The Ouaka River watershed is located in the central-east of the Central African Republic (CAR) and crosses the sub-prefectures of Bakala, Bambari and Kouango. The watershed covers an area of 29,730 km². The Ouaka River, 611 km long, is one of the tributaries of the Oubangui River which separates the Central African Republic (CAR) from the Democratic Republic of Congo. The hydrographic network of this basin is of the dendritic to subparallel type, and the geomorphology is marked by a plain whose altitude varies between 368 m and 805 m. The area is dominated by a Sudano-Aubangian climate, with alternating dry seasons (December to February) and rainy seasons (March to November). Average annual rainfall is 1372 mm and the average annual temperature is 26.8˚C.

Figure 1. Geological map of the CAR showing the study zone ([12]).

Geologically, the Ouaka River basin is bounded to the north by the Pan-African orogenic chain, and to the south by the northern edge of the Congo Craton. The rocks represented in the study area correspond to Archean and Palaeo-Proterozoic assemblages ([12]; Figure 1). According to [5], the special feature of the Archean is the presence of a wide band of heterogeneous granite (3.6 to 2.8 Ga), 25 to 50 km long, which crosses the Bakala region in a NW-SE direction. This heterogeneous granite is made up of monzonites, tonalites and granodiorites. A large buttonhole of migmatites appears in the Bakala region. The Lower Proterozoic is represented by the essentially quartzite series, clearly metamorphosed but not migmatized, which covers a large area of the Ouaka river basin ([39]). These rock formations are eroded and are found in the riverbed.

Superficial sediment sampling campaigns were carried out during the dry season in the Ouaka River catchment area. All samples were taken from accessible banks, at a depth of 25 cm to 40 cm. The Global Positioning System (GPS) was used to pinpoint the precise geographical locations of the sampling points (Table 1). A total of 34 sediment samples were taken in the Ouaka River watershed, including 17 upstream of the town of Bambari (B1, B2, B3, B4, B5, B6, B7, B8, B9, B10, B11, B12, B13, B14, B15, BRS1 and BRS2) and 17 downstream (K1, K2, K3, K4, K5, K6, K7, K8, K9, K10, K11, K12, K13, K14, K15, KRS1 and KRS2), were kept in coded plastic bags for laboratory analysis.

At the MIPROMALO laboratory in Yaoundé, Cameroon, the particle size analysis of the sand samples was used to determine the particle size distribution using conventional techniques ([54]; [43]; [8]; [34]). After washing each sample, the sandy fractions were separated and treated with hydrochloric acid and hydrogen peroxide to remove carbonates and organic matter, then dried in an oven at 60˚C. After drying, each sand sample is sieved on a column of fifteen vibrating sieves from the AFNOR series, with mesh sizes ranging from 0.063 mm to 4 mm.

The reject from each sieve was weighed and, using the formulae of [27] and [17] based on the unit Φ (Φ = −3.3219 log (Q) with “Q” denoting grain diameter in mm), five particle size distribution parameters were calculated ([6]; [54]; [43]; [44]; [46]). These are the median (Md), mean (Mz), standard deviation or sorting (So), skewness (Sk) and kurtosis (Kg).

Table 1. Sampling sites in the Ouaka River watershed.

2.3.1. Mean

The mean or average grain defines the average size of sand grains. It reflects the average power of palaeoflows ([15]; [32]; [13]). It is defined by the relationship:

Depending on the different values of the mean, the following facies can be distinguished:

Very coarse sands: 2000 μm > Mz > 1000 μm ou −1 Φ < Mz < 0 Φ; Coarse sands: 1000 μm > Mz > 500 μm ou 0 Φ < Mz < 1 Φ; Medium sands: 500 μm > Mz > 250 μm ou 1 Φ < Mz < 2 Φ; Fine sands: 250 μm > Mz >125 μm ou 2 Φ < Mz < 3 Φ; Very fine sands: 125 μm > Mz > 63 μm ou 3 Φ < Mz < 4 Φ; Silts and clays: Mz < 63 μm ou Mz > 4 Φ.

2.3.2. Median

The median (Md) or the average is linked to the size of the grains as a function of the energy of deposition of the medium. It expresses 50% coarse elements and 50% fine elements. It defines the central tendency of the granulometric distribution. Its formula is:

Md = Φ50. (2)

2.3.3. Sorting (So) or Classification Index

Sorting is a mathematical measure used to classify and assess sorting actions during transport and deposition. Sorting values reflect fluctuations in kinetic energy and therefore in the speed of the deposition agent ([45]; [49]; [18]). It is defined by the following formula:

Standard deviation values provide information on sand classification and highlight 6 sediment classes ([17]; [18]):

Very well classified sand: So < 0.35; Well classified sand: 0.35 < So < 0.50; Fairly well classified sand: 0.50 < So < 0.71; Medium-grade sand: 0.71 < So < 1; Poorly classified sand: 1 < So < 2; Very poorly classified sand: 2 < So < 4.

2.3.4. Asymmetry or Skewness (Sk)

Skewness is the analysis of the particle size distribution of sediments at the point where they were deposited, according to whether they are fine or coarse (generally represented mathematically by positive or negative skewness). Its formula is:

Depending on the difference in slope, different types of curves can be distinguished:

Very asymmetrical curves towards coarse gravel: −1.00 < Sk < −0.30; Asymmetrical curves towards coarse gravel: −0.30 < Sk < −0.10; Almost symmetrical curves: −0.10 < Sk < 0.10; Asymmetrical curves towards fines: 0.10 < Sk < 0.30; Very asymmetrical curves towards the ends: 0.30 < Sk < 1.00.

2.3.5. Acuity Index or Kurtosis (Kg)

Kurtosis expresses the ratio between the extent of the central part and the extent of the extremities of the granulometric distribution on the frequency curve. It therefore measures the angularity of the frequency curve. It is defined by the following formula:

Multi-modal frequency curves are not suitable for calculating the acuity index ([9]). This index will therefore only be calculated for unimodal frequency curves. Kurtosis values can be used to classify 6 types of frequency curve ([17]; [18]):

Very platykurtic curve: Kg < 0.67;Platykurtic curve: 0.67 < Kg < 0.90;Mesokurtic curve: 0.90 < Kg < 1.11;Leptokurtic curve: 1.11 < Kg < 1.50;Very leptokurtic curve: 1.50 < Kg < 3;Extremely leptokurtic curve: Kg > 3.

The test of [55] used by [52] is used to determine the transport mode of fluvial sediments. Three modes of transport (suspension, saltation and rolling) are highlighted by placing the grain size on the x-axis and the cumulative percentages on the y-axis.

The most complete mode of transport is that of a tractive current represented by the C-M diagram (C = percentile and M = median) of [36], [22] and [52]. This diagram can be used to interpret sediment grain size variations in terms of transport dynamics. It can also be used to identify the types of currents that transport material. The five segments of this diagram each correspond to a particular mode of transport: NO (bottom rolling deposits), OP (bottom rolling and graded suspension deposits), PQ (graded suspension and bottom rolling deposits), QR (graded suspension deposits) and RS (homogeneous or uniform suspension deposits).

The [33], [34] and [49] Md-So diagram is used to relate the granulometry of sediments to the environment in which they were deposited. This method makes it possible to deduce the origin of the sands (rivers, beaches, continental dunes, coastal dunes) from the location of the point clouds obtained. [19] and [52] Sk-So diagram can also be used to characterize sediments from a palaeo-environmental point of view.

The interpretation of variations in energy and fluidity factors, based on the statistical method for studying sediments, appears to have an excellent correlation with the various processes and the deposition environment ([45]; [44]; [46]). It is in this context that the analysis of the linear discriminant function of sediments was established with the aim of characterizing the depositional environment on the basis of the following equations (where Mz is the granulometric mean, So is the graphical standard deviation included sorting, Sk is the asymmetry and Kg is the graphical kurtosis):

Discrimination between wind and beach environments (Y1)

Y1 = −3.5688 Mz + 3.7016 So² − 2.0766 Sk + 3.1135 Kg (6)

If Y1 is less than −2.7411, it is a wind deposit and if it is greater than −2.7411, it is a beach.

Discrimination between beach and shallow agitated marine environment (Y2)

Y2 = 15.6534 Mz + 65.7091 So² +18.1071 Sk + 18.5043 Kg (7)

If Y2 is less than 65.3650, it is a beach deposit and if it is greater than 65.3650, it is a shallow agitated marine environment.

Discrimination between shallow, agitated marine environments and river environments (Y3)

Y3 = 0.2852 Mz − 8.7604 So² − 4.8932 Sk + 0.0482 Kg (8)

If Y3 is less than −7.419, the environment is fluvial, and if it is greater than −7.419, the environment is shallow marine.

Discrimination between rivers and turbidity (Y4)

Y4 = 0.7215 Mz + 0.403 So² + 6.7322 Sk + 5.2927 Kg (9)

If Y4 is less than 10.000, the environment is fluvial and if it is greater than 10.000, the environment is turbid.

Twenty out of thirty-four sediment samples collected in the field were washed under tap water (i.e. 200g on each sample) on a 50 μm sieve to eliminate particles smaller than this mesh diameter.

After washing, the samples were oven-dried at 105˚C for 24 hours. After drying, the dry sediment obtained was divided into four parts in order to obtain a sample that was well representative of the sediment as a whole.

Bromoform (CHBr3) with a density of 2.89 is the dense liquor used for mineral separation. The separating funnel is filled about 3/4 full with bromoform. The sand is poured into the funnel, and with the tap closed, a glass stirrer is used to impart a rotary movement to the liquor, in order to make the heavy minerals fall to the bottom of the funnel. The liquor is rotated several times until separation is complete. After a few minutes’ rest, the tap is opened and the heavy minerals are collected in the funnel fitted with the filter paper. As soon as the heavy minerals have been recovered, the tap is closed. The filter paper lets through the bromoform, which is collected in the Erlenmeyer flask. On a hot plate, the slides are baked and soaked in Canada balsam. Once the balsam has turned yellow, more heavy minerals are added to the slide until the desired concentrate is obtained. The slide thus obtained is immersed in an alcoholic solution to remove the balsam. Ideally, 100 to 200 transparent grains per slide are ready to be observed and counted under a polarising microscope.

The “ZTR” index, which is the combined percentage of zircon, tourmaline and rutile among non-opaque heavy minerals, was calculated using the formula of [23]:

ZTR index < 75% implies immature to sub-mature sediments and ZTR > 75% indicates mineralogically mature sediments ([23]; [52]).

The proportions of the different granulometric facies of the Ouaka River sediments are shown in Table 2. Analysis of this Table 2 revealed that sediments in the upstream zone are dominated by medium sand particles with an average concentration of 41.2% ± 14.44, followed by fine sand (average concentration of 21.65% ± 10.79), silt-clay (average concentration of 21.45% ± 10.5), coarse sand (average concentration of 11.05% ± 9.01) and gravel (average concentration of 4.66 ± 4.24). In the downstream zone, on the other hand, the average concentration of coarse sand (33.43% ± 12.8) is dominant, followed by silt-clay (30.92% ± 6.79), medium sand (24.87% ± 7.9), fine sand (6.41% ± 1.07) and gravel (4.37% ± 1.98). Generally speaking, we note a spatial evolution of deposition from medium sand particles upstream to fine sand particles downstream, with the presence of a high concentration of coarse sand downstream. This spatial distribution of granulometric facies is a function of the hydrodynamics that drain the sand particles, whose diameters decrease regularly from upstream to downstream. However, the presence of a few coarse materials in the downstream zone is linked to inputs from secondary watercourses, which may be responsible for the sedimentation of these materials in this zone.

The cumulative semi-logarithmic curves for sediments in the Ouaka River have a generally sigmoidal shape (Figure 2(a), Figure 2(b)). Overall, these curves show a predominance of the fraction between 0.02 and 2 mm (sand and silt-clay). To a first approximation, and based on the observation of these granulometric curves, the sand is predominantly made up of medium grains in both the upstream and downstream zones. This granulometric facies indicates that deposition occurred as a result of moderate to strong variation in the transport current ([53]; [52], [18]). In the same vein, [43] states that hyperbolic or sigmoid curves indicate a progressive reduction in the strength of the transport agent from upstream to downstream. Such facies are most often found in sediments deposited by surface currents or bodies of water. Sedimentation occurred by free accumulation. These analyses reinforce the impact of hydrodynamics, highlighted by the variation in the proportions of the different granulometric facies in the Ouaka River sediments.

Table 2. Distribution of granulometric facies (%) in Ouaka River sediments.

Av: Average; Sd: Standard deviation.

Figure 2. Cumulative granulometric curves of sediments from Ouaka River: (a) upstream sector; (b) downstream sector.

3.2.1. Graphique Mean Size (Mz)

The mean graphical size of the river sediments varies from −0.19Φ to 1.99Φ, with an average of 0.83Φ ± 0.91 in the upstream zone and from −0.03Φ to 1.97Φ, with an average of 0.28Φ ± 0.44 in the upstream zone (Table 3). Analysis of Table 3 shows that medium to fine sands are dominant in the Ouaka River. Coarse to very coarse sands can be observed in certain sampling sites in the upstream zone, particularly in the localities of Gono and Takobanda. Medium sands dominate in the localities of Bakala and Bambari, and also in the downstream zone. The presence of coarse particles in this river corroborates the theory of significant particle inputs from secondary watercourses, particularly tributaries.

3.2.2. Sorting (So) or Classification Index

The sorting values (So) for sediments in the Ouaka River, presented in Table 3, vary between 1.07Φ and 1.34Φ (mean 1.22Φ ± 0.08) upstream and between 1.08Φ and 3.61Φ (mean 2.09Φ ± 0.66) downstream. These results are typical of river environments and provide ample evidence that medium to very fine sands are very poorly classified and coarse to very coarse sands are poorly classified. The poorly classified sands were transported by a slightly irregular current, whereas the highly classified sands were transported by a fairly irregular current. This fairly irregular current comes from the tributary streams of the Ouaka River, which influences the classification of the sediments. This poor grading may be linked to the presence of numerous obstacles such as islands, which reduce hydrodynamics and concentrate particles of all sizes in different sectors of the Ouaka River.

3.2.3. Skewness (Sk)

The sediments of the Ouaka River have skewness index (Sk) values that vary between −2.1 and −0.08, with an average of −0.36 ± 0.48 in the upstream zone and between −0.61 and −0.08, with an average of −0.26 ± 0.15 in the downstream zone (Table 3). The exceptions are samples B2 = 0.29 and B7 = 0.02 in the upstream zone, where the curves are almost symmetrical and asymmetrical towards the ends, respectively. Generally speaking, the skewness values of the sands are mostly negative. This trend reflects a preponderance of asymmetrical curves towards coarse sand and very asymmetrical curves towards coarse sand. This indicates the influence of several sources of sedimentary input (tributary materials, runoff and gully water) with currents of varying intensity but generally strong in the Ouaka River.

3.2.4. Kurtosis (Kg)

The upstream zone of the Ouaka River records very varied values of the acuity index or kurtosis (Kg). Most of the values oscillate between 2.54 and 3.93, characteristic of very to extremely leptokuric frequency curves. On the other hand, the Kg values of sediment samples B2, B5, B6 and B7 show platykurtic curves and those of B3, B8 and BRS2 show leptokurtic curves. In the downstream zone, very leptokurtic curves dominate (1.7 to 2.93). With the exception of samples K7 = 1.23 is leptokurtic and K15 = 1.09 is mesokurtic.

This variation in kurtosis values may be due to occasional variation in the flow characteristic of the depositional medium ([46]). The values thus obtained provide information on the quality of the grading and also on the multiplicity of different sources of input (several granulometric stocks). Taken together, this suggests a highly contrasted hydrodynamic variation from various sources of supply. Such variation may also be due to seasonal changes in the hydrological regime of the watercourses (period of sediment collection in the dry and rainy seasons in the Ouaka River), linked to rainfall variability ([52]).

Table 3. Granulometric parameters of sediments in the Ouaka River.

Av: Average; Sd: Standard deviation.

Application of the test of [55] used by [52] to sediments from the Ouaka River produced the results shown in Figure 3(a) and Figure 3(b). This test shows the relationship between sediment grain size and transport mode. Three populations of sand are found in the study zone:

The population A includes sands transported by rolling;The population B includes sands transported by saltation;The population C is characterized by the transport of grains by suspension.

Figure 3. Semi-logarithmic grain size curves of sandy facies: (a) upstream sector; (b) downstream sector.

Table 4. Percentage of sediment by different modes of transport.

Analysis of Table 4 shows that the population B is the largest in the Ouaka Rver watershed, at 48% in the upstream sector and 49% in the downstream sector. This demonstrates the hegemony of saltation transport of sands and silts. Transport by rolling only concerns the coarse fractions, as shown by population A. Population A is not very dominant in the upstream and downstream sectors, with concentrations of 42% and 31% respectively. Transport by suspension (population C) only concerns fine fractions (silt). Their concentrations vary between 10% upstream and 20% downstream.

Table 5 has been drawn up to corroborate the relationship between the different types of sediment transport and the different particle size classes.

In the upstream sector, gravel was mainly transported by rolling (76%), followed by coarse sand (47%), medium sand (47%), fine sand (41%) and silt-clay (14%). Saltation transport carried coarse sand (53%), medium sand (53%), fine sand (59%) and silt-clay (51%). Finally, silt-clay (35%) was only transported by suspension. On the other hand, in the downstream sector, the mode of transport of gravel (88%) and coarse sand (56%) is dominated by rolling. Transport by saltation mainly ordered the transport of medium sands (94%) and fine sands (94%). As in the upstream sector, silt-clay (65%) is mainly controlled by suspension transport.

Table 5. Specification of the type of transport according to particle size classes.

[37], reproduced by [22] and [52], is used to link the nature of sediment types to the energy of the transport medium.

Figure 4. Diagram C-M for sediments from the Ouaka River.

The results (Figure 4) of this relationship show that the dispersion of the samples analyzed from the Ouaka River marries the portions of bottom rolling deposits (NO) and bottom rolling deposits and graded suspensions (OP). Examination of these data shows that the sediments in this river are characterized by bottom roll generated by bottom turbulence, linked to hydrodynamics, corresponding to the movement of coarse elements such as gravel. The mode of transport by suspension corresponds to settling conditions in which the speed of the bottom is zero or too low to produce any sort of classification. These settling conditions are characteristic of silty-clay fractions.

So vs Md and So vs Sk diagrams, characteristic of the evolution of granulometric parameters, are used to determine the depositional environment of Ouaka River sediments. They are also used to characterize sediments from a paleoenvironmental perspective.

The So vs Md diagrams for the upstream sector (Figure 5(a)) and the downstream sector (Figure 5(b)) show the dispersion of sediments in the river domain. This means that 100% of the sediment grains originate from the Ouaka River watershed.

Figure 5. Diagrams So vs Md and So vs Sk: (a) and (c) upstream sector; (b) and (d) downstream sector.

Sediment dispersion in the So vs Sk diagrams, shows the importance of a river-like environment in the upstream (Figure 5(c)) and downstream (Figure 5(d)) sectors. Inputs from the dune domain have a negligible influence on sediment deposition in the Ouaka River.

The different values of the linear discriminant functions of the Ouaka River sediments are shown in Table 6. Analysis of Table 6 shows that the discrimination values between the wind and beach environments (Y1) range from 0.53 to 16.96, with an average of 10.27 in the upstream sector, and from 1.16 to 22.75, with an average of 22.19 in the downstream sector. These Y1 values are greater than −2.7411, indicating a beach deposition environment for Ouaka River sediments.

The values of the discriminant function between beach and shallow agitated marine environment (Y2) for the sediments analyzed range from 103.4 to 171.9 (mean 145.7) in the upstream zone and from 99.61 to 199.9 (mean 175.5) in the downstream zone. They are higher than Y2 = 65.3650, suggesting a shallow, agitated marine environment. With the exception of sample K11 in the Liwa locality (north of the study area), the value is 58.37 < 65.3650, indicating a beach deposit.

Figure 6. Environmental discrimination diagrams based on the linear discriminant function Y1 vs Y2: (a) upstream sector; (b) downstream sector.

Discrimination between shallow, agitated marine environments and river environments (Y3) is also calculated. The results of this calculation range from −12.99 to −8.49 (mean −10.93) in the upstream zone and from −38.48 to −7.98 (mean −36.81) in the downstream zone. These results are lower than Y3 = −7.419, indicating a dominant east-fluvial environment. All but two samples, B3 = −0.27 at Yamalé and K11 = −1.79 at Liwa, suggest the influence of a shallow marine deposit.

Finally, calculation of the discrimination function between fluvial and turbid environments (Y4) shows that in the upstream zone, 10 out of 17 samples exceed Y4 = 10 (i.e. 58.8%) and in the downstream zone, 12 out of 17 samples exceed Y4 = 10 (i.e. 70.5%). The results of this calculation indicate the dominance of a turbid depositional environment and a weak influence of a fluvial depositional environment.

The Y1-Y2 binary diagrams studied in the upstream and downstream sectors (Figure 6(a), Figure 6(b)) clearly showed the dominance of an agitated marine beach deposition environment.

The Y2-Y3 binary diagrams (Figure 7(a), Figure 7(b)), on the other hand, indicate an agitated fluvial environment, influenced by a shallow marine beach environment in agitated waters.

Figure 7. Environmental discrimination diagrams based on the Y2 vs Y3 linear discriminant function: (a) upstream sector; (b) downstream sector.

On the basis of the above, certain results demonstrate the influence of a marine and beach deposit in the middle of a continental environment. These results lead us to formulate two paleoenvironmental hypotheses regarding the presence of these sediments of marine origin:

These sediments could be carried by runoff-induced hydro-sediment transport. Indeed, the geographical location of the Ouaka watershed makes it a corridor for the passage of hydro-sedimentary flows; These sediments may be derived from limestone residues formed in a marine environment. The presence of carbonate formations in the study area is confirmed by the work of [39] and [5]. Geochemical analyses to demonstrate the presence of carbonate clasts in these sediments could be carried out at a later date.

Table 6. Values of linear discriminant functions for sediments in the Ouaka River.

In the sediments of the Ouaka River, the most abundant heavy minerals (Table 7) in the upstream part are opaque minerals (63% - 72%), tourmaline (9% - 13%), rutile (5% - 10%), sillimanite (3% - 8%), kyanite (1% - 5%) and garnet (1% - 5%). On the downstream side, after the opaque minerals (72% - 79%), we find tourmaline (6% - 12%), rutile (2% - 7%), sillimanite (2% - 5%), anatase (1% - 5%) and garnet (1% - 4%). Other heavy minerals such as zoïsite, augite, epidote, hypersthene and diopside are weakly/rarely observed in some samples.

With regard to the ZTR index, all the samples analyzed from the Ouaka river catchment had a ZTR < 75%, indicating immature to sub-mature sediments.

Analysis of Table 7 below shows that in the upstream zone, tourmaline, rutile and sillimanite are the main non-opaque heavy minerals, the most abundant in the samples. Anatase and kyanite are concentrated in the Gono, Takobanda and Bakala localities. Garnet showed a high frequency in samples B1 and B14. In the downstream zone, anatase is added to tourmaline, rutile and sillimanite to form the four main non-opaque heavy minerals. In addition, grenate tends to be more frequent in the southern samples of the study area, particularly in K3, K7, K10, K14 and K15.

Table 7. Distribution of heavy minerals in Ouaka River sediments.