Measurements were conducted near the village of Bangou Koiré (13˚36'3.52"N, 1˚52'44"E), 30 km northwest of Niamey (Figure 1). This semi-arid environment is characterized by two types of winds. The monsoon is a wind from the southwest with an average direction of N213˚34˚. It blows mainly between April and September, with maximum daily speeds varying regularly between 8 and 18 m∙s⁻¹. The harmattan generally begins to blow in October and retreats around March. Its direction is approximately N131˚63˚ and maximum daily speeds are regularly below 8 m∙s⁻¹ (Figure 2). Cumulative annual rainfall is approximately 500 mm between May and September [24].

Measurements were taken in a field cultivated with millet (Pennisetum glaucum) (Figure 3). The field was cultivated from 2010 to 2018, then left fallow in 2019 before being cultivated again in 2020, mainly with millet combined with beans. Millet is usually sown in pockets in June after the first rain. The millet

Figure 1. Localization of the study site in the Sahel (top) and land uses around the measurement sites at Bangou Koiré (bottom). SE are anti-erosive layouts of the grassy surface type (Google Earth image, 2019; modified).

Figure 2. Daily average wind direction and daily maximum speeds (over 5 minutes) measured at 2.5 m above ground level at Bangou Koiré from April 2020 to March 2021. North is 0˚.

Figure 3. Illustrations of vegetation cover in the Bangou Koiré field between 2020 and 2021.

field is then weeded once in July and again in August. The millet ripens and its ears are harvested in September. After the harvest, the field is opened to livestock, which graze on the crop residues. The field surface is then dominated by millet and bean plants, as well as weeds. Vegetation cover gradually increases, reaching its maximum at the end of the rainy season in October. As the stems and leaves of cowpeas are harvested for fodder, the field surface is mainly covered with millet residues during the dry season, which are either lying down or standing up. Soil cover by crop residues has decreased and become very low since February (Figure 3).

Wind transport over a surface can be defined by horizontal flux. Several authors have measured horizontal flux on different land uses in the Sahel [24][25][27][28][32][43]-[46]. Wind transport was determined using BSNE (Big Spring Number Eight) sand traps [47] (Figure 4). BSNE traps have been widely used to characterize erosion flux in southern Niger [25]-[28][32][48], and China [49]. Three BSNE are installed on a mast at 5, 15, and 30 cm above the ground. Five masts equipped with three BSNE are placed in the millet field, four of which form a 30-meter square, with the fifth placed in the center.

Figure 4. A three BSNE mast installed in the Bangou Koiré millet field: A. side view; B. front view: 1, 2, and 3 are respectively BSNE at the bottom, middle, and high levels above the soil; 4: mast; 5: weathervane.

The measurements were performed over a year, from April 2020 to March 2021. Sediments eroded by wind and trapped in the BSNE were collected every two weeks. Twenty-four collections of sediments were made. The collection interval for eroded sediments represents the main constraint. This constraint does not affect the seasonal or annual cumulative flux but masks the temporal variability. However, it should be noted that a fixed duration for the collection of eroded sediments is common in studies of wind erosion of soils in the Sahel [26][42].

After every two weeks, the sediments trapped in the BSNE are collected and weighed. A flux density Q(z) (kg∙m⁻²) is determined at each BSNE by dividing the mass m(z) of sediments trapped in a BSNE by its opening area s(z) (Equation (1)).

Q(z) = m(z)/s(z)(1)

The flux density is highest near the ground and follows a power law with the height (Equation (2)) [24]-[28][31][32][50][51].

where a is the flux density at height z = 0, and b is a dimensionless parameter.

Integrating the flux density over the saltation height set at 0.4 m (Equation (3); [52]) gives the horizontal erosion flux (Fh) expressed in kg∙m⁻¹ per event.

Wind erosion directly affects the top few millimeters of the soil [28]. The soil analysis therefore focused on the surface layer of the field. Thus, five samples were taken from the top 5 centimeters and mixed to form one composite sample representative of the surface soil of the millet field. This composite sample was analyzed in the laboratory of the National Institute of Agronomic Research of Niger (INRAN).

The transport of nutrients in wind flux was determined using the masses trapped in the five BSNE traps at the top of the masts. These masses were grouped together to form a sample to be analyzed per collection of eroded sediments. However, the trapped masses were too small and insufficient to allow analysis for the period from August 2020 to March 2012. That is another constraint linked to the soil analysis techniques used at the INRAN laboratory. So between August and March, the nutrient transport is considered to be nil due to the very small amount of eroded sediments.

The analyses, which concerned both wind flux and soil samples, were carried out at the soil laboratory of the National Institute for Agricultural Research of Niger (INRAN). The parameters analyzed were: 1) carbon content determined by the Walkley and Black method [24][53][54]; 2) nitrogen content by the Kjeldahl method [24][53]; 3) phosphorus content using the Bray method [24][55]; 4) three granulometric fractions were determined using the Bouyoucos method [24][56]: mean to very coarse sands, very fine to fine sands, and the clay-silt fraction.

Flux measurements are shown in Figure 5. The shape of the curve is very similar to those obtained on traditional fields cultivated without organic fertilizer in the Sahel Millet is usually sown in pockets in June after the first rains [23]-[28][45][57]. The period between April and July is characterized by intense erosion and corresponds to the start of the monsoon and rainy season. This period of erosion is characterized by the occurrence of very strong wind velocities (Figure 2) linked to convective events in a context of minimal protection of the soil surface by crop residues (Figure 3). Between April and July, the fluxes recorded on the millet field, with an average of 11.94 kg∙m⁻¹∙d⁻¹, represented 93.28% of the annual fluxes. Only 2.39% of the annual flux was recorded during the Harmattan, representing an average flux of 0.15 kg∙m⁻¹∙d⁻¹, compared to 97.61% during the monsoon (Table 1). This confirms that wind transport is mainly active at the beginning of the monsoon on cultivated areas in the Sahel.

The annual flux was 1 468 kg∙m⁻¹, while in Kilakina (South-east Niger), which records 200 mm less rainfall than Bangou Koiré, the maximum annual flux was 2 900 kg∙m⁻¹ on a millet field [25], i.e. twice that of Bangou Koiré. In Banizoumbou, under similar rainfall conditions between 1996 and 1998, a maximum cumulative total of 601 kg∙m⁻¹ was measured, while [27] measured 370 kg∙m⁻¹ between 2006 and 2008, which are respectively 2 and 4 times lower than in Bangou Koiré. In fact, the cultivated soils at Banizoumbou had greater vegetation cover than at Bangou Koiré. With these annual horizontal fluxes, soil losses of 20 T∙ha⁻¹∙yr⁻¹ were recorded on the Banizoumbou field [27]. Soil losses on the millet field in Bougou Koiré, where flux can reach four times those of the Banizoumbou fields, could therefore be well above 20 T∙ha⁻¹∙yr⁻¹.

Figure 5. Intra-annual dynamics of horizontal flux measured at Bangou Koiré on the millet field from April 2020 to March 2021.

Table 1.Comparison of seasonal horizontal flux measured on the millet field between 2019 and 2021.

Figure 6 and Table 2 show the comparative distributions of particle size fractions of soil and wind-flux sediments generated by erosion on the same soil. This comparison shows that the content of fine particles (clay and silt) is higher in wind-flux sediments than in soil. In fact, the fine fraction (clay and silt) and fine sand contents are 6.16% and 50.63%, respectively, in the soil. In the flux, these contents are 12.46% and 76.16%, respectively, i.e., more than 2 and 1.5 times higher than the content available in the soil.

Wind erosion has therefore led to granulometric fractionation by transporting and most likely depleting the soil of fine particles. Granulometric fractionation, although less intense than that observed in Bangou Koiré due to wind erosion, has been observed in studies carried out in Niger [49], and China [50]. [58], for example, showed that on cultivated soils on the Ordos Plateau (China), wind erosion depletes the soil of erodible particles with diameters between 60 and 400 µm. Similarly, [59] used wind tunnel measurements to show that increased wind flux causes significant entrainment of particles and aggregates with diameters of less than 100 µm. Losses of fine particles could lead to nutrient losses and reduce the productivity of Sahelian soils. The nutrient content of Sahelian soils is known to be low and concentrated in the surface horizon, which is subject to wind erosion [23][31][32].

Contrary to our measurements at Bangou Koiré, [58] has highlighted an increase in the fine clay-silt fraction of soils due to wind erosion. It should be noted that the very high wind flux (Figure 5) at Bangou Koiré could indicate significant saltation, which could have released and consequently transported more fine particles.

The measurements have shown that fluxes contained less coarse sand than the available content in the soil (Figure 6). In fact, the coarse sand content in the soil was four times higher than in the fluxes. There was therefore less coarse sand in the fluxes measured. In fact, the weight of coarse particles is such that they are transported close to the soil surface, where crop residues no doubt trap them more effectively. Wind erosion of the millet field surface therefore tends to increase the sandy skeletal fraction in soils, making them more unreactive and poorer.

Figure 6. Comparison of particle size fractions determined in annual flux and in millet field soil.

Table 2.Distribution of particle size fractions between the flux and the soil of the millet field.

Figure 7 and Table 3 show the concentrations of fertilizers transported in wind flux and those available in millet field soils. It appears that the nitrogen, organic carbon, and phosphorus contents in the flux are, on average, 10 times higher in wind flux than in soils (Table 3). The nitrogen, organic carbon, and phosphorus contents in the soil are 0.634%, 0.1%, and 33.57 ppm, respectively. In the flux, these contents are 3.064%, 1.71%, and 260 ppm, respectively, which are more than 5, 17, and 8 times higher than the available content in the soil.

At the village level in the Sahelian environment, [56] have shown that nutrient loss through wind erosion is higher than that caused by water erosion. Nutrient losses from a field during an event of wind erosion can represent up to 73% of the nitrogen and 100% of the phosphorus needed for plant production [60]. [61] also showed that soil losses during just two erosion events resulted in losses of 18.3 and 6.1 kg·ha⁻¹ of nitrogen and phosphorus, respectively.

Figure 7. Distribution of fertilizer contents (N, C org, and P) in the annual flux and in the soil of the millet field.

Table 3. Fertilizer contents (N, C org, and P) and their comparative ratios in the annual flux and the millet field soil.

Nutrient transport showed significant seasonal variability (Figure 8; Table 4). There appears to be a period of intense transport between April and July, which corresponds to the beginning of the monsoon and rainy season. This period of high nutrient mobilization can be explained by the occurrence of very strong winds (Figure 2) and minimal protection of the soil surface by vegetation (Figure 3). During this period (April–July), all nutrient transport takes place on the millet field. However, the period from April to May was very characteristic. This period generated 49%, 82%, and 60% of the annual transport of phosphorus, nitrogen, and organic carbon, respectively, on the millet field, for only 21% of the horizontal flux mobilized on the field.

During the second part of the monsoon (August to September) and throughout the Harmattan, nutrient transport is practically non-existent on the millet field. Indeed, during this period, wind erosion on the field was low, with only 2.39% of the annual flux recorded.

Figure 8. Comparative trends in horizontal fluxes and associated nutrients: A. phosphorus, B. carbon, and nitrogen on the millet field during the wind erosion cycle 2020-2021.

Table 4.Seasonal variation of fertilizer contents (N, C org, and P) transported in the horizontal flux of the millet field.