The study was carried out in Lower Casamance, more precisely in the commune of Coubalan, located south of the Tenghory arrondissement, in the Bignona Department of the Ziguinchor region (Figure 1). The climate here is coastal South Sudanian, with two seasons: dry and rainy [9] . Average monthly temperatures range from 21.5˚C to 35˚C between 1990 and 2016 [9] . Average annual rainfall over the 1980-2018 series is 1302.04 mm [10] . The Tenghory district is characterized by a predominance of tropical ferruginous and ferralitic soils [11] cited by [10] and [12] . Lower Casamance boasts the densest forests in Senegal, with trees reaching heights of 15 to 20 meters. There are mangrove areas as well as dense dry forests composed mainly of evergreen species [13] [14] cited by [12] and [15] .

Random stratified sampling was carried out for the different land-use types in all 13 villages of the commune of Coubalan. The sampling unit for fields and fallows is a square plot of 50 m × 50 m, i.e. a survey area of 2500 m² as recommended by [16] for the study of woody vegetation in agroforestry systems. As for forests, plots measuring 30 m × 30 m were set up, i.e. an area of 900 m² [17] [18] [19] . The plots were, therefore, installed randomly according to land use as follows: 20 plots at field level, 8 plots at fallow level and 12 plots at forest level.

In each survey, all species were inventoried and the dendrometric parameters of each individual were measured. Total height was measured using a suntoo dendrometer, trunk diameter at breast height using a forestry compass, and cross diameter of the crown (East-West and North-South direction) using a metric tape.

Young trees were counted to assess regeneration. Woody species with a diameter of less than 5 cm at 1.30 m are considered to be regenerating [20] . Individuals of the Arecaceae family with stipes less than 1.30 m high are considered part of the regeneration [21] cited by [22] .

The data obtained was entered and the tables and graphs were produced using Excel. XLSTAT Version 2014 software was used to perform the Analysis of Variance (ANOVA), followed by the Newman-Keuls test of comparison of means at the 5% significance level to determine the different groups. A Principal Component Analysis (PCA) was performed to establish the discriminating characteristics of the different land-use patterns.

Various formulas were used to calculate vegetation parameters, including:

Specific frequency

The frequency of occurrence provides information on the distribution of a species in a stand.

It can be expressed as an absolute value or as a percentage. In percentage terms, it is calculated using the following formula:

F = N r i / N r × 100

F = frequency of presence expressed as a percentage (%); Nri = number of surveys where species i is present; Nr = total number of surveys.

Species richness

This is assessed on the basis of total species richness, which is the total number of species present in the stand considered in a given ecosystem [23] .

Density

This is the number of individuals per unit area. It is expressed in number of individuals/ha. Observed or actual density is obtained by dividing the total number of individuals in the sample (N) by the surface area sampled (S).

D = N / S

D = density (ind/ha); N = total number of individuals in the stand; S = sample area sample area (ha).

Basal area

This is the surface area occupied by the tree at the base of its trunk. It is expressed in square meters per hectare, and is calculated as follows:

S t = ( ∑ π ( d 1.30 / 2 ) 2 ) / S

St = basal area; d1.30 = trunk diameter in m at 1.30 m from the ground; S = sample area (ha).

Woody cover

This is the surface area of the tree crown projected vertically onto the ground. It is expressed in square meters per hectare. Woody cover is calculated using the following formula:

C = ( ∑ π ( d m h / 2 ) 2 ) / S

C = woody cover; dmh = average crown diameter in m; S = sample area (ha).

Shannon weaver index (H')

Consider both abundance and species richness, and is used to assess the distribution of individuals according to species. It ranges from 0 to 4.5. The index is minimum when all individuals belong to the same species. It is maximum when each individual represents a distinct species [24] . It can be expressed in bits and its formula is:

H ′ = − ∑ p i . log 2 p i

Pi = relative abundance of each species.

Evenness index (E)

Provide information on the distribution of species abundance in the stand. According to [25] , the regularity index is a more rigorous term of comparison. It ranges from 0 to 1. It tends towards 0 when all individuals correspond to a single species. It tends towards 1 when each species is represented by the same number of individuals [23] cited by [26] . This index is given by the following formula:

E = H ′ / H max

H': Shannon Weaver Index; H_max = log2(S).

Importance Value Index (IVI)

This is a synthetic, quantified expression of the importance of a species in a stand. To assess specific preponderance in tropical forests, this index is often used [27] cited by [28] . For better interpretation, the IVI has been expressed as a percentage (%) and is defined as the arithmetic mean, for species i, of relative density (Dr), relative frequency (Fr) and relative dominance (Domr).

IVI = ( D r + F r + D o m r ) / 3

Dr = (Ni/N) × 100 (Dr = relative density (%); Ni = sample size of species i; N = total sample size).

Fr = (Fi/F) × 100 (Fr = relative frequency (%); Fi = frequency of presence of species i (%); F sum of frequencies of all species in the sample).

Domr = (Sti/St) × 100 (Domr = relative dominance (%); Sti = basal area occupied by species i; St = total basal area of species in sample).

Stand regeneration rate

IL is given by the percentage ratio between the total number of seedlings and the total number of trees in the stand.

SRR = (Total number of seedlings)/(Total number of trees in stand) × 100.

The woody flora surveyed in the commune of Coubalan enabled us to identify 53 species, divided into 48 genera belonging to 22 botanical families.

In the fields, 37 species in 34 genera and 17 families were identified; in the fallow land, 28 species in 27 genera and 15 families were identified; and in the forests, 43 species in 41 genera and 19 botanical families were identified (Table 1).

The species with the highest relative abundances in the fields of the Coubalan commune are Icacina oliviformis (53.77%), Guiera senegalensis (5.43%), Terminalia macroptera (2.99%), Acacia ataxacantha (1.96%) and Cassia sieberiana (1.46%).

In fallow land, the most abundant species is Guiera senegalensis with a relative abundance of 79.58%. It is followed by Cassia sieberiana (5.01%), Azadiracta indica (4.55%), Combretum micranthum (4.36%) and Terminalia macroptera (1.65%).

In forests, the most abundant species are Guiera senegalensis (21.12%), Combretum micranthum (20.57%), Terminalia macroptera (18.18%), Combretum nioroense (14.96%) and Cassia sieberiana (3.88%) (Figure 2).

With regard to structural parameters, analysis of variance revealed a significant difference in observed density (P < 0.0001), cover ratio (P < 0.0001) and basal area (P = 0.003) between the different land-use types.

In fact, forest density (1321.3 ± 635.8 ind/ha) is statistically higher than that of the other land uses, which show no significant difference between them. Forest density is almost four times that of fallow land (335.5 ± 297.4 ind/ha) and more than thirty-four times that of fields (38.6 ± 26.7 ind/ha) (Table 2).

As with density, forest cover was statistically higher (79.25 ± 21.08%) than other land uses, which showed no significant difference between them. Woody vegetation cover in the forest is thus 7 times higher than in the fields (11.373%) and more than 5 times higher than in the fallow land (14.731%).

As for basal area, as for density and cover rate, it is significantly greater in forests (13.09 ± 0.1 m²/ha) compared to fallow (1.18 ± 0.013 m²/ha) and fields (2.32 ± 0.024 m²/ha) (Table 2).

Legend: ind/ha: individual per hectare. In the same column, values accompanied by the same letters are not significantly different.

In terms of dendrometric parameters, statistical analysis revealed a highly significant difference in woody diameter and height (P < 0.0001) between the different land-use types (Table 2).

Average trunk diameter and average woody height are highest in fields, with values of 20.4 cm and 7.4 m respectively, followed by forests, with average trunk diameter of 5.4 cm and average height of 4.2 m. The lowest values for dendrometric parameters were recorded in fallow land, with 3.4 cm and 3.05 m respectively for average trunk diameter and average woody height (Table 2).

Analysis of Table 3 shows that the Shannon Weaver index (H') is highest in forests (3.34 bits), followed by fallow land (1.73 bits) and fields (1.37). As for the regularity index (E), it is higher in forests (0.62) and lower in fallow land (0.28).

It thus appears that specific diversity is higher in forests and lower in fallow land.

The regeneration rate of the woody stand in the Coubalan commune is 96.11%. This rate varies from one type of land use to another. Fallow land has the highest regeneration rate at 98.49%, followed by fields at 98.32% and forests at 87.51% (Figure 3).

The species with the highest Importance Value Indices (IVIs) in the different land-use types are shown in Table 4.

In the fields, the 5 species with the highest Importance Value Indices are Icacina oliviformis (18.74%), Guiera senegalensis (15.89%), Faidherbia albida (12.71%), Anacardium occidentale (7.02) and Parkia biglobosa (6.63%).

In fallow land, the 5 most ecologically important species are Guiera senegalensis (31.83%), Faidherbia albida (16.39%), Azadirachta indica (9.90%), Cassia sieberiana (5.29%) and Parkia biglobosa (4.24%).

Legend: H': Shannon index; E: Regularity index.

Legend: Dr: Relative density; Fr: Relative frequency; Domr: Relative dominance.

The 5 most ecologically important forest species are Combretum micranthum (13.15%), Guiera senegalensis (9.52%), Terminalia macroptera (9.24%), Khaya senegalensis (8.04%) and Cola lateritia (6.35%) (Table 4).

Horizontal structure

Analysis of Figure 4 shows that in the fields of the Coubalan commune, the woody stand is characterized by a predominance of individuals in the [5 - 15] diameter class. This class alone accounts for 40.12% of the stand. It is followed by the [15 - 25] and [35 - 45] classes, comprising 26.54% and 12.35% of the stand respectively.

In terms of fallow land, the [5 - 15] diameter class is still the best represented, accounting for 72.13% of the stand. This is followed by the [15 - 25] and [25 - 35] classes, which account for 14.75% and 6.56% respectively.

As for the forest, like the two previous land-use types, its stand is dominated by individuals belonging to the [5 - 15] diameter class, with 80.9% of the woody stand. This class is followed by [15 - 25], [25 - 35] and [35 - 45] with 8.8% and 5.9% of the stand respectively.

The woody stand structure in forests and fallow land is L-shaped, while that in fields is exponential and decreasing. These structures are all characteristic of a young, balanced stand, reflecting good regeneration of the stand, with recruitment of young individuals towards the intermediate classes being greater in the fields.

Vertical structure

In the woody stand of the different land-use types in the Coubalan commune, the [2 - 4] height class is the most represented, with 27.0%, 89.61% and 58.28% of individuals in fields, fallows and forests respectively (Figure 5). This class is followed by the 4-6m height class, with 17.71%, 7.23% and 27.58% of individuals in fields, fallows and forests respectively (Figure 5).

As with the horizontal structure, the vertical structure is characteristic of a young, balanced stand.

Figure 6, obtained from the Principal Component Analysis (PCA) performed on the basis of the matrix of three land-use types and six woody vegetation parameters, shows the discriminating characteristics of each land-use type. We thus distinguish:

- Fields, characterized by vegetation consisting essentially of large-sized (H) and large-diameter (D) woody plants.

- Forests, characterized by high vegetation structural parameters (basal area, cover ratio and density) and high species richness.

- And fallow land, characterized by low structural parameters and low species richness.

The figure also shows that stand density is strongly correlated with woody vegetation cover rate, and woody height is strongly correlated with trunk diameter.