Assessment of plant communities' pattern and diversity along a land use gradient in W Biosphere Reserve, Benin Republic Laurent Gbenato Houessou1,2*, Anne Mette Lykke3, Oscar Semadegbe Teka2, Aristide Cossi Adomou2,4, Madjidou Oumorou5, Brice Sinsin2 1 Laboratory of Ecology Botany and Plant Biology, Faculty of Agronomy, University of Parakou (Benin) 2 Laboratory of Applied Ecology, Faculty of Agronomic Sciences, University of AbomeyCalavi (Benin) 3 Department of Bioscience, Aarhus University, Vejlsøvej 25, 8660 Silkeborg (Denmark) 4 National Herbarium, Faculty of Sciences and Technics, University of Abomey-Calavi, Abomey-Calavi (Benin) 5 Laboratory of Research in Applied Biology, Polytechnic School of Abomey-Calavi, Department of Environment, University of Abomey-Calavi, (Benin) *Corresponding Author: houessoulaurent@gmail.com Abstract Human disturbance on vegetation is an important concern in biodiversity conservation. In this study we assessed how anthropogenic disturbance affected plant communities pattern, diversity, life form and chorotype composition along a land use gradient. Vegetation relevés were performed along a land use gradient (park-buffer zone-communal land) at W Biosphere Reserve in Benin. Non-metric multidimensional scaling (NMS) was used to assess plant communities patterns. Indicator species were determined for each plant community and land use. Plant community diversity, life forms and chorotypes composition were assessed and compared among land uses using one-way analysis of variance. NMS ordination showed a good separation between relevés of the park and those from the communal land while relevés of buffer zone were mixed within the park and communal land relevés. There was no significant difference between species richness among land uses types (F = 0.68; p = 0.529, ANOVA test at a level of significance of 5%). The Pielou evenness for the plant communities was higher in the park (E= 0.69±0.04) and buffer zone (E = 0.61±0.13) than in the communal lands (E = 0.44±0.02) while Shannon index showed no clear pattern along land use gradient. Therophytes abundance was significantly higher in the communal land while hemicryptophytes abundance was significantly higher in the park. Wide-distributed species abundance was significantly higher in the communal land whilst Sudanian species showed significantly higher abundance in the park. We concluded that monitoring of the indicator species of the plant communities and their traits are relevant tools for managers to follow-up changes in plant communities. Introduction Aside environmental factors, disturbance is considered as a factor affecting plant community structure, distribution, composition and functionality (Biswas and Mallik, 2010; Nacoulma et al., 2011). Human disturbance on vegetation is nowadays an important concern in biodiversity conservation and current global change (O’Connor, 2005, Southworth et al., 2016). The effects of this disturbance could lead to: (i)- rarity and vulnerability of some species (Adomou et al., 2006), (ii)- high occurrence of invasive species (Aboh et al., 2008), (iii)- ecosystem loss or habitat fragmentation (Thompson et al., 2017). Moreover, human disturbance could completely change the species composition of the original plant communities, which in some cases can become irreversible (Kassi N’Dja West African Journal of Applied Ecology, vol. 27(2), 2019: 61 - 78 62 West African Journal of Applied Ecology, vol. 27(2), 2019 and Decocq, 2008; Lindenmayer et al., 2017). In tropical region, many anthropogenic factors were targeted as inducing notable changes in vegetation. Flamenco-Sandoval et al. (2007) outlined clearing for agriculture, i.e. slash and burn cultivation, grazing and tree logging as important driving forces contributing to land use and land cover change and accordingly vegetation composition. For instance, most studies on vegetation fire underlined meaningful effect of fire on vegetation pattern in savannas and plant communities’ structure, composition as well as their traits (van Wilgen et al., 2007). Meanwhile grazing systems effects on plant communities’ diversity, structure, composition, life form and productivity were evidenced by previous studies (Lezama et al., 2014). In that way, Hendricks et al. (2005) observed perennial species substitution by annual species near livestock camp in South Africa. Likewise, O'Connor et al. (2011) underlined longterm decrease of forbs richness under a high stocking rate. As disturbance results, patches of vegetation in different stages of succession are distributed across the landscape. Most previous ecological studies have documented the successional vegetation patterns as temporal and spatial change in vegetation composition (Fournier et al., 2001; Kassi N’Dja and Decocq, 2008). Following time scale, the vegetation composition goes from, vegetation dominated by pioneer plant species to secondary pseudo-stable vegetation with more competitive species (Kassi N’Dja and Decocq, 2008). Depending on the interaction between anthropogenic disturbance intensity and abiotic factors, plant community traits as well as their composition shift over time and space at each vegetation stage and reflect the ongoing process in the plant community (Bangirinama et al., 2010). Therefore, the knowledge on the change occurring in plant communities’ characteristics across land use can enable to understand how far human disturbance affect plant community’s composition and diversity and provide reliable tools for phytodiversity monitoring and management. Except for researches already reported on human disturbance on tree species communities or herbaceous communities more often separately (Shackleton, 2000; Nacoulma et al., 2011), little is known about how anthropogenic disturbance shapes the whole plant communities’ pattern and affects community diversity and floristic composition. Hence, in this study we focused on a gradient going from communal land to core area of a biosphere reserve, where anthropogenic disturbance is considered as absent. Overall, we aim to describe changes in floristic composition along a land use gradient. More specifically, the study aims i)- to assess plant community patterns, ii)- to determine change in indicators species along the land use gradient, iii)- to determine alpha and beta diversity of plant communities, iv)to investigate the change in plant communities traits (life form and chorotypes) in order to provide managers with simple and reliable tools for monitoring and evaluating the success of ongoing conservation actions with respect to phytodiversity. Material and methods Study Area The study was conducted in the W Biosphere Reserve in Benin (WBR) (11°26’-12°26’ N; 2°17’-3°05’ E, Figure 1). The WBR is Houessou et al: Assessment of plant community pattern and diversity along a land use gradient in W Biosphere Reserve composed of the park and its adjacent hunting zones and is under the administration of the National Centre for Wildlife Reserves Management (CENAGREF) that outlines and implements management and conservation actions of the reserve (Clerici et al., 2007). The reserve is located in the Sudanian centre of endemism (White, 1983), where climate is characterized by one rainy season (May to October) and a dry season (November to April). The mean annual rainfall experienced ranges from 900 mm to 1100 mm. The mean monthly temperature ranges from 25 to 35°C and values of the relative air humidity range from 81% in August to 26 % in February (ASECNA, Unpubl. data). Overall, soils are tropical ferruginous type and characterized by moderate fertility (Viennot, 1978). Anthropogenic activities (livestock grazing, cropping, uncontrolled fire and logging) are strictly prohibited inside the reserve. At the periphery of the reserve, a 5 km land belt (buffer zone) is set up to stop anthropogenic pressure from the communal lands on the park. In the buffer zone, crop growing, Non-timber Product Forests (NTFPs) harvesting and livestock grazing are allowed but subjected to restrictions. In contrast to the park and buffer 63 zone, there is no restriction with respect to human activities in the communal lands. Thus, this latter is subjected to high anthropogenic disturbance. Cropping based on shifting cultivation and livestock breeding based on extensive use of pastureland represent the most important socio-economic activities of the local populations. The main cultivated crops are cotton, sorghum, corn and millet. Cattle, sheep and goat are the main livestock farmed. Uncontrolled fires are frequently applied by the local populations in order to favour perennial grasses regrowth, for poaching and for land cleaning according to local perception. The population density in the peripheral WBR is about 20.0 inhabitants km-2 (INSAE, 2013). Vegetation types occurring in the reserve are composed of a mosaic of savannas (shrub, tree, grass savanna and woodland) and gallery forest (CENAGREF, 2008). In the communal lands, vegetation is dominated by croplands and fallows and degraded savanna. Roughly the area is divided in three land uses (park, buffer zone, communal lands) presenting a gradient of land use from protected to nonprotected area. Then we assume that the area is suitable for studying the land use gradient effect on plant communities’ pattern and Figure 1: Location of W Biosphere Reserve 64 West African Journal of Applied Ecology, vol. 27(2), 2019 diversity. Data Collection Landsat 8 OLI/TIRS satellite image of October 2017 (path 192 and row 52) was processed by using normalized difference vegetation index "NDVI" to enhance vegetation contrast. Then we performed the maximum likelihood supervised classification which enable to identify the main patches of vegetation in each land use type (i.e. park, buffer zone and communal lands). Based on the processed image, we set up a total of 120 stratified random plots in five main vegetation types (woodland, gallery forest, shrub/tree savanna, grass savanna and fallow) in the three land uses types. Trees and shrubs sampling was carried out in 30 m x 30 m plots and herbaceous floristic composition was carried out through phytosociological relevés in subplots of 10 m x 10 m (Weber et al., 2000). In each plot we recorded directly in the field: (i)- exhaustive species list, species naming conventions were taken from Benin flora (Akoègninou, 2006); (ii)- percentage cover for each species; (iii)vegetation types; (iv)- soil texture using visual assessment (clayey soil; silty soil; sandy soil, gravely soil); (v)- level of perturbation (grazing disturbance) based on a visual assessment of the level of clipped vegetation and cattle footprint presence. Data Analysis Plant Community Ordination, Classification and Indicator Species Determination A presence absence data matrix of the 120 vegetation relevés was analysed using PC-Ord (McCune and Mefford, 2006). Non-metric multi-dimensional scaling (NMS) based on Sørensen (Bray-Curtis) distance measure was used for the vegetation ordination (Kruskal, 1964). We used NMS autopilot to determine the number of axes which gave the best configuration of the relevés in the ordination space (McCune and Grace, 2002). Cluster analysis was used to classify the plant communities in each land use type (with Sørensen distance measure and flexible beta linkage method). The number of plant communities was determined using indicator species analysis, which implied selecting the number of clusters that had the smallest average p-value and the highest number of indicator species (McCune and Grace, 2002). A cover-abundance data matrix was used for the indicators species determination. For each plant community and land use type, the indicator species were selected numerically following the method of Dufrêne and Legendre (1997). The indicator species determination in each land use type was done by adding to the initial data matrix of relevés an additional variable describing the land use type to which the relevé belong. The same rule was used for the plant communities’ indicator species determination. The Indicator Species Analysis Package in PC-Ord was run to determine the indicator value (IV) of each species in each land use type and in each plant community. The indicator value is the combination of the species relative abundance (Ai in %) and relative frequency (Bi in %) in each land use type and in each plant community (IV = Ai x Bi). The Monte Carlos test of permutation was performed on the indicators values to determine the plant species which indicator value was significant. The indicator species were represented by the species which had the highest indicator value and significant Monte Carlos test (p < 0.05). In the specific case of land use indicators Houessou et al: Assessment of plant community pattern and diversity along a land use gradient in W Biosphere Reserve species determination, we selected plant species which indicators values probability of Monte Carlos test is <0.01 and had a high IV value in the considered land use type comparing to the other land use. Intra-community diversity of the Plant Communities The intra-community diversity (α-diversity) of the plant communities was assessed using the species richness, the Shannon diversity index (Shannon, 1949) and the Pielou evenness (Pielou, 1969). Plant community species richness Species richness was determined by counting the number of species recorded in the relevés describing each plant community. We computed the total number of species recorded per plant community and estimated the species richness for woody and herbaceous layers. Plant community Shannon index - It was estimated as: where pi is the relative abundance of the species i in a given plant community and S the species richness of the community. Plant community evenness - It was computed as: where S is the total number of species per plant community and H’ is the value of the Shannon index. E values range from 0 (dominance of few species in the community) to 1 (evenly distribution of plant species in the community). 65 Beta Diversity of the Plant Communities In contrast to alpha diversity, beta diversity (β-diversity) is considered as the intercommunity diversity i.e. within plant communities (Magurran, 2004). β-diversity describes the change (turn over) in species composition between two plant communities. We used it hereafter to determine floristic change between land use compositions. It was estimated as: where Sørensen similarity index = 2c / (a+b); and a = number of species recorded only in the plant communities of land use A, b = number of species recorded only in a plant communities of land use B and c = number of species shared by both communities of land use A and B. The values of β-diversity range from 0, for a complete similarity, to 1, for an absence of similarity. We assumed that if β-diversity > 0.5, the plant communities in two land use types were floristically different (Mwaura and Kaburu, 2009). Plant Community Composition in Life Forms We assigned life forms to species using those defined by Raunkiaer (1934): Therophyte (THERO); Hemicryptophyte (HEMI), Chamaephyte (CHAM), Cryptophyte (CRYP) and Phanerophyte (PHAN). Afterward, the abundance of each life form type (CLFi) was calculated for each plant community according to the formula: where ni = number of the species with the life forms i in plant community and S = total number of species in the community. In 66 West African Journal of Applied Ecology, vol. 27(2), 2019 addition, we estimated the percentage cover of each life form type according to the formula: where ri = percentage cover of the species with the life i in the plant community and R = total percentage cover of all species in the community. Plant Community Composition in Chorotypes For chorotypes composition assessment, we used the classification defined by White (1983) and grouped species into three main chorotypes i.e. Sudanian species (S); Wide distribution species (WD), corresponding to afro-american, pantropical and paleotropical plant species, and Continental distribution species (CD) corresponding to afro-malagasy, afro-tropical, pluri-regional African species and sudano-zambesian species. On this basis, the composition in chorotype for each plant community was estimated following the rules described above for life forms. Both percent abundance and percent cover were assessed. Statistical Analysis The data were log-transformed to meet the assumptions of normality and homogeneity of variance for each of the estimated parameters (Dagnelie, 2011). One-way analysis of variance (ANOVA) test was performed in Minitab 18.1 to determine if there were significant differences between species richness, life form composition and chorotype composition for the plants communities along the gradient going from communal land to the park. Results Plant Communities Pattern and Classification In the 120 sample plots inventoried, an overall of 338 plant species belonging to 57 plant families were noticed. The NMS ordination diagram shows a good separation between relevés of the park and those from the communal land. Relevés of the buffer zone were mixed within the park and communal Figure 2: Diagram of the projection of the 120 plots in the first two axes of the non-metric multidimensional scaling (NMS). A two-dimensional solution was obtained by NMS autopilot for a best configuration of the plots. The final stress = 18.14 and final instability = 0.0032 with 200 iterations and 50 runs with randomized data. Legend: Ap = Park; BZ = Buffer zone and CL = Communal land Houessou et al: Assessment of plant community pattern and diversity along a land use gradient in W Biosphere Reserve 67 Figure 3: Clustered plants communities per land use based and common plant communities between land uses as revealed by indicator species analysis. Legend: P1, P2, P3, P4, P5 = Plant communities clustered in the park, C1, C2, C3, C4 = Plant communities clustered in the communal. B1, B2, B3, B4 = Plant communities clustered in the buffer zone land (Figure 2). NMS ordination on the relevés for each land use followed by cluster analysis allowed classification of the park relevés into 5 main plants communities (P1, P2, P3, P4, P5), buffer zone relevés into 4 communities (B1, B2, B3, B4) and communal land relevés into 4 communities (C1, C2, C3, C4) (Appendix). Based on the indicator species analysis, results showed that the Loudetia togoensis & Bulbostylis abortiva community was found in park (P5), buffer zone (B4) and communal land (C1). The Andropogon tectorum & Costus spectabilis community (B3 in the buffer zone; P4 in the park) and the Crossopteryx febrifuga & Andropogon gayanus community (B2 in the buffer zone; P2 in the park) were both present in the buffer zone and in the park (Figure 3 & Appendix). Species Richness, Shannon Index and Pielou Evenness of Plant Communities Overall, the mean (± standard error) species richness recorded per plant community was lower in the park (128.6 ± 45.2) comparatively to the buffer zone (135.1 ± 42.6) and the communal land (143.2 ± 40.9). However, there was no significant difference between species richness among land uses (F = 0.08; TABLE 1 Mean alpha diversity of the plant communities in each land use Vegetation Layer Herbaceous Woody Overall Alpha diversity Species richness Shannon index (bits) Pielou evenness Species richness Shannon index (bits) Pielou evenness Species richness Shannon index (bits) Pielou evenness Land use Park Buffer Zone 120.6±29.96 117±42.44 3.27±0.41 3.46±0.28 0.67±0.08 0.59±0.23 21.25±12.03 19.25±6.06 4.42±0.64 4.13±0.91 0.81±0.22 0.72±0.11 128.6±45.2 135.1±42.6 4.08±0.29 3.94±0.44 0.69±0.04 0.61±0.13 Significance (Anova Communal land test at 5% level) 132.5±37.04 F=0.68; P =0.529 3.15±0.65 0.45±0.01 19±7.84 F = 0.42; P = 0.669 2.05±0.37 0.44±0.19 143.2±40.9 F = 0.08; P = 0.925 3.96±0.61 0.44±0.02 68 West African Journal of Applied Ecology, vol. 27(2), 2019 p = 0.925; Table 1). At a significance level of 5%, ANOVA test showed that there was no significant difference between species richness for the herbaceous layer (F = 0.68; p = 0.529) and the woody layer among the three land use types (F = 0.42; P = 0.669). Considering the woody layer, the mean Shannon index value of the plant communities was higher in the park (H’ = 4.42 ± 0.64) and buffer zone (H’ = 4.13 ± 0.91). In contrast, the woody species diversity was low in the communal land (H’ = 2.05 ± 0.37) with an uneven distribution in communal land (Pielou evenness < 0.5). Considering the herbaceous layer, the diversity was intermediate in the three land use types. However, Pielou evenness displayed low value in the communal land showing the dominance of few species in the communal communities Beta diversity among Land Use Type β-diversity displayed high floristic similarity between the buffer zone and the park (β = 0.10). Moreover, we found that plant communities in the park and buffer zone were floristically different from those in the communal land (β > 0.5). Park and buffer zone shared 193 species (58 %) while the number of shared species between the buffer zone and communal land was 108 species (32 %). Park and communal TABLE 2 Indicators value of the plant species in each land use Land use Park Buffer Zone Communal Land Species Park Androgon gayanus 55 Diheteropogon amplectens 30 Hyparrhenia smithiana 32 Loudetia arundinacea 24 Lannea barteri 16 Loxodera ledermannii 14 Aganope stuhlmannii 35 Andropogon chinensis 29 Andropogon tectorum 16 Tinnea barteri 22 Indigofera paniculata 15 Aspilia angustifolia 8 Chasmopodium caudatum 10 Indigofera bracteolata 5 Polygala arenaria 13 Hyparrhenia smithiana 5 Sorghastrum bipennatum 1 Ximenia americana 4 Andropogon schirensis 12 Andropogon pseudapricus 6 Commelina erecta 13 Flueggea virosa 0 Pennisetum polystachion 1 Senna obtusifolia 0 Setaria pumila 0 Sida acuta 0 Indicator Value Buffer zone Communal land 30 0 2 0 11 0 0 0 0 0 0 0 5 0 15 0 0 0 6 0 0 0 39 1 46 1 50 5 42 8 34 1 41 0 34 0 34 0 28 3 35 4 1 51 2 70 0 36 1 76 1 31 P-value 0.0002 0.0004 0.0004 0.0004 0.0008 0.002 0.0024 0.0032 0.0034 0.0034 0.0042 0.0002 0.0002 0.0002 0.0002 0.0004 0.0004 0.0004 0.0006 0.0014 0.003 0.0002 0.0002 0.0002 0.0002 0.0002 Houessou et al: Assessment of plant community pattern and diversity along a land use gradient in W Biosphere Reserve 69 TABLE 2 continued Indicators value of the plant species in each land use Land use Species Communal Land (continue) Vitellaria paradoxa Tephrosia pedicellata Triumfetta rhomboidea Sida cordifolia Desmodium hirtum Euphorbia hyssopifolia Euphorbia convolvuloides Wissadula amplissima Paspalum scrobiculatum Dichrostachys cinerea Park 0 0 0 0 0 0 3 0 0 1 lands shared 30 % of the species (106 species). Indicator Species Change among Land Use Floristic analysis based on the indicator value of the plant species in each land use type revealed that Androgon gayanus, Diheteropogon amplectens, Hyparrhenia smithiana, Loudetia arundinacea, Loxodera ledermannii, Andropogon chinensis and Andropogon tectorum yielded high indicator value in the park (Table 2). Postcultural or ruderal species such as Senna obtusifolia, Tephrosia pedicellata, Triumfetta rhomboidea, Sida cordifolia, Desmodium hirtum, Indicator Value Buffer zone Communal land 1 52 1 81 0 64 0 16 0 19 0 16 1 24 0 17 0 13 1 43 P-value 0.0002 0.0002 0.0002 0.0008 0.0012 0.0016 0.0028 0.0044 0.005 0.0066 Euphorbia hyssopifolia, Dichrostachys cinerea, Euphorbia convolvuloides presented high indicator value in the communal land. Buffer zone displayed a pioneer species as well as perennial Poaceae with high indicator value: Andropogon schirensis, Hyparrhenia smithiana, Andropogon pseudapricus, Commelina erecta and Indigofera bracteolata, Ximenia americana. Life Forms Composition of Plant Communities Life forms composition of the plant communities showed that the percent abundance as well Figure 4: Life forms composition of the plant communities: (a) – Percentage of abundance of the life forms in the plant communities; (b) – Percentage cover of the life forms in the plant communities Legend: **P < 0.01, ***P < 0.001, ns for P>0.05; Ap = Park; BZ = Buffer zone; CL = Communal land; THERO= Therophyte; HEMI= Hemicryptophyte, CHAM = Chamaephyte, CRYP = Cryptophyte and PHAN = Phanerophyte 70 West African Journal of Applied Ecology, vol. 27(2), 2019 Figure 5: Chorotypes composition of the plant communities: (a)- Percent abundance of the chorotypes in the plant communities; (b)- Percent cover of the chorotypes in the plant communities. Legend: **P < 0.01, ***P < 0.001, ns for P>0.05; Ap = Park; BZ = Buffer zone; CL = Communal land; WD = Wide distribution species; S = Sudanian species; CD = Continental distribution species as the percent cover of hemicryptophytes and therophytes were significantly different among the three land uses at a significance level of 1% (p < 0.01, Figure 4a & 4b). Park presented higher percent abundance (16.24 ± 4.64 %) and higher percent cover (42.00 ± 13.89 %) in hemicryptophytes compared to the buffer zone and communal land. Therophytes yielded high percent abundance (55.20 ± 4.16 %) and high percent cover (68.86 ± 7.90 %) in the communal land. At a significance level of 5%, ANOVA test showed that land use type had no significant effect on the phanerophyte percent abundance (F =3.31; p = 0.079; Figure 4a). However, the cover of the phanerophytes was significantly different among the land use types at a significance level of 1% (F = 7.76; p = 0.009, Figure 4b). Results showed no significant difference between the land use types with regard to cryptophytes and chamaephytes. types (ANOVA test at a significance level of 5 %). Sudanian species percent abundance (F = 46.49, p < 0.001, Figure 5a) and wide distribution species percent abundance (F = 17.99, p< 0.001, Figure 5a) were significantly different among the land use types based on ANOVA test at a significance level of 1 %. Plant communities in the park, sheltered a high percent abundance of Sudanian species (39.7 ± 2.86 %) and those in communal land exhibited low percent abundance (17.49 ± 4.53 %) in Sudanian species. Plant communities in the communal land yielded higher percent abundance (32.69 ± 6.09 %) of wide-distribution species than those in the park. Plant communities in the buffer zone presented an intermediate situation. The same tendency was observed for percent cover of the different chorotypes’ composition in the plant community (Figure 5b). Chorotypes Composition of Plant Communities Regarding the chorotypes’ composition of the plant communities, we found no significant difference for the species with continental distribution among the land use Discussion Plant Communities Pattern and Floristic Change In this study, we focused on potential Houessou et al: Assessment of plant community pattern and diversity along a land use gradient in W Biosphere Reserve floristic change of plant communities along a gradient of land use (park-buffer zonecommunal land). Results showed a clear difference between communal land relevés, where anthropogenic disturbance occurred, and park relevés where the vegetation was undisturbed. Buffer zone relevés were mixed within park relevés and communal land relevés. These results highlight the human disturbance influence on plant communities’ distribution and composition (Koulibaly et al., 2006; Liu et al., 2009). Plant communities pattern and indicator species analysis showed that, although there were plant communities exclusive to each land use yet some of them were shared between the land uses. The Loudetia togoensis & Bulbostylis abortiva community was found to be common to the three land uses suggesting that this community may be less affected by disturbance. Indeed, this plant community thrives on shallow and poor soil which is named "bowe" (Padonou et al., 2014). This type of soil is not suitable for agriculture purpose and is hence set apart during land clearing for cultivation. However, the herbage on "bowe" is grazed by cattle in the communal lands. Plant communities in the communal lands derived from secondary successional pattern as evidenced by the indicators’ species of the plant communities in this land use type. For instance, Tephrosia pedicellata and Triumfetta rhomboidea were described as indicating overgrazed sites near to hamlets (Fournier et al., 2001) while Spermacoce stachydea, Digitaria horizontalis and Schizachyrium exile are generally associated with young fallow (1-3 years) on poor soils (Sinsin, 1993). Plant species such as Dichrostachys cinerea, Piliostigma thonningii and Flueggea virosa are indicators of old fallow (5-10 years) on 71 soil which is recovering its fertility. In the park, indicator species analysis revealed that perennial Poaceae such as Andropogon gayanus, Andropogon schirensis, Andropogon tectorum and Loxodera ledermannii were the main indicator species in the park. Although variation can occur depending on the soil, these species are often described as characterising the ultimate stage of succession, which is maintained by the annual cyclic fire in savanna (Fournier et al., 2001). Indicators species in the buffer were both represented by perennial grasses and ruderal species. This highlights the dynamics of indicator species in the buffer area from perennial Poaceae species to postcultural and ruderal species characteristic of disturbed area. We deduced that indicator species for plant communities shift along the gradient of disturbance depending on the level of disturbance affecting the communities. β-diversity highlighted similarities between plant community composition in the buffer zone and the park. In contrast, plant communities in the communal lands differed floristically from those in the buffer zone and park. These results suggest that human disturbance influence the composition and distribution of plant species and communities as demonstrated by earlier studies (Kassi N’Dja and Decocq, 2008; Biswas and Mallik, 2010). In fact, in the communal lands adjacent to the park, agriculture and cattle breeding emerged as the main activities practiced in the area (Clerici et al., 2007). It is well known that extensive vegetation clearing for agriculture as it is practised around the park leads to fallow and semi-natural vegetation changing the original composition of the plant communities (Kassi N’Dja and Decocq, 2008). In addition, high intensity grazing modifies both the structure 72 West African Journal of Applied Ecology, vol. 27(2), 2019 and composition of the plant communities by selective grazing of species and trampling as highlighted by Haarmeyer et al. (2010). Plant Communities Species Richness, Shannon Diversity Index and Pielou Evenness Our findings showed that there was no significant difference in plant communities species richness between the different land uses. However, species richness of the plant communities in the communal lands was numerically higher comparing to the park and to the buffer zone. This suggests that although disturbance affects plant community composition along the gradient of land use, the pattern of species richness in the plant communities did not follow the gradient going from communal land to the park. This result could be supported by the intermediate disturbance hypothesis which predicts that during the course of vegetation succession, the species richness was higher at the intermediate stage than at the final stage (Wilkinson, 1999; Biswas and Mallik, 2010). In our case, plant communities in the communal lands were at earlier stage and intermediate stage of succession while plant communities in the park were at the final stage and more stable. Therefore, it is not surprising to find high species richness in the plant communities in the communal land comparing to the park. However, our findings were contrary to those of Nacoulma et al. (2011) who in a related study found that the species richness of the plant communities in the protected area was significantly higher than in the communal land. Nonetheless, our results were corroborated by Shackleton (2000) who found higher species richness for plant communities in communal lands comparing to conservation areas. We found that plant communities in the park and buffer zone displayed high diversity for the woody layer while the diversity was lower in the communal lands. In fact, the lower diversity of the woody species in the communal lands might be linked to the selective exploitation of woody species. During the land clearing for cultivation, woody species in the farmland are cut down. Only a few species with high economic value mainly Vitellaria paradoxa, Parkia biglobosa, Adansonia digitata, Bombax costatum, Tamarindus indica or fodder species such as Pterocarpus erinaceus, Afzelia africana, Khaya senegalensis and Stereospermum kunthianum are set aside by the farmers in the communal lands (Bonou, 2008) for household needs. Therefore, woody species are less diversified in the communal lands comparatively to the park and buffer zone. Regarding the herbaceous layer, results showed that diversity was intermediate within the three land uses. Nonetheless, communal land plant communities presented an uneven species distribution in contrast to the park and buffer zone. This suggests that evenness distribution of plant species communities is more sensitive to disturbance as showed by the significant correlation between plant communities evenness land use type, vegetation type and pastoral pressure. This might be explained by the fact that disturbance results in the emergence of new plant species after land abandonment. Communal lands plant communities are dominated by few pioneer species which behave like invasive species at the first stage of succession. In addition to land cultivation which results in invasion of pioneer species, grazing eliminates most of the preferred grazed species and favours the dominance of some invasive species such as Hyptis suaveolens (Aboh et al., 2008). Houessou et al: Assessment of plant community pattern and diversity along a land use gradient in W Biosphere Reserve Life Form of Plant Communities The study highlighted that there was a floristic change in plant community’s life forms along the disturbance gradient going from communal lands to the park. Overall, the percent abundance as well as the cover of the hemicryptophytes and therophytes changed significantly between the communal lands, buffer zone and park. Hemicryptophytes abundance as well as their cover were significantly higher in plant communities in the park compared to the buffer zone and communal lands. Conversely, therophytes abundance and cover decreased significantly from the park to the communal lands suggesting that disturbance may favour the occurrence of therophytes in detriment of hemicryptophytes. Considering the phanerophytes percent abundance, our results showed no significant difference among the three lands uses suggesting that disturbance did not affect phanerophytes abundance. This result was similar to those obtained by Banda et al. (2006) who found lower density of tree stands in national park in Tanzania comparatively to the open area where human disturbance occurred. Our findings did not show any difference between plant community composition in chamaephytes and cryptophytes composition along the gradient from the park to communal lands. Ultimately, we concluded that the life form composition of plant communities can be used as indicator for phytodiversity monitoring of mainly hemicrytophytes, therophytes and phanerophyes in our study area. Chorotypes of Plant Communities The chorotypes are regarded in plant ecological studies as important traits of vegetation, which described the phytogeographical affinity of the plant communities (White, 1983; 73 Adomou et al., 2006). Our study revealed that the Sudanian species proportion decreased significantly from the plant communities in the park to the communal lands, while the proportion of wide distribution species increased along the gradient. Previous studies also found an important proportion of wide distribution species in secondary vegetation (Adomou et al., 2006; Bangirinama et al., 2010). Species with continental distribution showed no significant difference between the land uses. Chorotypes composition of a plant community can be used as an indicator of disturbance. High occurrence of wide distribution species indicates a high level of degradation in the community while high occurrence of the Sudanian species indicates a relatively undisturbed community in the Sudanian region. Conclusion This study illustrates change in diversity and species compositions of plant communities along a land use gradient. Our results support intermediate disturbance hypothesis and highlight the relevant indicators of plant community attributes to monitor change occurring in plant communities due to human disturbance at species or habitat level. At species level we found that floristic composition change as expressed by indicators species of the plant communities could be monitored at local scale to detect early change in vegetation. At habitat level, species richness and Shannon diversity index are not relevant at least at local scale, for phytodiversity monitoring, although Pielou evenness could be successfully used. The study also documents life forms and chorotypes composition of the plant communities as relevant indicators to 74 West African Journal of Applied Ecology, vol. 27(2), 2019 be monitored by managers for phytodiversity conservation. Acknowledgments This research was funded by the SUN project (Sustainable Use of Natural vegetation in West Africa) (EU FP6 INCO-dev 031685) and International Foundation for Science through IFS Grantee D4762. We thank “W” Biosphere Reserve Managers for providing us with field facilities during data collection. We remain grateful to Bio Sibo for field assistance and Redmond Sweeny for linguistic corrections. References Aboh, B.A., Houinato, M., Oumorou, M. and Sinsin, B. (2008). Capacités envahissantes de deux espèces exotiques, Chromolaena odorata (Asteraceae) et Hyptis suaveolens (Lamiaceae), en relation avec l’exploitation des terres de la région de Bétécoucou (Bénin). Belgian Journal of Botany, 141: 113-128 Adomou, A.C., Sinsin, B. and van Der Maesen, L.J.G. (2006). Phytosociological and chorological approaches to phytogeography: a meso-scale study in Benin. Systematics and Geography of Plants, 76: 155-178. Akoègninou, A., van Der Burg, W.J. and van Der Maesen, L.J.G. (eds), (2006). Flore analytique du Bénin. Backhuys Publisher, Wageningen. Banda, T., Schwartz, M. and Caro, T. (2006). Woody vegetation structure and composition along a protection gradient in a miombo ecosystem of western Tanzania. Forest Ecology and Management, 230: 179185. Bangirinama, F., Bigendako, M.J., Lejoly, J., Noret, N., DE Nanniere, C. and Bogaert J. (2010). Les indicateurs de la dynamique post-culturale de la végétation des jachères dans la partie savane de la réserve naturelle forestière de Kigwena (Burundi). Plant Ecology and Evolution, 143: 138-147. Biswas, S.R. and Mallik, A.U. (2010). Disturbance effects on species diversity and functional diversity in riparian and upland plant communities. Ecology, 91: 28-35. Bonou, A. (2008). Estimation de la valeur économique des Produits Forestiers Non Ligneux (PFNL) d’origine végétale dans le village de Sampéto (commune de Banikoara). Msc Dissertation, University of Abomey-Calavi, Abomey-Calavi. CENAGREF, (2008). Rapport de dénombrement pédestre dans le Complexe Parc W Bénin-Edition 2008. MAEP/ ECOPAS, Kandi Benin. Clerici, N., Bodini, A., Eva, H., Gregoire, J.M., Dulieu, D. and Paolini, C. (2007). Increased isolation of two Biosphere Reserves and surrounding protected areas (WAP ecological complex, West Africa). Journal for Nature Conservation, 15: 26-40. Dagnelie, P. (2011). Statistique théorique et appliquée: Inférence statistique à une et à deux dimensions. Tome 2. De Boeck, Belgium. Dufrêne, M. and Legendre, P. (1997). Species assemblages and indicator species: the need for a flexible asymmetrical approach. Ecological Monographs, 67: 345-366. Flamenco-Sandoval, A., Ramos, M.M. and Masera, O.R. (2007). Assessing implications of land-use and land-cover change dynamics for conservation of a highly diverse tropical rain forest. Biological Conservation, 138: 131-145. Houessou et al: Assessment of plant community pattern and diversity along a land use gradient in W Biosphere Reserve Foumier, A., Floret, C. and Gnahoua, G.M. (2001). Végétation des jachères et succession post-culturale en Afrique tropicale. In : Floret C, Pontanier R (eds) La jachère en Afrique tropicale. John Libbey Eurotext, Paris. Haarmeyer, D.H., SchmiedeL, U., Dengler, J. and Bösing, B.M. (2010). How does grazing intensity affect different vegetation types in arid Succulent Karoo, South Africa? Implications for conservation management. Biological Conservation, 143: 588-596. Hendricks, H.H., Bond, W.J., Midgley, J.J. and Novellie, P.A. (2005). Plant species richness and composition a long livestock grazing intensity gradients in a Namaqualand (South Africa) protected area. Plant Ecology, 176: 19-33. INSAE, (2013). Quatrième recensement général de la population et de l’habitation 2012, RGPH4. Cotonou. Kassi N’Dja, J.K. and Decocq, G. (2008). Successional patterns of plant species and community diversity in a semi-deciduous tropical forest under shifting cultivation. Journal of Vegetation Science, 19: 809-820. Koulibaly, A., Goetze, D., Traore, D. and Porembski, S. (2006). Protected versus exploited savannas: characteristics of the Sudanian vegetation in Ivory Coast. Candollea, 61: 425-452. Kruskal, J.B. (1964). Nonmetric multidimensional scaling: a numerical method. Psychometrika, 29: 115-129. Lezama, F., Baeza, S., Altesor, A., Cesa, A., Chaneton, E.J. and Paruelo, J.M. (2014). Variation of grazing-induced vegetation changes across a large-scale productivity gradient. Journal of Vegetation Science, 25: 8-21. Lindenmayer, D., Thorn, S. and Banks, S. (2017). Please do not disturb ecosystems 75 further. Nature Ecology & Evolution, 1: 1-3. Liu, B., Zhao, W., Wen, Z., Teng, J. and Li, X. (2009). Floristic characteristics and biodiversity patterns in the Baishuijiang River Basin, China. Environmental Management, 44: 73-83. Magurran, A.E. (2004). Measuring biological diversity. Blackwell publishing, London. McCune, B., & Grace, J.B. (2002). Analysis of ecological communities. MjM Software Design, Gleneden Beach, Oregeon. McCune, B. and Mefford, M.J. (2006). PCORD, Multivariate analysis of ecological data, Version 5.3.1. Software, Gleneden Beach. Mwaura, F. and Kaburu, H.M. (2009). Spatial variability in woody species richness along altitudinal gradient in a lowlanddryland site, Lokapel Turakan, Kenya. Biodiversity and Conservation, 18: 19-32. Nacoulma, B.M.I., Schumann, K., Traoré, S., Bernhardt-Römermann, M., Hahn, K., Wittig, R., and Thiombiano, A. (2011). Impacts of land-use on West African savanna vegetation: a comparison between protected and communal area in Burkina Faso. Biodiversity and Conservation, 20: 3341-3362. O’Connor, T.G. (2005). Influence of land use on plant community composition and diversity in highland sourveld grassland in the southern Drakensburg, South Africa. Journal of Applied Ecology, 42: 975-988. O'Connor, T.G., Martindale, G., Morris, C.D., Short, A., Witkowski, E.D.T.F. and Scott-Shaw, R. (2011). Influence of Grazing Management on Plant Diversity of Highland Sourveld Grassland, KwaZulu-Natal, South Africa. Rangeland Ecology & Management, 64: 196-207. Padonou, E.A., Adomou, A.C., Bachmann, 76 West African Journal of Applied Ecology, vol. 27(2), 2019 Y., Lykke, A.M. and Sinsin, B. (2014). Vegetation characteristics of bowé in Benin (West Africa). Journal of Plant Sciences, 2; 250-255. Pielou, E.C. (1969). An introduction to mathematical ecology. Wiley, New York. Raunkiaer, C. (1934). The life forms of plants and statistical plant geography. Clarendon, Oxford. Shackleton, C.M. (2000). Comparison of plant diversity in protected and communal lands in the Bushbuckridge lowveld savanna, South Africa. Biological Conservation, 94: 273-285. Shannon, C.E. (1949). The mathematical theory of communication. In: Shannon CE, Weaver W (eds) The mathematical theory of communication. University of Illinois Press, Urbana. Sinsin, B. (1993). Phytosociologie, écologie, valeur pastorale, production et capacité de charge des pâturages du périmètre NikkiKalalé au Nord-Bénin. PhD Dissertation, Université Libre de Bruxelles, Bruxelles. Southworth, J., Zhu, L., Bunting, E., Ryan, S.J., Herrero, H., Waylen, P.R. and Hill, M.J. (2016). Changes in vegetation persistence across global savanna landscapes, 1982-2010. Journal of Land Use Science, 11: 7-32. Thompson, P.L., Rayfield, B. and Gonzalez, A. (2017). Loss of habitat and connectivity erodes species diversity, ecosystem functioning, and stability in metacommunity networks. Ecography, 40: 98-108. van Wilgen B.W., Govender, N. and Biggs, H.C. (2007). The contribution of fire research to fire management: a critical review of a long-term experiment in the Kruger National Park, South Africa. International Journal of Wildland Fire, 16: 519-530. Viennot, M. (1978). Notice explicative de la carte pédagogique de reconnaissance de la R.P. Bénin. Feuille de Kandi-Karimama. ORSTOM, Paris. Weber, H.E., Moravec, J. and Theurillat, J.P. (2000). International Code of Phytosociological Nomenclature. Journal of Vegetation Science, 11: 739-768. White, F. (1983). Vegetation of Africa: a descriptive memoir to accompany the UNESCO AETFAT UNSO vegetation map of Africa: UNESCO, Paris. Wilkinson, D.M. (1999). The disturbing history of intermediate disturbance. Oikos, 84: 145-147. Houessou et al: Assessment of plant community pattern and diversity along a land use gradient in W Biosphere Reserve 77 APPENDIX Description of clustered plant community in each land use Plant communities P1 = Combretum glutinosum & Loxodera ledermannii community Frequent species Cochlospermum tinctorium(61%) Tephrosia bracteolata(58%) Soil texture Significant indicators species Plant communities’ description Siltygravelly soil Detarium microcarpum Communities weakly met in the park on silty soil with sometimes presence of gravel. The vegetation was under tree savanna or woodland dominated by Isoberlinia doka and Burkea africana (IV= 64.1; P= 0.0082) Burkea africana(53%) Combretum glutinosum Annona senegalensis(38%) (IV=40.4; P=0.024) Pteleospsis suberosa(37%) Loxodera ledermannii Detarium microcarpum(36%) (IV=49.6; P=0.0166) Isoberlinia doka (36%) Andropogon gayanus(30%) P2 = Andropogon gayanus & Crossopteryx febrifuga community Andropogon gayanus(100%) Silty soil Andropogon gayanus Indigofera dendroides(80%) (IV = 62; P = 0.0002) Ampelocissus leonensis(80%) Crossopteryx febrifuga Combretum molle(80%) (IV = 48.3; P = 0.0064) Siphonochilus aethiopicus(80%) Andropogon schirensis Grewia cissoides(80%) (IV = 54.8; P = 0.0066) Communities under tree savanna vegetation with Vitellaria paradoxa, Isoberlinia doka and Aganope stuhlmannii Chamaecrista mimosoides(80%) P3= Hyparrhenia involucrata & Indigofera leprieurii community Hyparrhenia involucrata(73%) Siphonochilus aethiopicus(73%) Sandy-silty Hyparrhenia involucrata soil (IV=65.9; P=0.0002) Combretum glutinosum(67%) Indigofera leprieurii Chasmopodium caudatum(58%) (IV=41.5; p=0.0272) Combretum collinum(58%) Pennisetum polystachion Indigofera dendroides(58%) (IV=38.6; p=0.0431) Communities met on different type of soil. The vegetation was represented by shrub/tree savanna dominated by Combretum spp, on silt-sand soil and Acacia hockii on clay-silt soil Polygala arenaria(51%) P4 = Andropogon tectorum and Costus spectabilis community Pandiaka heudelotii (79%) Combretum molle (76%) Gardenia ternifolia (69%) Soil with silt and clay Andropogon tectorum (IV= 99.5; P =0.0002) Vigna gracilis Lannea acida (68%) (IV = 77.7; P = 0.0008) Strychnos spinosa (64%) Costus spectabilis Pterocarpus erinaceus (60%) (IV=62.3; P =0.0044) Isoberlinia doka (60%) Lannea acida Daniellia oliveri(60%) (IV=56.7; P=0.0028) Stereospermum kunthianum(60%) P5 = Loudetia togoensis & Bulbostylis abortiva community Loudetia togoensis(100%) Lannea microcarpa(75%) Silty soil on crust Loudetia togoensis (IV = 100; P = 0.0002) Spermacoce filifolia(50%) Bulbostylis abortiva Andropogon pseudapricus(50%) (IV = 97.8; P = 0.0002) Ophioglossum costatum(42%) Sporobolus festivus Ipomoea eriocarpa(45%) (IV = 65.1; P = 0.0034) Polygala arenaria(45%) Spermacoce filifolia Combretum spp(45%) (IV = 65.2; P = 0.0038) Vegetation on deep soil on upland represented by woodland forest dominated by Isoberlinia doka, Pterocarpus erinaceus and in the herbaceous layer by Andropogon tectorum, Beckeropsis uniseta. The community was also met under gallery forest with Lannea acida and Daniellia oliveri on the tree layer and Andropogon tectorum, Rottboellia cochinchinensis at herbaceous layer. Plant communities on crust lateritic soil (less deep soil) dominated on herbaceous layer by Loudetia togoensis and scattered by woody species such Lannea microcarpa and combretim spp 78 West African Journal of Applied Ecology, vol. 27(2), 2019 APPENDIX continued Description of clustered plant community in each land use C1 = Loudetia togoensis & Bulbostylis abortiva community Combretum glutinosum (78%) Lannea microcarpa (50%) Silty soil on crust (IV = 68.7; P = 0.002) Spermacoce filifolia (45%) Loudetia togoensis Ophioglossum costatum(42%) (IV = 75.7; P = 0.002) Loudetia togonensis (39%) Lannea microcarpa Ipomoea eriocarpa (37%) (IV = 42.2; P = 0.0078) Polygala arenaria (35%) Ophioglossum costatum (IV = 33.3; P =0.0084) Andropogon pseudapricus (25%) C2 = Piliostigma thonningii & Flueggea virosa community Bulbostylis abortiva Setaria pumila(72%) Annona senegalensis(69%) Piliostigma thoningii(57%) Vitellaria paradoxa(53%) Hibiscus asper(50%) Silty soil sometimes with relative dominance of clay Dichrostachys cinerea (IV = 83.4; P= 0.0002) Flueggea virosa (IV = 75.9; P= 0.0002) Piliostigma thonningii (IV = 95.1; P= 0.0002) Pennisetum polystachion(48%) Flueggea virosa(46%) Vegetation on less deep soil (bowé in French). Tree layer was almost absent. The herbaceous strata height was about 30 cm. Due to soil condition, that plant communities was not used for cultivation. Nonetheless it was used pasture for cattle grazing Rare vegetation in communal land, represented by old fallow (5 to 10 years). The tree layer was almost absent and the shrub layer was about 5 m and dominated by Piliostigma thoningii and Dichrostachys cinerea Dichrostachys cinerea(43%) C3 = Digitaria horizontalis & Spermacoce stachydea community Setaria pumila(78%) Commelina benghalensis(76%) Indigofera hirsuta(69%) Sandy soil sometimes silty-sandy soil Detarium microcrapum(47%) Digitaria horizontalis (IV = 69.7; P =0.0064) Schizachirium exile (IV = 49.1 ; P = 0.0204) Ageratum conyzoides(45%) Spermacoce stachydea Leucas martinicensis(35%) (IV = 35.4; P = 0.0342) Crotalaria retusa(34%) Mitracarpus hirtus(25%) Young fallow within farmland. The herbaceous layer was dominated by Setaria pumila, Digitaria horizontalis. The tree layer resulted from those set apart by farmers and the shrub layer after three years was abundant and grew from coppices Celosia trigyna(25%) C4 = Tephrosia pedicellata & Detarium microcarpum community Tephrosia pedicellata(90%) Spermacoce stachydea(85%) Pandiaka heudelotii(85%) Silty soil and sometimes gravel (IV = 38.1; P = 0.0077) Detarium microcarpum Hackelochloa granulari (70%) (IV = 64.5; P = 0.001) Detarium microcarpum(72%) Triumfetta rhomboidea Hyparrhenia involucrata(65%) (IV = 76.1; P= 0.0024) Brachiraria deflexa(60%) Brachiaria deflexa Aspilia kotschyi(80%) Prosopis africana(85%) Burkea africana(75%) Pennisetum polystachion(70%) Combretum nigricans(70%) Crossopteryx febrifuga(60%) Common vegetation in communal lands derived from overgrazing. Tree layer was about 5 à 8 m dominated by Combretum spp., Terminalia spp. and Detarium microcarpum (IV = 62; P= 0.068) Digitaria horizontalis(50%) B1 = Burkea africana & Indigofera bracteolata community Tephrosia pedicellata Gravelly with silts. Often, presence of block of stone Burkea africana (IV = 63.8; P= 0.0066) Combretum glutinosum (IV = 50.9; P= 0.009) Indigofera bracteolata (IV =71.4; P= 0.0108) Common vegetation on soil with outcrop stone. The tree layer was dominated by species such Prosopis africana, Burkea africana. The herbaceous layer was limited in cover Microchloa indica(50%) Bombax costatum(50%) Cissus populnea(40%) Legend: Note that based on indicator species analysis B2 = P2; B3 = P4 and B4 = P5 P1, P2, P3, P4, P5 = Plant communities clustered in the park, C1, C2, C3, C4 = Plant communities clustered in the communal. B1, B2, B3, B4 = Plant communities clustered in the buffer zone