Geology of Lom Pangar watershed is dominated by anciens syn-tectonic granites as well as migmatitic and garnetiferous gneiss crossed by NE-SW oriented faults ( Figure 1 ).

A band of gold mineralized shale, following the same orientation with mayor linar structures, appears in the SE of Lom Subwatershed and is subjected to prolific artisanal gold mining.

Particle size is a fundamental property of any sediment, soil, or dust deposit. It affects their entrainment, transport, and deposition, and therefore provides important clues to the sediment provenance, transport history, and depositional conditions (Pye & Blott, 2004; Wanogho et al., 1987) . Particle size analysis provides a valuable method for characterizing soils and sediments, aiding in comparison and environmental interpretation (Chappell, 1998) . Because of their large specific surface areas, particles like sand, silt, and clay, can play a significant role in adsorption processes in various environmental contexts as they have more sites available for adsorption. This means they can adsorb more molecules or ions. Finer particles, such as clay, have a larger surface area compared to coarser particles like sand. As a result, finer sediments often exhibit higher adsorption capacities (Garnier-Laplace et al., 1997) .

The influence of sediment grain size on the magnetic parameters and heavy metal contents is strong (Abrahim & Parker, 2008; Li et al., 2011) . Sediments could be a potential source of heavy metals that can be released into the overlying water via natural and anthropogenic processes, where they could harm the drinking water quality and human health (Pélédé et al., 2018; Suresh et al., 2012) . Due to the weathering of rocks and anthropogenic activities, elements are produced and transported by Rivers and redistributed to adjacent lakes and their sediments. One of the most important processes affecting this migration of elements is interaction with geological materials. The accumulation and distribution of elements depend mostly on the characteristics of the geological material such as mineral species and grain sizes. The fine fraction (less than 62.5 μm) is consistently more contaminated than bulk samples (Taylor, 2007) . Metals have low solubility in water and are adsorbed and accumulated on bottom sediments (Jain et al., 2008) .

Lake sediments serve as fundamental components of our environments, providing essential nutrients for living organisms. They also act as sensitive indicators for monitoring contaminants, serving as both sinks and carriers for pollutants within the aquatic environment (Bai et al., 2011) . Therefore, the analysis of lake sediments plays a crucial role in evaluating the pollution status of aquatic environments.

The hyporheic zone, which refers to the transitional area between the surface water of a river and the groundwater beneath ( Figure 2 ), is an ecologically rich and dynamic region where surface water and groundwater interact.

This zone typically occurs in the sediment or substrate beneath the waterbed that are characterized by a unique set of physical, chemical, and biological conditions. The hyporheic zone plays a vital role in various ecological processes (Schmutz & Moog, 2018; Seliger & Zeiringer, 2018; Touchette et al., 2015) , including nutrient cycling, filtration of contaminants, and habitat for a diverse array of aquatic organisms. Additionally, it serves as a refuge for many species during periods of environmental stress, such as drought or extreme temperatures.

Sediments can damage turbines and other mechanical equipment through erosion of the coating on the blades ( Figure 3 ), leading to surface irregularities and more serious material damage (Azrulhisham & Azri, 2018; Rai et al., 2020) . One of the factors contributing to the gradual decline in the performance as well as the failure of hydropower plants over the years is silt erosion (Tandel et al., 2023) .

Many other authors have also worked on the impact of sediment on turbines and mechanical equipment of dams (Sharma et al., 2021; Sharma & Gandhi, 2023; Tandel et al., 2023; Tarodiya et al., 2022) .

Abrasive particles in the run-of-river hydropower plants reduce unit efficiency, increase maintenance costs, and may cause turbine downtime and associated production losses.

One of the factors contributing to the gradual decline in the performance as well as the failure of hydropower plants over the years is silt erosion. To deal with this situation during the design, operation, and maintenance of hydro plants, knowledge of turbine wear needs to be improved with regard to relevant suspended sediment parameters including the suspended sediment concentration (SSC) and particle size distribution (PSD). Measurements of SSC and PSD were conducted at the intake weir and settling basin of a run-of-river hydropower plant in Bentong, Malaysia, using Laser In-Situ Scattering and Transmissometry (LISST) instruments. The results indicate that all samples were predominantly composed of silt and fine sand particles, with a mean size range of less than 50 μm and concentrations ranging from 10 to 30 mg/l (Azrulhisham & Azri, 2018) . The identification of the continental origin of the deposit suggests the need to develop solutions in the upper watershed to reduce the siltation rate (Amarjouf et al., 2015) .

Laser diffusion is widely used as a particle size technique for materials ranging from a few hundred nanometers to several millimeters, typically from 0.02 μm to 2000 μm ( Figure 4 ).

When it comes to studying environments that contain a significant quantity of silt and clay, particle size analysis by laser diffraction was mainly developed since the end of the 1990s is the technique preferred to mechanical analyses based on the interactions between light and matter (Delanghe et al., 2017) . The angles of light scattering by the analyzed particles depend on the particle size. It is thus possible to obtain in a few minutes an estimate of the particle size proportions of particles present in different types of samples (solid or liquid). This method also makes it possible to work on smaller sample volumes than sieving.

The study area is shown in Figure 5 . It belongs to the department of Lom and Djerem, East Region of Cameroon, at about 120 km N of the capital city of Bertoua. It is located between 05˚37'38" - 07˚22'40"N latitudes and 13˚26'25" - 14˚05'19"E longitudes, with an average altitude of 700 m.a.s.l. The site is part of the savannah tropical (Beck et al., 2018) , commonly called sub-equatorial “Guinean” climate. Heavy rains fall from August to mid-December. The mean annual rainfall is 1578 mm/year. Temperatures are relatively constant over the year with an annual average of 23˚C ± 1˚C. Winds are low in the study site both in frequency and intensity. The “calm” period (≤1 m/s) represents 44% of the year. South-west winds dominating in the rainy season represent 26.3% of the time and the northern winds of the dry season 7.7%. Sampling sites have been chosen according to the longitudinal zonation of water quality conditions in the future reservoir with smooth shapes).

The study site consists of the transitional and lacustrine sections of the Lom Pangar. It belongs to a whole program of the dam monitoring with 8 sampling stations. The choice of the sections and stations used in this paper is because of their vicinity to the outlet of the dam that carries the turbines and mechanical equipment, as well as low water energy.

The other sections of the dam include the riverine with stations L1 and P1, and the downstream stations with D1, D2 and D3 are not considered in this article as the focus is on sediments reaching the outlet. L represents the Lom River. It’s a major contributor to the Sanaga where the downstream stations D1, D2 and D3 are going. It rises in the Central Africa Republic, in the southeastern edge of Adamaoua, towards the 1400 m ( Figure 6 ) and 70 km east of Meiganga, and flows down through the Adamaoua plateau with a difference in altitude of 590 m and slope up to 20%, to reach Sanaga River at 380 km downstream.

Its flow rates in low hydrologic regimes vary between 10 to 42 m³/s over the year. Pangar River (P), the main tributary of Lom River (L), rises at the foot of the Ngaoundal in the western part of the Adamawa Plateau. P rises at over 100 km far and joins the Lom at around 5 km upstream of the site of the dam outlet.

Sampling was conducted using the standard method outlined by AFNOR (2009) and AFNOR ISO 13320-1 (2009). A plastic shovel was used to collect swallow sediments on the tidal zone, at the water borders from 04 to 09 April 2023, with the beginning of the dam filling period. This period was chosen because of the closure of the dam weirs that further significantly reduces the speed of water flow to deposit more finer particles. Samples were collected from 6 points across 3 stations representing the transitional and lacustrine sections of the dam. Upon collection, each sample was immediately placed into sampling bags in bulk and gently squeezed to remove the excess of water. Subsequently, samples were properly conditioned and transferred to the laboratory of Physico-chemical Analysis in Environmental Engineering, at the Department of Hydraulic, Hydraulic Machines, and Environmental Engineering of the Faculty of Energy at the Politéhnica University of Bucharest, Romania.

Each sample was weighed and its pH measured using a pH-meter. Subsequently, the samples were dried in an oven at 40˚C to remove moisture, and their weight was recorded every 4 hours until the variation in mass became negligible. Dried samples were crushed ( Figure 7 ), and sifted to obtain a fine and homogeneous powder. Sporadic large fragments and visible debris were picked out manually and the processed samples stored in clean and labeled containers. A fraction of 3 grams from each sample was transferred to the laboratory of Science and Engineering of Oxide Materials and Nanomaterials within the same university. Here, the samples underwent sieving using a 3 mm mesh to obtain the desired particle size.

The granulometric analysis was done on the finer sample fraction mixed in ethyl alcohol on the Laser Diffusion Technique with the Malvern Instruments Mastersizer 2000 © Measuring unit ( Figure 7 ), following the manufacturer’s prescription and standard methods (Delanghe et al., 2017; Fournier-Sowinski et al., 2012) . Bulbs are removed in the system and the dispersion of aggregate into an ethyl alcohol milieu is done with a pump rotating at 2500 r/min. Three measures were done and the average value was taken.

These data offer a comprehensive understanding of the characteristics and behavior of particles in a solution, which is crucial for particle characterization across diverse fields, including environmental science. Table 1 presents the detailed insights gleaned from laser diffusion samples analyzed during the study.

The obscuration percentage provides insight into the degree of light scattering or absorption by particles in the solution. The optimal obscuration range that provided the best measurement performance is set in the table and is between the normal known range (10% - 20%). In this case, values above 20% suggest a higher concentration of particles or larger particle sizes, which can impact the clarity or opacity of the solution. Thus, the highest value (18.5%) in the sample E3-P2G may be due to the influence of it coarser grain tendency observed at the right bank of the River Pangar.

The Particle Refractive Index (PRI) indicates how much particles bend or refract light. It appears consistent at 1.650. Consistent values across samples suggest uniformity in particle composition or size, while variations may indicate differences in particle properties. Thus, each sample analysed is consistent in its particle properties such as its mineral component.

The absorption coefficient represents the extent to which particles absorb light at a specific wavelength. Its value is 0.1 and appears low and consistant for all samples. Higher values of this parameter suggest greater light absorption, which could be due to factors like particle composition or size distribution. It corroborates with the above-mentioned PRI that displays each sample as consistent in properties.

The Dispersant Refractive Index (DRI) is consistent here with a single low value of 1.330, close to the above-mentioned PRI for all samples which is also a single low value 1.650 for all sample. For a dispersant medium, it shows how light interacts with suspended particles and helps appreciate the quality of the results of the lazer diffusion. Matching or closely aligned values between DRI and PRI optimize dispersion stability and minimize light scattering. Thus the ethyl alcohol used as dispersant in the analysis optimised the stability of the milieu for best refraction of particles.

This parameter indicates the level of deviation between observed and expected values in a model or analysis. Values range between 0.639 on a sample E5-L3G and 4.914% on a sample E1-LPG. Lower values suggest a better fit of the model to the data, indicating a more accurate characteristic of particle properties. Thus, a relatively high value on sample E1-LPG may be due to its relatively coarse grain tendency.

The Volume concentration reflects the proportion of the solution occupied by particles. Changes in concentration can affect various properties such as viscosity, optical properties, and stability of colloidal suspensions. This parameter is very low and ranges between 0.0095% and 0.0170%. Thus, all sample analysis milieu are close in terms of concentration and fit with initial analysis parameters.

Specific surface area measures the extent of particle a surface relative to its mass. This represents the total surface area per unit mass of the sediment particles. It is an important parameter for understanding various physical, chemical, and biological processes occurring at the sediment water interface. Fine-grained sediments generally have higher specific surface areas compared to coarse-grained sediments due to their larger surface area per unit mass. Generally, higher values indicate larger surface area available for interactions, such as adsorption, chemical reactions, or agglomeration, which can influence particle behavior and performance in various applications. But this is not the case at the Lom Pangar given their range between 0.765 m²/g for E4-P2G to 1.21 m²/g in E6-L3D with high-value trends in the left bank (finer grain size trend). This corroborates with the fact that the specific surface area of sediment can vary widely depending on many factors such as particle size distribution, mineralogy, surface roughness, and porosity. Thus, it appears that geology with particle roughness and mineralogy may have played a major role on the parameter and this aspect keeps attention to carrying further studies to understand.

Table 2 displays the insights data into the particle size distribution of the sample being analyzed such as quartiles, percentiles, and statistics parameters.

Not Applicable (D₉₅ is neglectable as well as indices using it when clayed fraction is greater than 10%).

The median diameter represents the particle size at which 50% of the cumulative distribution is below that size. It describes the central tendency of the particle size distribution and gives an estimation of the average coarseness of the sediment by integrating all the particle size classes. It varies between 8.761 µm in E6-L3D and 15.751 µm in E4-P2D µm. Thus sediments analyzed are dominated by clays and finer silts. However, as mentioned in paragraph 4, this parameter is perfect only with the unimodal distribution and might deviate if particule are bimodal or multimodal distribution (White et al., 2007) .

The mode defines the size of sieves that have a maximum particle. It represents de size with the maximum concentration ratio of the particles. In sediment granulometry, the mode refers to the particle size that appears most frequently within a sample. Unlike the median or mean, which represent central tendencies, the mode identifies the most common size class or size range in the sediment. River Pangar mentioned the lowest values at the right bank E3-P2G (7.59) and the higher values at the left bank E4-P2D to 19.95. However, the Particles size distribution frequency curve ( Figure 10 ), described later on section 4.4, displays multimodal particle sizes explaining the tendency of particle to differentiate more or less into specific class sizes.

The surface weighted mean particle size is expressed as the surface mean diameter of particles. It is particularly useful when considering properties of materials that are influenced by surface area, such as reactivity, dissolution rate, or surface adsorption. It expresses the average particle size of the sample from the Equation (1):

$D\left[3,2\right]=\frac{{\displaystyle \sum \left(D_i^3\cdot S_i\right)}}{{\displaystyle \sum \left(D_i^2\cdot S_i\right)}}$ (1)

where:

D_i is the diameter of the individual particle i and S_i the surface associated with particle i.

D[3, 2] values are very low and range from 4.960 µm in the E6-L3D to 7.839 µm in E4-P2G confirming the dominant fraction of clays and finer silts in sediments.

In application, this formula is highly accurate only in the conditions of spherical particles and homogenous material (particles and composition) and narrow size distribution (Li et al., 2005) . When samples deviate much from these conditions, accuracy may also deviate. Further, the accuracy of the analysis technique is to be considered, especially when particles tend to agglomerate or aggregate and make the interpretation of the results more complex as it may not be accurately reflecting the true D[3, 2]. However, despite these limitations, it remains a valuable parameter for characterizing particle size distributions, especially in applications where surface area plays a crucial role. It’s often used in conjunction with other particle size parameters to provide a comprehensive understanding of particle size characteristics.

The Span measures the width of the particle size distribution and indicates the uniformity or spread of particle size within the sample. It is calculated in the Equation (2).

$\text{Span}=\frac{D_{90}-D_{10}}{D_{50}}$ (2)

where:

D₉₀ is the diameter below which 90% of the cumulative volume distribution is found,

D₁₀ is the diameter below which, 10% of the cumulative volume distribution is found,

D₅₀ is the median diameter or the diameter below which 50% of the cumulative volume distribution is found.

This dimensionless quantity varies between 3.213 in E6-L3D and 15.881 in E3-P2G. A smaller value indicates a narrower particle size distribution, suggesting that the majority of particles in the sample are of similar sizes. The high-value trend of the Span is registered in the right band of the dam with the Pangar tributary. These observations corroborate with other parameters. As per the D₅₀, this parameter is perfectly accurate in the case of unimodal distribution and less in the multimodal differentiated distribution as analysed samples.

Uniformity is assessed by examining how closely the particle size distribution resembles a single peak or distribution curve, indicating a more uniform distribution of particle sizes. It refers to the degree of consistency or homogeneity in particle size distribution class within the sample. A more uniform particle size distribution suggests that the sediment sample contains particles that are relatively similar in size, without significant variations or outliers. Values greater than zero here, vary from 1.04 at E6-L3D to 4.19 at E3-P2G. The high value in the right bank of the dam showing the heterogeneity trend of the grain size might be due to the coarser particle trend noticed on that bank. It should be noted that the geology of the Pangar subwatershed is mostly crystalline and may have produce the coarser trend of sandy particles than the Lom containing a band of gold shale-bearing, source of artisanal mining, but both subwatershed have the same elevation level. The Lom spring source is in the Central Africa Republic.

Fraction greater than 3000 µm obtained after sieving is displayed in Table 3 .

LP, P and L stand for study stations. E stands for collected samples, Cumul stands for cumulative granulometry.

This fraction is negligible in regard to their low percentage in the total bulk sample ( Figure 8 ). A relative high concentration (11%) is met at the sample E6-L3D. The point is located less than 1 km downstream of the River access point from the Ouami village. The relatively high concentration should be due to the coastal erosion from population activities, mainly board river access for a river crossing for fishing and traveling in that section of the river. Apart from that sample, it appears that the right band of the dam is seen to provide coarser particles than the left band as mentioned.

The low proportion of coarse particles can be explained by the change in transportation conditions. As they are mainly transported by cartage, a transport threshold should be reached, particularly for non-cohesive sediments (Powell & Soares Frazao, 2017; Swartenbroekx et al., 2013; White et al., 1975) . This included the flow rate of the river, the slope of the topography, the falling speed of particles, the roughness of the water bed and the characteristics of the particle itself. Such particle characteristics are the equivalent diameter, density, particle morphology and granulometric distribution.

The Lom Pangar dam is located in the low altitude of the watershed with a very low slope ( Figure 6 ). This combined with the fact that sampling was done at the beginning of the dam filling, contributes to slow the river flow rates towards the subcritical regimes with high loss of its specific energy. This phenomenon was observed in Dupuyer Creek, Montana, USA (Whitaker & Potts, 2007) and in the Lower Mekong River in Viet Nam (Bravard et al., 2014) .

This constitutes the major fraction of Lom Pangar sediment. They are composed of Silt > Clay > Sand, as displayed in Table 4 .

LP, P and L stand for study stations. E stands for collected samples, Cumul stands for cumulative granulometry.

In the fine particle fraction, silt is dominant with over 70% of the total sample ( Figure 9 ). Silt is the dominant particles size reaching the Lom Pangar dam. This shows the bed of the reservoir being clogged. Clay varies between 6% to 12%. Thus the percentile D₉₅ is insignificant as well as parameters SI, SKI, KG integrating it (Le Nindre & Serrano, 1988) .

Grain size curves are used to characterize the physical properties of sediments. They can provide information on the dynamics of geological and environmental processes.

The particularity of the Lom Pangar fine sediments is its multi-modal particle size distribution trend, presenting one major pic and many minor more or less shaped pics on the frequency curves ( Figure 10 ).

For each curve the highest pics represents the major mode, indicating a class of the size where accumulated a maximum number of particle. There are many more individualized lower pics on each curve going from E6-L3D with 3 poorly individualized lower pics at the left of the mayor pic, to E1-LPG with 2 well individualized lower pics at the right of mayor pic. These lower pics indicate that particles size are differentiated and are likely accumulated on either size of the major pic to form minor modes of minor class sizes.

This demonstrates the heterogeneity of sediments, indicated by a mixture of finer grains to sand, which corroborates with the relatively high uniformity parameter mentioned earlier. Such variability is a characteristic of fluvial sediments. The heterogeneity appears to be more pronounced in the right bank of the dam compared to the left bank.

The cumulative curve of sediment granulometry is displayed on Figure 11 .

It appears that all curves have the same trend with irregular highly stretched s-curve confirming the multi-modal frequency curve of the riverine sediments paleoenvironment. Samples from Pangar River tend to deviate at the top of the curve, indicating the relative coarse grain of that River branch.

Various indices parameters characteristics are computed and displayed in Table 5 . As mentioned in section 4, It more concerns elementary statistical analysis approaches, such as first and second statistical moments (mean, mode, skew and kurtosis) or percentiles (D₅₀, D₈₄ … etc.). However, the theoretical basis for such use remains weak because these approaches assume a lognormal and unimodal Gaussian sedimentary distribution, which is rarely like in this case of Lom Pangar. Semi-quantitative indices like sorting indices (Trask index, Inman index, Folk and Ward index, etc.) are often temporary solutions and tend to be correlated with the standard deviation. Nevertheless, despite their limitations, central tendency values (mean, median), shape indices (skewness and kurtosis) and dispersion indices (sorting, etc.) remain dominant in the interpretation and graphical representation of particle size distributions (Duquesne & Carozza, 2023; Folk, 1966; Fournier et al., 2014; Mercier, 2013) .

M₀ ranges from 9.47 at E1-LPG to 27.43 at E3-P2G. Values are low but relatively higher on the right bank of Pangar. It shows a lower spatial heterogeneity, indicating fewer differences between the D₂₅ and D₇₅ percentiles across the site. These values can be kept as a baseline to assess in the same condition, changes in spatial variability over time in the study areas keeping in mind that the right bank of the dam is relatively higher, due to the relatively short distance with the sediment source, compared to the Lom. This situation was also found in Osi River in southwestern of Nigeria (Lawal et al., 2017) .

The sorting index S₀ ranges from 0.31 to 0.53 mm showing that sediments are very well sorted in the lacustrine and transitional sections of the Lom Pangar (Trask et al., 1932) . According to the Trask sorting coefficient, this dispersion parameter is as worse as the value of S₀ is higher and parfait for values around 1, following the ranges of Table 6 .

As well as M₀, MΦ measures the heterogeneity of grain size within sediments. Negative values in Table 5 indicate a relatively uniform distribution of values across the study area with minimal variability between the lower D₁₆ and upper D₈₄ percentiles.

This is the Quartile Deviation Phi (QDΦ) expressing how grain size varies within the sample. It refers to the variability of sediment properties across different spatial scales. It is determined by the Krumbein-Index with the Equation (3).

$QD\Phi ={\log }_{10}{S}_0$ (3)

where S₀ is the Sorting-Index parameter.

For Lom Pangar, its value ranges from 0.906 to 1.683, confirming the low variability of the sediments grains size. This is explained by the well sorting particles found in the sorting index S₀.

In addition to the position and dispersion parameters, the asymmetry parameter, measuring the shape of the distribution on either side of the median, is necessary for the correct definition of a particle size curve. For Trask-Index, Skewness Sk varies from 61.908 to 242.5 µm.

For the Inman Index, SkΦ are very low and vary from −0.168 to 0.303 showing the high proposition of fine material in samples. Negative values correspond to dominance of finer sediments. The Inman Index also confirmed that sediments are fluviatile.

Particle size fractions were plotted on the Spark diagram ( Figure 12 ).

It appears that all sediments fall in the silted and sandy silted zone confirming the finer size of Lom Pangar sediments and corroborating with all previous parameters. As this method is based on arbitrary choices for class limits and sub-categories, it shows that categorization is simple and quick but works properly when dealing with samples of the same origin. If not it may become even challenging (Folk & Ward, 1957) in finding the most suitable texture classification scheme that maximizes the contrast between samples (Duquesne & Carozza, 2023) .