Study area
The study was carried out at four sites in classified forest, field and fallow along a phytogeographic level in Burkina Faso. The sites were Wayen (11°55'08"N and 3°43'33"W), Tiogo (12°10'36"N and 2°41'29"W), Bondoukuy (11°55'08"N and 3°43'33"W) and Boromo (11°44'47"N and 2°55'54"W) (Fig. 1). The vegetation in the classified forests is characterized by various savanna types including shrub savannas, tree savannas, woodlands, dry forests, and gallery forests (Traore et al. 2020). Fallow and field were dominated by agroforestry parks. Wayen and Tiogo sites are located in the North Sudanian sector with annual rainfall varies from 600 to 700 mm, and 40–70 rainy days per year. Soils are mostly lithosols. Bondoukuy and Boromo are located in the South Sudanian sector. Annual rainfall ranges from 800 to 900 mm, with 70 to 90 rainy days per year. Soils are mainly ferralsols (Bognounou et al. 2009). Subsistence agriculture, livestock and exploitation of NTFPs are the main livelihood activities in all sites. A wide variety of annual crops are cultivated: cotton, maize, sorghum, groundnut, cowpea and millet. These are often associated with multipurpose species trees like Vitellaria paradoxa, Parkia biglobosa, and Sclerocarya birrea (Cissé et al. 2019).
Description of study species
Sclerocarya birrea, a dioecious tree in the Anacardiaceae family, is found in West Africa, northern Cameroon, Sudan, Tanzania, and northern Kenya (Hall et al. 2002). Female trees bear yellow drupes, 3–4 cm in diameter and 15–25 g in weight (Shackleton et al. 2003). The leaves are borne in clusters at the apex of stout branchlets, and are alternate and compound, 8–38 cm long imparipinnate, bearing 3–18 pairs of opposite or subopposite leaflets (Arbonnier 2009). The flowers are small, shortly pediculate, whitish purple to red. The fruits are obovoid fleshy and juicy drupes, 23.5 cm in diameter, green becoming yellow at maturity with a weight of 18–35 g (Diallo et al. 2006). The species thrives with 500 to 1600 mm of annual rainfall over different soil types (Gouwakinnou et al. 2009). It occurs across a range of vegetation types, principally mixed deciduous woodland, wooded grassland and through the open dry savannas of northern Tropical Africa and the Sahelian region (Nyoka et al. 2015). In Burkina Faso, it occurs in Sudanian zone, sometimes forming monospecific stands (Tingueri et al. 2021).
Sampling design and data collection
Three land use types were considered regarding the different anthropogenic pressures: field, fallow and classified forest (Fig. 2). In this study, classified forests are legally designated areas of natural savanna or forest established by public authorities to limit human disturbances and protect resources, ecosystem functions, and services (Fig. 2c). The classified forests of Tiogo and Wayen were sampled in the North Sudanian sector, while the classified forests of Bale and Tuy were sampled in the South Sudanian sector. However, except the classified forest of Tiogo, which operates under participatory management with local communities, the other forests were fully controlled by the States services (Kagambega et al. 2019). All these classified forests are designated by IUCN Category IV of protected areas (IUCN 2012). Despite their status, illegal logging and NTFPs harvesting occur in these forests (Traore et al. 2020). Fields are currently active croplands under intensive cultivation (Fig. 2a). Fallows refers to previously cultivated lands left uncultivated for over five years to restore vegetation and soil fertility (Fig. 2b).
Data were collected during the rainy season (July to October 2023). In each site, the fields were visited with the owner's permission to identify accessible populations of S. birrea for measurements and understand their management. A total of 150 plots were selected using a stratified random sampling method across the three land-use types. In fields and fallows, 120 square plots (50 × 50 m) were established, while 30 rectangular plots (50 × 20 m) were set up in classified forest. Plot sizes followed the savannas vegetation surveys guidelines (Thiombiano et al. 2016). Sampling was based on the presence at least four trees of S. birrea individuals, with a minimum distance of 100 m maintained between plots within each land use type (Aleza et al. 2015). In each plot, all woody species were recorded to assess diversity and composition. Species that could not be identified in situ, were sampled for later determination using the catalogue of vascular plants of Burkina Faso (Thiombiano et al. 2012), and names were checked with World Plant Names Index (https://wfoplantlist.org/plant-list). Tree height and diameter at breast height (DBH ≥ 5 cm) were measured for all woody species.
To assess S. birrea regeneration, five subplots of 5 x 5 m were placed at each plot’s four corners and center. In each sub-plot, seedlings, suckers and coppices was recorded and classified following previous studies (Aleza et al. 2015; Kabré et al. 2020; Ouédraogo et al. 2025). Seedlings grow from germination of seed (Fig. 3a), suckers from lateral roots (Fig. 3b) and coppices are juvenile plant from cut or burned tree stumps (Fig. 3c).
Statistical analysis
Influence of land use on the diversity and composition of woody species associated with S. birrea stands
The composition, abundance, and taxonomic diversity of woody species associated with S. birrea were analyzed across land use types. Stand diversity was assessed using Hill’s framework (Hill 1973) with the BiodiversityR package (Kindt 2022), by calculating four indices:
N0 = S; with S the number of species in a plot;
N1 = eH’; with H’ Shannon’s index;
N2 = 1/D, D = is the Simpson’s diversity index;
Evenness (E) = H′/ln(S).
One-way analysis of variance (ANOVA) was used to assess the effect of land use on diversity indices (N0, N1, N2, and Evenness). When significant, Tukey’s HSD test identified pairwise differences between land use types. Species composition across land-use types was assessed using ANOSIM and visualized with NMDS based on species abundance and dominance data. The ecological importance of woody species co-occurring with S. birrea was evaluated using the Importance Value Index (IVI), which ranges from 0 to 300 (Mueller-Dombois and Ellenberg 1974). The IVI was calculated by summing three components:
Relative Frequency = frequency of a species/sum of all frequencies × 100;
Relative density = number of individuals of a species / total number of individuals × 100
Relative Dominance = total basal area of a species/basal area of all species × 100
Importance Value Index = Relative density + Relative dominance + Relative frequency.
For each land use type, the 20 species with the highest IVI were selected to illustrate patterns in dominant woody species composition.
Effect of land use on the density of adult, seedling, suckers and coppices of S. birrea
Structural parameters of S. birrea included basal area (G), density (N), mean diameter (Dg), and Lorey’s mean height (HL) were assessed. G\(\:=\frac{{\pi\:}}{40000\:\text{S}}\sum\:_{\text{i}=1}^{\text{n}}\text{d}\text{i}²;\) \(\:di\:\)the DBH of i-th tree (Philip, 2002). N= n/S, with n the average number of individuals per plot and S as the area expressed in hectares. HL expressed: \(\:\text{H}\text{L}=\frac{\sum\:_{i=1}^{k}gihi}{{\sum\:}_{i=1}^{k}gi}\) ; expresses the height of the individuals adjusted by the basal area. The natural regeneration was calculated \(\:Nr=\frac{\sum\:_{l=1}^{k}{N}_{l}{\stackrel{-}{N}}_{rl}}{n}\); with \(\:{\stackrel{-}{N}}_{rl}=\left(\frac{1}{{n}_{l}}\right)\sum\:_{i=1}^{{n}_{l}}{y}_{li}\); \(\:Nr\) is the mean density of S. birrea regeneration within land use, n the total number of sampling,\(\:\:{N}_{l}\) is the mean density of adults S. birrea, \(\:{y}_{li}\), the regeneration density withon ith plot of the stand (Bonou et al. 2009). ANOVA was used to compare structural parameters among land use types.
Size class distribution of S. birrea across land use
Diameter and height class distributions based on Condit et al. (1998) were used to interpret the status of S. birrea populations according to land use types. DBH data were computed and assembled in ten diameter classes of regular interval whereas five height classes were established for regeneration stratum. To assess population structure, the 3-parameter Weibull theoretical distribution model was applied due to its flexibility and simplicity (Glèlè Kakaï et al. 2016). The density function f(x) was define below:\(\:\:f\left(x\right)=\frac{c}{b}{\left(\frac{x-a}{b}\right)}^{c-1}\text{exp}\left[-{\left(\frac{x-a}{b}\right)}^{c}\right];\) where, a is the location parameter, b is the scale parameter, c is the shape parameter, x is the diameter. The Weibull distribution can take several forms depending on the value of the shape parameter. Value of c < 1 generally indicate populations with high regeneration potential (reverse J-shaped distribution), whereas values of c near or greater than 1 (flatter, unimodal, or left skewed distributions) indicate populations with a lower regeneration potential. Many factors, however, affect the c value and it, must be interpreted with caution, but generally the higher the c, the higher the tendency for population decline and vulnerability to extirpation (Lykke et al. 2025). A log-linear analysis was performed to test the adequacy of the observed structure to the Weibull distribution. The hypothesis of adequacy between both distributions is accepted if the probability value of the test is higher than 0.05. All statistical analyses were performed in R (R Core Team, 2022).