The study was conducted in central Senegal, in the administrative region of Kaffrine located between latitudes 12˚06 N and longitudes 15˚33 W. It is bordered to the north by the regions of Diourbel, Louga and Matam, to the south by the Republic of Gambia, to the east by the region of Tambacounda and to the west by the region of Kaolack (Figure 1). The study was conducted specifically in two rural communities in the Kaffrine region, Lour Escale and Ida Mouride. This is an ecological transition zone between the pastoral Sahelian north and the Sudanese south, which hosts or through which many ruminants transit during the dry season, and where the advancing agricultural front is putting strong pressure on available woody resources.

The dry tropical climate is of the Sudano-Sahelian type. The average monthly minimum and maximum temperatures are 16.2˚C (January) and 42.1˚C (May) respectively. The average annual temperature is around 29.6˚C. Rainfall is irregular: the average rainfall (between 1965 and 2008) is 704.32 mm at the reference station (Koungheul). Koungheul has been characterized by a persistent rainfall deficit since 1970, with a large proportion of biologically dry years, leading to a total amount of rain collected that is less than the annual average corresponding to 23 years of deficit. The rainy season occurs from May to October, which classically distinguishes two periods in the year: a dry period (P < 2T) of 7 to 8 months (from October to May) and a rainy season (P > 2T) of 4 to 5 months. The months of July, August and September, which account for 79.31% of rainfall, constitute the biologically wet period, with August being the wettest month (32%) [3] [4] .

The Kaffrine region belongs to the sedimentary formations of the continental terminal [5] . The soils are sandy, sandy-clay or clayey-sandy and support vegetation characterized from north to south by grassy savannah, shrubby savannah,

tree savannah and wooded savannah. Agriculture, mainly rain-fed, and extensive livestock farming are the main activities of the region.

Five villages have been selected for each rural community In Lour Escale, we have: Lour Village (LE), Koura Mouride (KM), Sobel Diam-Diam (SDD), Darou Dame Leye (DDL) and Yetty Khaye (YK). For the Ida Mouride site, the villages selected are Ida Village (IV), Fass Diabel (FD), Ndawène Saragnama (NS), Maka Katal (MK) and Arafat Mbayène (AM). These villages were selected on the basis of the size of their populations and their willingness to host transhumant herders.

For each village, 20 plots of 50 m × 50 m (2500 m²) in the rangelands were used to study vegetation parameters. In each plot, the flora inventory and tree measurements were carried out. Height and trunk circumference at 0.3 m from the ground to establish stand structure, crown diameter to assess crown cover, and the distance between the 2 closest trees were measured for each tree to calculate actual and theoretical densities for each site.

Species naming was based on the flora of Senegal [6] , the book of “vernacular names of plants” [7] and the catalog of vascular plants of Senegal. Synonyms have been updated on the basis of the enumeration of African flowering plants [8] .

All the data was entered and processed in an Excel spreadsheet. A floristic list was established and the specific richness of the flora, the density and some parameters of the woody stand were evaluated.

For the evaluation of the diversity parameters, we used the diversity index (H') of Shannon Weaver (1949) and the evenness (E).

The Shannon index is the most commonly used. It expresses the relative importance of the number of abundant species in a given environment. Its value gives an estimate of the uncertainty with which one can correctly predict the species to which the next individual collected belongs. The index is minimum when all individuals belong to the same species. It is maximum when each individual represents a distinct species [9] . It is given by the following relationship:

H' = −Σpilog2 pi

In terms of evenness, it provides information on the distribution of species abundances in the stand. According to [10] , the regularity index appears to be a more rigorous comparison term. It is between 0 and 1. It tends towards 0 when all the individuals correspond to a single species. It tends towards 1 when each species is represented by the same number of individuals [11] . Its value is obtained by using the following formula:

E = H'/H'_max

With H' = Shannon index; H'_max = log2S, S being the total species richness.

For woody stand structure, we used Sturge’s rule to define the number of classes and the interval between each class for a sample size of n.

Number of classes = 1 + (3.3 logn)

Class interval = (X_max − X_min)/Number of classes

With X_max and X_min, respectively the largest and smallest value of X in the statistical series.

From X_min we obtain the class limits or class boundaries by successive addition of the class interval.

- Floristic composition:

The flora of the two rural communities (RCs) studied has a total of 70 species, 51 of which were recorded in the RC of Lour Escale and 45 in Ida Mouride. They are unevenly distributed in 54 genera and 27 taxonomic families (Table 1).

In total, 26 species are common to both RCs, compared to 40 species present in one and the other RC.

RC: Rural community; IM: Ida Mouride; LE: Lour Stopover; 1: present; 0: missing.

On the generic level, the best represented families are the Ceasalpiniaceae and Rubiaceae (5 genera each). They are followed by Combretaceae and Euphorbiacea (4 genera each) and by Capparaceae, Fabaceae and Mimosaceae (3 genera for each family). 6 families are represented by two species and 10 families by only one species.

At the specific level, only the genus Combretum is represented by 4 species. 4 other genera are represented by 3 species. These are Acacia, Ficus, Gardenia and Terminalia. Considering the families, the Combretaceae family alone has 10 species, followed by the Rubiaceae with 7 species and the Ceasalpiniaceae and Mimosaceae (5 species each).

- Distribution and specific importance according to the selected sites:

The presence or absence of a species in a village is coded (Table 2) by the values (1 or presence and 0 or absence).

Examination of this table allowed us to classify the species into 3 groups:

The first group is made up of species present in six (6) to ten (10) sites; these are highly common species, and they could constitute the very common or true indifferent species, these species are very frequent and widely distributed throughout the sites; they are 22 in number.

The second group is made up of common but infrequent species present in three (2) to six (5) sites;

And the third group is made up of differential species that are present in only one (1) site. These species reveal specificities of the environment.

Specific diversity is assessed by calculating the specific richness per village and for each RC. The total and average species richness (TSR) for each site are determined and their values are given in Table 3.

We have presented some parameters of woody vegetation for each RC: these are density, species diversity, cover and regeneration of woody species (Table 4 and Table 5).

The level of organization

The Shannon index was 3 bits at the Lour Escale RC route set (Table 4). It ranged from 2.52 (at KM) to 3.13 (at YK). The SDD village showed the same index as that observed for all the routes in the RC.

However, regularity was lower at the RC level (0.52), while it was higher at SDD (0.61) and YK (0.59). It had the same value (0.53) for the villages of LE and DDL; this would suggest a similar level of organization for these two sites.

In the Ida Mouride RC, as in LE, there is also a difference between the Shannon diversity index values for the three sites. The values obtained show that diversity is greater at the NS (2.3) and IM village (1.68) sites (Table 5). It is in these two sites where the value of the Shannon index exceeds that of the RC (1.52). The regularity indices are also higher (0.65 and 0.62) for NS and IM village respectively.

- Woody cover

For the site, the basal area is 4.09 m²/ha. It is much higher at SDD (8.88 m²/ha), ie, double that obtained for all the rangelands of the RC, and lower at Koura Mouride (1.76 m²/ha).

The highest basal area values are found in the IM RC sites. They vary from 5.97 m²/ha in IM village to 14.92 m²/ha in the NS site.

The aerial cover is 20.73% when considering the whole of the Lour Escale CR. On the other hand, from one station to another, canopy cover values varied from 8.95% at DDL to 27.97% at YK, passing through values of 17.54% at KM and 17.94% at LE respectively. For the Ida Mouride RC, the overall coverage is 28%. It is slightly higher than that observed in Lour Escale. The villages of IM and LE had the highest coverage values.

- Density

The observed and theoretical densities show a significant difference in all two RCs studied. Theoretical to observed density ratios per site are very high. At the overall level of the Lour Escale RC, this ratio is 3, whereas it is higher in the Ida Mouride RC (7).

At the rural community level, this ratio is 2.72.

The average distances between feet used to calculate actual density by site vary from one RC to another and from one site to another.

- Regeneration

The number of seedlings counted in all the inventory plots is 15,797 or 82.93% of the woody stand of the Lour Escale RC.

When considering the sites, this rate varied greatly. Thus, it is 7.54, 11.51, 14.03, 16.26 and 33.56% of the stand for the DDL, KM, SDD, YK and LE village sites respectively.

The Ida RC, on the other hand, has a higher regeneration rate (90.48%). This rate varies from one RC site to another. The highest rate is found in FD (28.79%), while the lowest rate (5.24%) is obtained in NS.

The structure of the woody stand is characterized by the distribution of individuals in trunk diameter and tree height classes.

- Size distribution

When we consider the distribution of woody species according to their size in the Lour Escale RC (Figure 2), we realize that even if all classes are represented, almost all the individuals are grouped in the first two classes. The class 3 - 12 cm alone represents more than 78% of the individuals. Species such as Combretum glutinosum and Guiera senegalensis contributed the most to this proportion. They represent respectively 45% and 24% of the total contribution. The third dominant species, Cordyla pinnata, is absent in the first class. It contributes more to the higher classes.

In the Ida Mouride RC, the distribution is similar to that of Lour Escale, marked by a very high concentration of individuals in the first class (3 - 15) of diameter. This class represents nearly 80% of the woody stand and species such as Combretum glutinosum and Piliostigma reticulatum largely dominate it (45% and 25% respectively) (Figure 3).

- Height distribution

The height class distribution of woody plants shows a unimodal structure for both RCs studied (Figure 4 and Figure 5). All classes are represented even if the lower strata are dominant.

In the Lour RC (Figure 4), the 1 - 3 m stratum is by far the most important. There is a strong contribution of Combretum glutinosum and Guiera senegalensis in this one.

In the Ida Mouride RC (Figure 5), it is the second stratum (3 - 6 m) that contains the greatest number of individuals, even though the first stratum (1 - 3 m) has a significant presence.

As for Cordyla pinnata and Sterculia setigera, we note that they appear in the upper strata, unlike species such as Combretum glutinosum, Guiera senegalensis and Piliostigma reticulatum.