Local perception of ecosystem services and their conservation in Sudanian savannas of Burkina Faso (West Africa)

Context In Burkina Faso, Sudanian savannas are important ecosystems for conservation of plant diversity. Due to desertification and insecurity, population migration from the North has increased human density and anthropogenic pressure on southern savannas. This study aims to investigate knowledge of local populations on ecosystem services (ES) and perception of their conservation. Method Individual semi-structured interviews about knowledge on ES and ecosystem conservation issues were conducted. Informants were selected according to sociocultural groups and sex in three areas of different land use intensity: the communal area of Dano (CAD), the Total Wildlife Reserve of Bontioli (TWRB) and the Game Ranch of Nazinga (GRN). The use value and vulnerability index of each plant species were determined. A cluster analysis and a principal component analysis were carried out to identify the particular knowledge of different ethnic groups. Results Overall, 163 plant species were cited for fifteen ES. Provisioning services were most frequently cited (100%), regulating services second most frequently (92.47%). Entire plants were exclusively used for ES with non-material benefits (protection against wind, for shading, soil fertility, erosion prevention, tourism and religion). The ten species contributing most to ES provision were Vitellaria paradoxa, Parkia biglobosa, Diospyros mespiliformis, Adansonia digitata, Lannea microcarpa, Faidherbia albida, Khaya senegalensis, Afzelia africana, Ficus sycomorus, Pterocarpus erinaceus. Seven of them were identified as highly vulnerable. Around GRN, migrants and natives shared the same knowledge, while migrants in TWRB used the ES only to a small extent due to restricted contact with the native population. Migrants and natives of GRN had more knowledge on tourism and crafts services while the natives of CAD and TWRB made use of the services that sustain the quality of the agricultural land and meet their primary needs. To reduce further degradation, different communities suggested unanimously raising awareness of the importance of biodiversity and ecosystem conservation. The most quoted motivations to preserve ecosystems were vegetation sustainability and village development. Conclusion This study documented important local knowledge-based information to guide cultivation of local multipurpose species and initiation of communities to practice best management strategies for sustainable conservation of biodiversity.

can be classified in four main categories. The supporting services derive from general functioning of an ecosystem, the regulating services correspond to the direct services of ecological functions on site, and provisioning and cultural services refer to direct services of obtaining goods and social and spiritual well-being from ecosystems. While the provisioning services provide finished products of ecosystems, the non-material cultural services allow for developing and enriching knowledge systems, social relationships and aesthetic values [2]. Products of ecosystems used for provisioning services (such as food, fodder, wood, medicinal compounds) include non-timber forest products (NTFP) such as fruits, leaves, seeds, flowers, bark, medicinal herbs, as well as wood cut from trees for supply of energy and for construction. In West African semi-arid areas, local populations strongly depend on plant resources for meeting their daily needs [3][4][5][6]. Local people consider savanna ecosystems as their own good, as granary, pharmacy, pasture, place of religious worship and source of the strength of their territory [7]. Ecosystem functions and services do not only result from good ecosystem health, but also from the use that populations have made in various biogeographic and geo-economic contexts [8].
In Sudanian savanna ecosystems, climate and soil characteristics are favorable to the development of a diversified and dense vegetation cover [9]. In the semi-arid context of Burkina Faso, Sudanian savannas constitute a particular hot spot of plant species diversity [10,11]. Here, Zizka et al. [10] recorded 71% of all plant species of the country, with more than half of them being rare. Recent migration of human populations from the North to the South, fleeing desertification and climate change consequences on arable lands, has caused population density to increase in Sudanian savannas. In addition, terrorist attacks in the northern and eastern regions of the country have caused a new wave of migration and a raise in population density from 27.3 to 51.66 inhabitants/km 2 in the Southwest region between 1985 and 2019 [12]. This population growth has immediate consequences such as intensification of anthropogenic pressure on plant resources and protected areas [13]. Dimobe et al. [14 and 15] noted a significant decrease in natural vegetation from 1984 to 2013, followed by an expansion of croplands and habitat fragmentation in protected areas in the South Sudanian zone of the country [15].
Inclusive sustainable management of plant resources by the forest authority in collaboration with local populations could be a response to the strong anthropogenic pressure that ecosystems are facing, as the fundamental ecological role of forests is as important as their economic and social roles to local populations [3]. Indigenous people play a crucial role in biodiversity conservation through their traditional knowledge on species and habitats and their socioeconomic and symbolic practices as well [16]. However, many local populations suffer from consequences of large development projects and exploitation of natural resources. Among these consequences are land expropriation, loss of identity, language and culture [16]. Thus, local populations adapt their use to the change they perceive [8] in different ecosystems. They know that their survival essentially depends on their adaptation to the socio-environmental impact of climate change [17]. Better conservation of ecosystems requires good mastering of endogenous and scientific knowledge [7]. In addition, together with processes of decentralization and population self-management, the local scale is the relevant geographic and socio-economic space for conducting participatory development policies [18]. According to Holou and Sinsin [19], it would allow for answering one of the major concerns of African countries which is the rational and sustainable management of natural resources.
In the last decade, research questions on ES have been increasingly focused on provisioning services [4,5,[20][21][22][23][24][25]. In order to design effective vegetation management that prevents damage and promotes well-being of local people, it is important to take into account the needs of the latter. Thus, the present study aims to (i) understand knowledge of local populations on ecosystem services provided by plant species, (ii) identify factors that influence this knowledge, and (iii) understand local perceptions of sustainable management of plant communities.

Study area
The study was carried out in the South Sudanian phytogeographical sector of Burkina Faso [9]. The sampling areas were chosen according to variations in land-use intensity and identified through classification of landuse/land cover data from multi-temporal Landsat images (for methodological details see [14,15]), allowing for assigning them to a three-point scale with high, medium and low land-use intensity [26]. The communal area of Dano (CAD), located in the southwestern region of Burkina Faso and characterized by agrosystems and grazing land, corresponds to the high land-use intensity level (Fig. 1). The Total Wildlife Reserve of Bontioli (TWRB), also located in the southwestern region, consists of a protected area of IUCN category I under several human pressures, and corresponds to the medium land-use intensity level. The Game Ranch of Nazinga (GRN) and its ZOVICs (village hunting zones), which are protected areas of IUCN category VI, devoted to hunting and located in the South-central region, corresponds to the low land-use intensity level.
The native sociocultural groups are the Dagara and the Pougouli in CAD, the Dagara in TWRB and the Kassena in GRN. The dominant non-native sociocultural group encountered in all sites are the Mossi which have moved to these areas to practice agriculture on more fertile soils and recently for gold panning. Agriculture, livestock and gold panning constitute the main human activities in the study area.

Sampling design and data collection Ethnobotanical data
Seventeen villages were randomly selected in the three study sites following a random sampling scheme: seven villages in CAD, five around TWRB and five around GRN. The selection of informants was stratified, based on sociocultural groups (Dagara, Pougouli, Mossi, Kassena) and sex [28]. The Mossi were immigrants and the other ones indigenous (Table 1). Individual semi-structured interviews were conducted from December 2016 to March 2017 to collect the data following a questionnaire. In each village and for each sociocultural group, ten informants (five men and five women were selected randomly) of at least 20 years of age were interviewed with their consent, yielding a total sample of 240 informants (Table 1). Interviews were conducted in the local language of the informants and translated by a local translator. Informants were asked to list the plant species they use and the ES rendered, and the used organs were recorded. Informants were also questioned about availability and dynamics of the plant species, reasons for these dynamics, solutions in case of regression and motivations to conserve biodiversity. Each informant classed the items of suggested solution and motivation by preference order. For identifying the plant species cited in local language, in each study site, the "walk-inthe woods" method was used at the end of the interviews. This method consisted in field visits with members of the community who have good knowledge about plant species [29]. They were selected for having cited the largest number of utilized species during the interview phase. During the field visit, fresh samples of cited species were collected and pressed for identification using the floras of Berhaut [30], Lebourgeois and Merlier [31], Poilecot [32] and the field handbook of Arbonnier [33]. The identified samples were verified by comparison with samples of the Ouagadougou herbarium at Joseph Ki-Zerbo University.

Floristic data
A vulnerability index was calculated taking into account species frequencies in the study area [20,34,35]. For this purpose, vegetation surveys were conducted in 152 plots distributed in different vegetation types (grass savannas, shrub savannas, tree savannas, savanna woodlands, woodland and gallery forest) and the fields throughout the study area. The sampling unit was a 50 m × 20 m (1000 m2) in non-cultivated savannas, 50 m × 10 m (500 m) in gallery forest, and 50 m × 50 m (2500 m2) in cultivated areas [36]. These plot dimensions are used to take into consideration the spatial distribution of most species with unevenly spread individuals [36].

Data analyses
Firstly, useful species cited by informants were ranked according to the fifteen most cited ES of the four categories of provisioning, regulating, cultural and supporting ES [1,5]. Ethnobotanical indices were calculated to assess the importance of services provided by each species. They are: • the relative frequency of organ citation (RFO) with the adapted formula from Camou-Guerrero et al. where Ui denotes the number of different uses of a species and Ut the total number of people who cited the species.
Secondly, a nonparametric test of Kruskal-Wallis at the 5% threshold was carried out to compare the different ES quoted by the populations pertaining to age classes, gender, land use intensity and sociocultural groups, where χ 2 represents the approximate value of modal distribution. The degree of freedom, D.f = effectif-1 and a P value ˂ 0.05 indicate a significant difference in results. The analyses were processed with R software [38], with etno-botanyR and agricolae packages. A cluster analysis was carried out to determine the degree of similarity between the knowledge of different sociocultural groups; afterward, a principal component analysis (PCA) was performed with PcOrd9 software to assess the links between sociocultural groups (individuals) and knowledge on ES (initial variables).
Thirdly, the conservation status of the ten most used species was determined by the vulnerability index calculation [34] which is the average of the highest values of seven selected parameters (Table 2). According to Betti [34], if VI < 2, the species is assumed to be weakly vulnerable, if 2 ≤ VI < 2.5, the species is moderately vulnerable, and if VI ≥ 2.5, the species is highly vulnerable. An adaptation of the method of Lawrence et al. [39] was used to assess the orders of preference of causes of degradation, recommended solutions, and motivations for ecosystem preservation. The ranks given by each informant were converted into scores. The used scores were grades decreasing from the number of items in each question. For example, if 7 causes were cited for ecosystem degradation, the used scores were scores starting from 7 and decreasing in the order of the informant's citation. When an item was not cited by an informant, a score of 0 was given. For each item, the average score was calculated for each socio-cultural group, age group, and gender.
Vti: the average score given to a given item by a category of informants; Ti: the sum of the scores given to this item by this category of informants; Ni: the number of informants from this category of informants.

Diversity of used plant species and ecosystem services
Major ES providers to local populations were 163 species including 130 woody species and 33 herbaceous species (belonging to 122 genera and 42 families). In all categories of land-use intensity, the most dominant families were Fabaceae (35 species), Poaceae (17 species), and Malvaceae (13 species). Each ES concerned a great diversity of plant species. At least 60 species were used for 10 ES. Provisioning services were accomplished with the highest number of species (Fig. 2) and were the most cited (67%) by the informants. These services concerned medicinal use (120 species), fodder (76 species), and food supply (75 species). The species used in these three services had multipurpose uses. Diospyros mespiliformis, Khaya senegalensis, Vitellaria paradoxa and Saba senegalensis were quoted for 14 ES.
When considering the study area, the relative frequencies of number of citations varied significantly Vti = Ti/Ni between the different ES (χ 2 = 1849.6; Dl = 14; P value < 0.0001). By far the most quoted ES were food supply (25.98%) and medicinal use (20.89%), followed by fodder (7.33%), craft (6.10%) and energy supply (6.09%), which are all provisioning services (Fig. 3). Protection against wind (6.08%) and shading (6%) (regulating services) were important services, too. The relationships between ES and plant organs used for them (Fig. 4) showed that the entire plant was exclusively (100%) cited for six ES, notably soil fertility (supporting service), religion, tourism (cultural services), and shading, protection against wind and erosion prevention (regulating services).
Wood was exploited for the three provisioning services energy supply (100%), construction (100%) and craft (65%). Plant stalks were used only for craft and ceremony services. Plant leaves and fruits were cited for provisioning (food, fodder supply, medicinal use), regulating (pest control, water purification) and cultural (ceremonies) services with varying relative frequencies. Fruits were cited more often (70%) for food service. Flowers and seeds were used for food only. The highest number of organs (5) was cited for medicinal services, including bark and roots.
The use values (UV) computed for all quoted species showed that many species were exploited for several services. Woody species were the plant category with the highest UV (i.e., the most used species). Vitellaria paradoxa (UV = 3.775) was the species with the highest use value in the study sites. The three most used herbaceous species were Andropogon gayanus Kunth (UV = 0.723), Rottboellia cochinchinensis (Lour.) Clayton (UV = 0.400) and Hyptis spicigera Lam (UV = 0.370). In general, the ten species with the highest use values were in decreasing order: Vitellaria paradoxa, Parkia biglobosa, Diospyros mespiliformis, Adansonia digitata, Lannea microcarpa, Faidherbia albida, Khaya senegalensis, Afzelia africana, Ficus sycomorus, and Pterocarpus erinaceus (Table 3).

Variables influencing local knowledge of ecosystem services
With regard to the knowledge of local populations in different land-use intensity areas (    In order to understand a combined effect of land use area and sociocultural group, six sociocultural groups at the three land use sites were subjected to a cluster analysis, based on the citations of utilized ES (Fig. 5). These are: the groups of GRN Kassena, GRN Mossi, TWRB Mossi, CAD Pougouli, CAD Dagara and TWRB Dagara. The cluster analysis discriminated three groups, one comprising the Kassena and the Mossi at GRN, which is very different from the other two being composed of the natives at TWRB and CAD (the Dagara and the Pougouli) on the one hand and the Mossi at TWRB on the other hand.
Using the same data matrix, a principal component analysis grouped the 15 ES (initial variables) into 5 synthetic variables or principal components. The two first principal components are represented by the graph of the PCA ordination (Fig. 6). The first component (Axis 1) explains 88.08% of the total variation, the second component (axis 2) 8.04%. Therefore, these two axes explaining 96.12% of the total variation were used to describe relationships between sociocultural groups at different sites and ES. Axis 1 discriminated the natives of CAD and TWRB from the immigrant Mossi and the Kassena native to GRN. According to this axis, the Mossi of TWRB make use of the 15 ES only to a small degree and GRN residents mainly use the craft and tourism service while the natives of CAD and TWRB utilize the majority of services. Axis 2 discriminates ES utilization by the Mossi of TWRB from that by the GRN residents. Axis 2 underlines that the Mossi of TWRB make use of the 15 ES only to a small degree. The principal component analysis indicates that the utilization of ES is a function of the level of land use intensity and the economic benefits that people derive from plant formations. Thus, the GRN populations utilize the craft services related to tourism as they benefit from the financial income from tourism and participate in the management of the ranch. However, the natives of TWRB who do not benefit financially from the wildlife reserve of Bontioli and the natives of the anthropogenically shaped CAD utilize the ES that sustain the quality of the agricultural land and meet the primary needs of the rural populations.

Vulnerability of the species
The vulnerability indices of the used species in the study area varied from 1.57 to 2.71 (Table 3), indicating varying utilization pressures by local populations. Adansonia digitata, Afzelia africana, Diospyros mespiliformis, Khaya senegalensis, Lannea microcarpa and Pterocarpus erinaceus were most vulnerable with IV = 2.71. Eighteen other species were highly vulnerable with IV > 2.5 ( Table 3). The ten most utilized species were highly vulnerable, except for Faidherbia albida and Vitellaria paradoxa classified as moderately vulnerable (2 ˂ IV ≤ 2.5).

Local perceptions of availability and population dynamics of utilized species
The relative frequency of citation regarding the availability of all utilized species ( The population dynamics of the species are generally conceived as static because the proportion of opinions in favor of a decrease in species (40.08%) was approximately equal to that of the opinions in favor of their increase (40.86%). However, only 19.06% stated a constancy (Table 5). Among the most used species, V. paradoxa is the species with the highest positive dynamics (50%   increase and 20.51% constancy) while A. digitata is the one with the highest negative dynamics (52.3% decrease).

Perception of conservation of Sudanian savanna ecosystems by local populations
Regarding the variables leading to vegetation degradation, local populations agree that fire, cutting of fresh wood, and clearing for extension of cultivated areas were most important. In terms of importance, the majority of the sample ranked fire as the first cause of ecosystem degradation. However, for the Dagara of TWRB, the Kassena, and the Mossi of GRN, demographic increase is the most important cause (Table 6). Local populations ranked, by order of preference, five solutions for ecosystem conservation and four motivations for participating in sustainable management of ecosystems. For the suggested solutions, they ranked first "raising the awareness of local populations of the danger from degradation of natural resources," followed by "prohibition of fires by forest authorities" and "subsidy by the government" (Table 7). These solutions relate to a participation of local populations in the management of plant resources. The majority of the sample place awareness raising as the primary solution. However, the Pougouli, the Dagara of CAD, and the women place fire prohibition as the first solution, while the Dagara of TWRB place government subsidies as the first solution.
The motivation of local populations for sustainable management of plant resources is preferentially linked to their overall well-being due to the continuity of the provision of ES which will be ensured by long-term presence of the species, followed by the consideration of basic personal needs such as health, basic education, drinking water, and electricity provision by local authorities through the development of the village. Sustainability of vegetation is placed as the first motivation for sustainable management of plant resources by the majority of the sample. However, the elders, the Pougouli, the Mossi of GRN, and the men place village development as primary motivation (Table 8).

Diversity of utilized plant species and ecosystem services
Altogether, 163 plant species were cited by local populations as those providing them with different ecosystem services from Sudanian savannas. When considering the total number of species (1410) found by Zizka et al. [10] in the South Sudanian phytogeographic sector of Burkina Faso, only 11.6% of the potential flora of the study area are used by local populations. Each ES involves a great diversity of plant species: at least 60 species are used for 10 ES. This diversity of used resources makes it possible to overcome the problem of insufficient plant resources for a given service and could be exploited for the substitution of the most threatened species by those having a sufficiently high abundance in vegetation [23]. However, some specific ES involve specific species with specific property and characteristic. For example, the construction service is provided by stable and resistant species such as Khaya senegalensis (Desv.) A. Juss., Anogeissus leiocarpa, Burkea africana Hook. and Pterocarpus erinaceus Poir. [5]. In the study area, the most cited services were food supply followed by medicinal services. The importance of both services has been reported from investigations in the West [40], South [41] and North [35] of Burkina Faso and Côte d'Ivoire [42].
The relatively high percentage of use of fruits and leaves could be explained by their importance in various    services such as food, fodder supply and medicinal use. The high frequency of fruit citation (70%) for food supply shows the importance of fruit in the diet of local populations. Wild fruits contribute to a varied diet in terms of vitamin (A, B, C, D, and E) and micronutrient intake [41,43]. For example, the content of vitamin C in  fruit of Adansonia digitata and Detarium microcarpa is as high as in orange fruit; Moringa oleifera contains twice as much protein as yogurt. In addition, plant species used for food supply were also used for medicinal service. Sourabié et al. [44] reported anti-diarrheal effects of the fruit's pulp of Adansonia digitata and lowering of hypertension by Parkia biglobosa seeds.
The highest relative frequency (31%) of citations of the whole plant shows that local people are aware of the importance of vegetation and trees for their well-being, as the services they associate with the whole plant are regulatory, cultural, and supporting. These services are not destructive for plants and ecosystems but are rather conservative. This demonstrates local populations being  committed to the conservation of their environment which constitutes their living space. The 10 species with the highest use values were all woody species, and the shea tree (Vitellaria paradoxa) is a very popular species with the highest use value (3.775). In addition to its use in almost all services (14 of 15), it has a real use (UV) far more important than that of other species and a high socioeconomic value. This oleaginous species represents the fourth exportation product of Burkina Faso after gold, cotton, and livestock. Its high importance for populations was confirmed by other ethnobotanical studies [6,24,40,41,45,46]. Almonds and fruits of Vitellaria paradoxa, fruit pulp and seeds of Parkia biglobosa, and fruit pulp and leaves of Adansonia digitata are highly appreciated as non-timber forest products which provide income through their trade [47]. Shea butter (from Vitellaria paradoxa) and the African mustard, also called soumbala (from Parkia biglobosa), are transformed products with strong chains of added values [48,49]. Diospyros mespiliformis and Lannea microcarpa have highly appreciated edible fruits and medicinal uses. Faidherbia albida is a fodder woody species most appreciated by livestock breeders because it bears leaves and pods during the dry season when most of the woody species have shed leaves and fruits [35]. Faidherbia albida also has a high value of fertilizing cultivated or fallow soils [33] and is seen as a mystic plant by the Mossi [50]. Khaya senegalensis, Ficus sycomorus, Afzelia africana and Pterocarpus erinaceus are also fodder species with medicinal and cultural value [20,50,51]. Anti-malarial effects of Afzelia africana, Khaya senegalensis, Ficus sycomorus, Parkia biglobosa and Pterocarpus erinaceus are reported from Ghana [52]. The medicinal uses of the species most cited from our study area are also mentioned from other regions of Burkina Faso [35,44,53].
With regard to the high use value and the high demand for the products of most cited species, they constitute key species for local populations according to the definition by Clark and Sunderland [54]. However, once the value of a NTFP and the intensity of its use are extremely high, the resource is very likely to be overexploited, causing it to become locally extinct [54]. Gaisberger et al. [55] showed that overexploitation of species has emerged as the most important short-term threat. Overexploitation is the single most important threat for Parkia biglobosa (41.2%) and Vitellaria paradoxa (41.1%), and is only slightly exceeded by climate change in the case of Adansonia digitata (38.0%). The ethno-botanical use values correlated with the number of uses identified for each species and revealed the species preferred by local populations. However, the results must be taken with caution as the applied method does not distinguish between past, present and potential uses (some species may disappear because of anthropogenic pressure) [45,56].

Variables influencing the knowledge of ecosystem services
The three levels of land use intensity of this study design account for the economic benefit that local communities derive from them. The populations of the CAD and the TWRB live mainly from agriculture. The populations living near the TWRB engage in illegal activities such as farming, pasture and wood cutting in the protected area [14] as they do not benefit from economic benefits of the protected area in the same way as those of GRN. In fact, at least 18% of TWRB had been cleared [57] by local populations to install their fields. Forest administration has great difficulties to prevent neighboring villagers from using the resources in protected zones that they highly depend on [7,13] as long as no incentives are offered for compensation. In contrast, inhabitants living near GRN are employed with the forest officers to ensure functioning of this tourist attraction and vary their income by participating in forest management. Likewise, hunting in the village hunting zone around GRN and fishing in water points of the ranch provide populations with additional income. Therefore, providing biodiversity conservation actors with diversified sources of income is economically important to local communities [15,58]. The management of natural resources involving local communities contributes to better security of biodiversity through sustainable participatory management [59].
As for the sociocultural groups, the Pougouli and Dagara had more knowledge in ES referring to food supply, religion, wind protection, shading, medicinal use, soil fertilization, erosion prevention and energy supply. Their knowledge in religion services expresses the animist cult influence of Pougouli. According to the national statistics of population, the populations of the Southwest region have 64.9% of animists [60]. The high rate of religion ES is a means to preserve the surrounding vegetation and ecosystem. Religions are excellent channels for transmitting local knowledge through initiations (i.e., traditional and spiritual instructions) which are part of the education of the youngest in the preservation of the environmental values. To the animist communities such as Pougouli and Dagara, forests are the habitats of venerate spirits [50,51]. The good knowledge about species providing medicinal services may be explained by the preservation of their ancestral knowledge transmitted from generation to generation [3,16,61]. The use of species for food supply, wind protection, shading and energy supply ES reflects the poverty of these populations who heavily depend on income from agriculture and small livestock. The Dagara and Pougouli are introvert sociocultural groups little open to exterior influence and quite attached to their local environment. Pougouli and Dagara knowledge is also orientated to performing farming. They are essentially farmers and do not hesitate to transgress protected areas to install their fields [14]. The population of GRN pays little attention to medicinal services due to the fact that they live nearby a health center, leading to a loss of local knowledge about medicinal plants.
Local populations are subjected to increasing social (demographic and economic) and environmental pressures which have mostly led to a loss of knowledge [24,62]. TWRB Mossi had no specific knowledge about ES provided by Sudanian savanna species. The lack of specific knowledge about Sudanian vegetation of the Mossi migrants of TWRB could be related to the fact of having immigrated from the Sudano-Sahelian zone where the species composition is different. The ethnobotanical knowledge varies, effectively, across sociocultural groups because of cultural differences and social habits [4,63]. In addition, the migrant communities of TWRB are located outside the natives' villages which constitutes an obstacle for passing on and sharing inter-community knowledge.
On the contrary, the Kassena and GRN Mossi sociocultural groups possessed similar knowledge and had more knowledge in craft, energy and touristic services than the Mossi of TWRB, the Native of TWRB and CAD. This demonstrates that the Mossi migrants of GRN are well integrated and adapted to the native way of life. The Mossi migrants of GRN have replaced species which they originally used by species which serve the same purpose among the natives (the Kassena). Thus, the traditional use of species by migrant communities can be influenced more strongly by the environment than by cultural heritage [64].

Vulnerability of the species
The multipurpose use of species, the utilization of slow regenerating plant organs (as wood, seeds, bark, roots and flowers), and the preference (the high use value) that populations have for some species act to increase their vulnerability. The 10 species with high use value were highly vulnerable, except for Faidherbia albida and Vitellaria paradoxa classified as moderately vulnerable. The state of vulnerability of Adansonia digitata, Afzelia africana, Khaya senegalensis, Parkia biglobosa and Pterocarpus erinaceus is confirmed by Thiombiano et al. [65] who classify them threatened species of the southern Sudanian zone. According to these authors, Diospyros mespiliformis would be more threatened in the northern Sudanian zone. In the sub-Sahelian zone, Ouedraogo et al. [35] confirm high vulnerability of Adansonia digitata, Diospyros mespiliformis, Faidherbia albida, Khaya senegalensis, Lannea microcarpa, Parkia biglobosa, Pterocarpus erinaceus and Vitellaria paradoxa. All the 10 most used species of this study are also considered endangered by local populations of northern Benin [66]. Nevertheless, for the vast majority of plant species in Burkina Faso insufficient data are available for a full IUCN assessment [11,27]. Globally and according to the IUCN red list [67], the conservation status of Pterocarpus erinaceus has been decreasing and has become Endangered, indicating a high risk of extinction. Afzelia africana, Khaya senegalensis and Vitellaria paradoxa conservation status are classified Vulnerable, indicating risk of extinction. The status of Parkia biglobosa, Diospyros mespiliformis, Adansonia digitata, Lannea microcarpa, Faidherbia albida and Ficus sycomorus is Least Concern. Although these species have different levels of vulnerability in other regions of Burkina Faso [11,35] due to utilization by local populations, they are of great interest to the populations of our study sites. It would then be appropriate to think about their sustainable management through rational use motivated by the perceptions of local populations.

Local perceptions about availability and population dynamics of the most used species
Perceptions of local populations on the availability and population dynamics of used species follow the same trend in the study area. Local populations stated that Vitellaria paradoxa was the most abundant species and showed an increase, whereas Adansonia digitata was very rare and showed a decrease. Traoré et al. [20] found that, in the Southwest of the country, perception of the state of resources by the local Senufo is consistent with the findings in the field. Indeed, 64% of the species cited by the Senufo as being rare are part of the rare species revealed by the calculation of the rarity index at the end of floristic inventories. Thus, the least variation in the availability of important species is perceived by local populations. The socioeconomic importance, availability and population dynamics of species define the management of traditional agroforestry systems. According to Assogbadjo et al. [21], species perceived by local communities as threatened are integrated into traditional agroforestry systems. However, the populations admit general degradation of the ecosystems.

Local perception of conservation of Sudanian savanna ecosystems
According to local populations, fire, deforestation and clearing constitute the three main causes of Sudanian vegetation degradation. Local populations use fire as a tool for hunting, clearing of village surroundings and field preparation [63]. Deforestation by wood cutting results from the demand for energy, craft and construction services. The clearing of natural vegetation is practiced for installing new and extending existing agricultural land. Local populations are conscious of the causes of the degradation of plant resources, being essentially them provoking ecological pressure originating from increasing needs of growing populations [7,13,14]. While the development of agriculture has made it possible to free oneself from dependence on wild food, the expansion of agricultural land strongly contributes to ecosystem and biodiversity degradation [68]. Land use changes effectively have a negative impact on biodiversity due to habitat loss or fragmentation [20]. In the area of high land use intensity (CAD), species diversity is low and only species which provide ES are conserved.
Raising the awareness of local populations in the face of degrading natural resources is the first solution unanimously suggested by the communities. Given that climate change is not well perceived at the local scale, awareness raising may contribute to reducing anthropogenic pressure (fire, wood logging). Local population's awakening of awareness as to the vulnerability of plant resources used every day and their implication for natural resources management in collaboration with forest authorities are steps that would enhance sustainable conservation of plants. The success of biodiversity conservation efforts often depends on local populations, especially when these communities are the key players in ecosystem management [69]. In fact, high biodiversity existing in native territories around the world is the result of traditional knowledge and management practices [16]. Fair collaboration of local populations with the forest authority for conservation of plant resources could be achieved, for example, by meeting certain basic needs such as the creation of water reservoirs and the construction of health centers and schools. Sustainability of vegetation allows for diversification of incomes, especially for women who collect and sell NTFP to complete their incomes [24]. Local processing of NTFP by women before selling contributes to the value chain of products and generating incomes. For example, the almond of Vitellaria paradoxa is transformed into shea butter and the seed of Parkia biglobosa is transformed into soumbala (African mustard) [48,49].

Conclusion
This study has highlighted the importance of ES of Sudanian savannas and the importance of long-term preservation of numerous plant species. It shows that local populations do know their environment well and that they are aware of the causes of degradation of plant resources. Therefore, local people should be involved in programs and projects addressing sustainable management and conservation of Sudanian savanna ecosystems. The behavioral change can be achieved by awareness raising and education. Education of local populations needs to involve best management strategies and promotion of domestication and regeneration of local multipurpose species. Also species with low utilization value should be protected for conserving future biological resources because species being less important today could be sought for in the future to replace others having become rare in the meantime.
However, in face of local population vulnerability and their strong dependence on natural resources, it is often difficult to consolidate conservation and rational and sustainable use of biodiversity. In order to achieve effective awareness raising among local populations for sustainable management of plant resources, it would be necessary to offer them solutions and motivations adapted to their perceptions and their consents. This approach will enable full collaboration of the sociocultural communities.
Abbreviations ES: Ecosystem services; TWRB: Total wildlife reserve of Bontioli (medium land use intensity); CAD: Communal area of Dano (high land use intensity); GRN: Game ranch of Nazinga (low land use intensity.