table 6.Arsenic concentrations (ppm) in groundwaters [9].

The water requirements of the plant are estimated at 740,000 L/year. The annualized resources total 65,551,120 L/year: 26,320 L of meteoric water (dry year, 50 m² roof), 60,268,800 L of surface water (Katie), and 5,256,000 L of groundwater (minimum discharge of the first borehole). The resources-to-needs ratio is therefore about 89, a very largely surplus availability (Table 7).

Table 7.Comparison of the available annual reserves with the annual water requirements of the plant.

Note: the discharges of the Katié and of the boreholes were converted into annual volumes (multiplication by 24 × 365).