The determination of the leaching fractions required regular monitoring of the daily reference evapotranspiration (ETo) of the environment and associated calculations over the entire study period. Thus, climatic data from the meteorological station of Lake Kalassane were used to evaluate ETo of the environment according to the Penman-Monteith method described in Equation (1).

ETo = 0.408 Δ ( Rn − G ) + 900 T + 273 u 2 ( es − ea ) Δ + γ ( 1 + 0.34 u 2 ) (1)

where ETo is the reference evapotranspiration (in mm/day), Rn is the global radiation (in MJm²/day), G is the soil heat flux (in MJ/m²/day), T is the daily mean air temperature at a height of 2 m (in ˚C), v is the wind speed at a height of 2 m (in m/s), es the saturation vapour pressure (in kPa), ea the vapour pressure at temperature T (in kPa), Δ the slope of the saturation vapour pressure curve (in kPa/˚C), γ the psychrometric constant (in kPa/˚C) and 900 a constant for a daily time step.

The applied leaching fractions were calculated on the basis of the ETo of each previous day. They are represented by the following Equations (2)-(4)

T 0 = ETo (2)

T 1 = ETo + ( ETo ∗ 10 100 ) (3)

T 2 = ETo + ( ETo ∗ 20 100 ) (4)

where ETo is the reference evapotranspiration in mm, T0 the first leaching fraction in mm, T1 the second leaching fraction in mm and T2 the third leaching fraction in mm.

Following the estimation of the leaching rates, the results obtained were converted into litres and then extrapolated to the plot area (4 m²) to obtain the volume of water to be applied to each elementary plot. The application of the leaching doses was carried out using a 10 L graduated bucket, which served as a tool for measuring the volume of water to be applied, and a watering can to ensure the correct distribution of the leaching fraction on the plot to be treated.

Soil moisture analyses were carried out in the laboratory using the gravimetric method on a weekly basis. A total of 144 soil samples were taken and analyzed at a rate of 18 samples per week. The samples were taken from the depth profile of the soil on the 0 - 10 cm and 10 - 30 cm horizons. Glass capsules were used for weighing and drying the samples. Weighing was carried out using a 0.1 g precision balance. Similarly, an oven heated to 105˚C for 4 hours was used for drying. Thus, for each sample, a repetition of two weighings is made: 1) a first weighing of the wet mass of the sample and 2) a second weighing of the dry mass of the sample after the drying in the oven. Thus, the water content of the sample is obtained as a result of the ratio between the water mass of the sample and the dry mass of the sample (Equation (6)).

H = m h u m + m s e c m s e c (6)

Source: [26].

where H is the water content (%), m_hum is the mass of the wet sample (g) etm_sec the mass of the dry sample (g).

To assess the existence of a relationship between salinity and soil moisture, simple linear regression models were fitted to the EC and soil moisture data at the 5% significance level. For each model, EC represented the variable to be explained and soil moisture as a function of treatments, the explanatory variable.

Irrigation waters naturally contain mineral salts at defined composition and concentrations. However, in this study, in situ salinity analyses were carried out on irrigation waters through the evaluation of electrical conductivity. The result is presented in Table 4. The analysis of the electrical conductivity of the irrigation well water indicates a slight salinity of the latter, averaging 2.283 dS/m. A similar result was reported by [11]. This slight salinity of the groundwater noted in the basin is probably related to the nature of the water table. Indeed, this groundwater belongs to the Quaternary sands water table, which, on the one hand, suffers from overexploitation marked by a progressive drawdown in depth of its piezometer and, on the other hand, from the saline intrusion of sea water [11]. This phenomenon of saline intrusion can sometimes extend several kilometers inland. Thus, it spreads salt water over the land, resulting in the contamination of groundwater near the sea.

However, the use of this water in irrigation is in principle without consequences on the soil and crops as it is below the salinity threshold of irrigation water (3 dS/m) retained by FAO [23]. For example, [27] confirmed that the irrigation water salinity level of 3 dS/m can ensure good yield of tomato when the irrigation rates are daily. Similarly, [28], who tested several levels of irrigation water salinity combined with different frequencies on a silty-clay textured soil, indicates that irrigation with slightly salty water at a high frequency is favourable for good plant development.

Although irrigation with moderately salty water can be practiced, special attention is needed especially in conditions of high evaporation where irrigation is intermittent for fine-textured soil [29]. In this sense, irrigation management techniques with salty water have been proposed by researchers [20] [27]. Thus,

based on the sensitivity of plants to salinity at the growth stage, researchers suggest the use of fresh water at the initial stage of plant growth. A mixture of agricultural drainage water with good quality irrigation water is also considered as an alternative as well as alternating good quality irrigation water with brackish water [27] [30].

The results of the soil EC analyses (Table 4) indicate a generally light soil salinity. However, a difference in salinity characterized by an ascending vertical gradient is observed on the soil profile. The spatial distribution of soil salinity is also marked by heterogeneity between the different treatments. The EC of the surface horizon (0 - 10 cm) averages 3.65 dS/m while that of the deep horizon (10 - 30 cm) is 2.72 dS/m (Table 4). On the other hand, the soil EC is 3.12 dS/m in T0, 4.38 dS/m in T1 and 3.84 dS/m in T2 (Figure 3).

These results corroborate those of [11], who prove the existence of slightly salty soils in the lowlands near the ocean in the Niayes of Saint Louis.

The origin of salt in the area could be attributed to the climatic conditions of the environment and the estuarine position of the Gandiolais area. Indeed, the area has suffered a rainfall deficit marked by a considerable decrease in annual rainfall, which fell from 684 mm in 1960 to 271 mm in 2009 [31]. In addition, combined with other deleterious phenomena, notably the opening of the rift and the rise of the salt wedge, the area is facing phenomena of transport by mass flow of soluble salts from the soil solution to the surfaces, thus favoring the concentration and accumulation of salts after water evaporation [12].

The difference in salinity noted on the soil profile has already been observed by [12] [13]. It would be due to the dynamics of the groundwater table and the high evaporation of the environment [32]. Indeed, in environments where the water table is shallow, notably in the Niayes area [33], soluble salts are rapidly put back into movement in the whole soil profile by the rise of the water table or

by capillary action, due to direct evaporation or through vegetation [21]. Consequently, the concentration of salt varies from one horizon to another in the same profile as a result of cation exchange reactions between the soil and its solution, but also as a result of the dissolution rates and speeds of the various ions and salts [21].

The heterogeneity of the spatial distribution of salinity can be explained by the topography which influences the natural leaching of salts from the soil, but also the transfer of solutes. In addition, vegetation can also act as an agent in the spatial distribution and modification of the soil due to surface modifications caused by shade and the level of evapotranspiration on the one hand, and modifications at the soil interface influenced by plant roots and contributed organic matter on the other [34].

Although salinization reduces soil and crop productivity, lower soil salinity is not a limiting factor for plant development. For example, [35] evaluated the performance of onions under saline stress. The results showed that growth and bulb yield of cultivars were only affected at a salinity threshold of 4 dS/m. The same experiment repeated on tolerant cultivars of safflower (Carthamustinctorius L.) indicated an inhibition of seedling growth at a salinity threshold of 5.1 dS/m.

The richness of the soil absorbing complex in sodium ions and the associated risks of physical degradation are characterized by the soil sodium absorption rate (SAR).

The SAR describes the imbalance in the soil absorption complex between divalent cations (calcium and magnesium) and sodium [5]. It also makes it possible to define, at a threshold value, the soil sodicity term [36]. In this study, the results of the soil SAR assessment indicate a low sodium uptake rate for both horizons: 1) 3.21 meq/100g for the 0 - 10 cm horizon and 2) 3.29 meq/100g for the 10 - 30 cm horizon. This low level of exchangeable sodium is justified by the abundance of divalent cations (Ca²⁺ and Mg²⁺) on the absorbing complex. Calcium and magnesium are respectively estimated at 7.34 and 7.35 meq/100g for the 0 - 10 cm horizon and at 14.99 and 7.13 meq/100g for the 0 - 30 cm horizon, while sodium and potassium are respectively estimated at 2.22 and 16.69 meq/100g for the 010 cm horizon and 7.66 and 1.00 meq/100g for the 10 - 30 cm horizon. The simple soil facies illustrated by the piper diagram (Figure 4) shows this dominance of divalent cations (Ca²⁺ and Mg²⁺) over monovalent cations (Na⁺ and K⁺). It should be noted that weakly hydrated ions, in this case Mg²⁺ and especially Ca²⁺ are more energetically retained at the absorbing complex than strongly hydrated cations such as K⁺ and Na⁺ [13].

Calcium is naturally present in arid and semi-arid areas due to relatively low leaching. In contrast, in coastal areas, this abundance is explained by the nature of the CaCO₃-rich parent material [37]. However, in Gandiolais, described as a carbonate environment [13] dominated by sandy soils [11] [12], the low rate of sodium uptake could be related to the nature of the soil. Indeed, starting from

the saline character of the soil, which is less sensitive to destructuration due to its flocculating power and the sandy texture of the soil, residual or irrigated sodium ions are likely to be leached deep into the soil during irrigation or rainy seasons.

The evolution of soil moisture content under the three leaching treatments (T0, T1, and T2) is illustrated in Figure 7. Initially, the soil moisture values indicate a disparity in mass water content at the soil surface and at depth. It is respectively 12.2%, 12.75% and 10.88% for T0, T1, and T2 in the 0 - 10 cm horizon (Figure 7(a)). In the 10 - 30 cm horizon, a much higher moisture content is initially noted compared to the 0 - 10 cm horizon, i.e. 15.5%, 17.81% and 14.3% for T0, T1 and T2 respectively (Figure 7(b)).

Table 6 summarizes the rates of variation of soil moisture according to the leaching treatments. Indeed, the monitoring of soil moisture (Figure 7) shows a clear overall increase in soil moisture content for the different horizons and according to the leaching rates applied. However, the detailed analysis of Figure 7 indicates that the increase in soil moisture slowed down in the sixth week of leaching treatments and tended towards saturation. This trend is noted both on the 0 - 10 cm horizon and on the 10 - 30 cm horizon.

In the 0 - 10 horizon (Figure 7(a)), the greatest increase in soil moisture was noted with T2, with a rate of variation of 41.05% (Table 6), followed by T1, which showed a smaller increase in gravimetric moisture than the T2 treatment, with a rate of variation of 27.36%. T2 had a smaller increase in soil moisture content of all treatments, with a variation rate of 23.92%. In the 10 - 30 cm horizon (Figure 7(b)), a similar trend is noted for the leaching treatments applied. However, soil moisture is higher at depth than at the soil surface. The rates of change recorded are 33.96%, 25.25% and 42.03% for T0, T1 and T2 respectively.

The results show that the soil is likely to have a high water retention capacity. It reacted favourably to the various continuous leaching inputs throughout the study period. This can be explained by the soil properties such as texture, structure, mineralogical composition and organic matter composition. These parameters play an important role in the hydric behaviour of the soil by acting on the size of the soil pores as well as on the ability of the soil to retain water to a greater or lesser extent [40]. In addition, the difference in water content noted between the treatments is explained by the difference in the volumes of leaching fractions applied for each treatment. Indeed, the leaching rates were higher in T2 followed by T1 and T0. Thus the soil reacted according to the volumes of water applied, marked by a higher gravimetric moisture in T2, followed by T1 and T0.

Furthermore, the decrease in water retention that occurred in the sixth week of treatment could be explained by the fact that the soil reached the re-wetting point from this week onwards. In addition, a difference in moisture content between the surface horizon and the depth of the soil was also noted. This can be explained, on the one hand, by the transfer of water from the soil to the atmosphere by evaporation under the effect of temperature and, on the other hand, by the infiltration and retention of water in the depths under the effect of gravity.

To assess the level of salinity influence of soil moisture in relation to leaching fractions, simple linear regression models were fitted with soil moisture and electrical conductivity data (Figure 8). In the 010 cm horizon (Figure 8(a)), the prediction results show an overall significance for all models, i.e. a value of 0.027 for T0, 0.01 for T1 and 0.007 for T0. Thus, the fitted regression lines show negative slopes for all treatments, indicating a decrease in salinity with increasing humidity. However, the evaluation of the coefficients of determination R² shows a similarity in the predictions between models T1 and T2 with R² values of 0.84 and 0.86 respectively. Furthermore, the models predict that the variability of soil salinity in T1 and T2 is explained by 84 and 86% respectively by the leaching treatments, whereas in T0 it is estimated at 74%.

In the 10 - 30 cm horizon (Figure 8(b)), all fitted models are also statistically significant with p-values of 0.057, 0.012 and 0.006 for T0, T1 and T2 respectively. The slopes of the fitting lines are negative indicating a decrease in soil salinity as soil moisture increases with leaching fractions. Furthermore, the coefficients of determination (0.64 for T0, 0.83 for T1 and 0.88 for T2) of the models indicate that 64%, 83% and 88% of the variability in salinity is explained by the variability in soil moisture according to the leaching treatments.

The decrease in soil salinity is inversely correlated to the moisture leaching levels. Soil salinity decreased in all treatments but at relatively different levels. This decrease is explained by the principle of leaching which emphasizes the need for water to leach salts beyond the root zone. Indeed, the leaching fraction corresponds to the part of the applied irrigation water that is leached below the root zone. In fact, the higher the leaching fraction, the more likely it is to decrease soil salinity. Furthermore, the response of the soil to leaching treatments is dependent on a number of parameters including the salinity level of the irrigation water, which must be lower than that of the soil, the hydraulic conductivity of the soil and the frequency of irrigation [38].