Ecological assessment of riparian forests in Benin: Phytodiversity, phytosociology, and spatial distribution of tree species Armand Kuyéma NATTA Promotor: Prof. Dr. Ir. L.J.G. van der Maesen Hoogleraar in de Plantentaxonomie, Wageningen Universiteit, Nederland Prof. Dr. Ir. B. Sinsin Professor of Ecology, Department of Environment Management, Faculty of Agronomic Sciences, University of Abomey-Calavi, Benin. Co- promotor: Dr. A. Akoègninou Department of Vegetal Biology, Faculty of Science and Technique, University of Abomey-Calavi, Benin. Promotiecommissie: Prof. Dr. Ir. A. de Gier (ITC – Enschede, Nederland) Prof. Dr. F.J.J.M. Bongers (Wageningen Universiteit, Nederland) Prof. Dr. J. Lejoly (Free University of Brussels, België) Prof. Dr. S. Porembski (University of Rostock, Germany) Ecological assessment of riparian forests in Benin: Phytodiversity, phytosociology, and spatial distribution of tree species Armand Kuyéma NATTA Proefschrift ter verkrijging van de graad van doctor op gezag van de rector magnificus van Wageningen Universiteit Prof. Dr.Ir. L. Speelman in het openbaar te verdedigen op woensdag 26 november 2003 des namiddags te vier uur in de Aula NATTA Armand Kuyéma (2003). Ecological assessment of riparian forests in Benin: Phytodiversity, phytosociology, and spatial distribution of tree species Ph.D. Thesis Wageningen University, with summaries in English, French and Dutch. ISBN 90-5808-954-1 Key words: Riparian forests, flora, diversity, endangered species, structure, plant community, ordination, classification, rivers, streams, spatial distribution, ecological factors, sampling designs, Benin, West Africa. To My mother, Gwendoline, Kévin. CONTENTS Chapter 1 General introduction 1 Chapter 2 Study area 7 Chapter 3 Riparian forests, a unique but endangered ecosystem in Benin 15 Chapter 4 Assessment of riparian forest fragments plant diversity in West African savanna regions: an overview from Benin 27 Chapter 5 Structure and ecological spectra of riparian forests in Benin 41 Chapter 6 A phytosociological study of riparian forests in Benin (West Africa) 53 Chapter 7 Spatial distribution and ecological factors determining the occurrence of Pentadesma butyracea Sabine (Clusiaceae) in Benin (West Africa) 73 Forest structural parameters and floristic composition spatial variation and modelling across rivers in Benin 83 Ouémé and Comoé: forest-savanna border relationships in two riparian ecosystems in West Africa 105 Chapter 8 Chapter 9 Chapter 10 Assessing the density of Khaya species through Simple Random Sampling, Stratified Sampling and Systematic Sampling 117 Chapter 11 General discussion 129 Chapter 12 General conclusion 143 Summary 149 Résumés 155 Samenvatting 161 References 167 Riparian forests plant species list 193 Acknowledgements 211 Curriculum vitae 213 List of publications 215 Chapter 1 GENERAL INTRODUCTION A.K. Natta Chapter 1: General Introduction Chapter 1 GENERAL INTRODUCTION 1.1. RESEARCH BACKGROUND The loss and fragmentation of tropical forest is the single greatest threat to the world’s biological diversity (Whitmore 1990, Huston 1994). In 1992, the Convention on Biological Diversity highlighted that measures must be implemented for the conservation of natural ecosystems, especially for tropical forests, which are famous for being the most species rich ecosystems on earth. Although many species have been described, very little is known about their ecology (Sayer & Wegge 1992). In Africa the most species-rich forests, at least for woody plants, are in the wetter areas of West Central Africa (from the base of Mt Cameroon South East into Gabon), and in the North East of Madagascar (Gentry 1992). Unfortunately this diversity is characterised by a deforestation rate of 0.7 %, more than twice the world average of 0.3 % (FAO 2000). In West and Central Africa, the rain forest block is interrupted between South East Ghana and South West Nigeria. Biogeographers call this discontinuity the Dahomey-Gap. Innumerable animals and plants are common to Central Africa and upper Guinea, showing that the two forests blocks have been connected in the past. However, this does not mean that the forests are identical. The two rain forests blocks are characterised by special endemic species to confirm that they have not only been isolated but are intrinsically different (Kingdon 1990). On the other hand, the West-African countries located in the Dahomey-Gap, such as Benin, are characterised by the absence of tropical rain forests and a low percentage of dense semi-deciduous forests. In consequence the need to conserve the biological diversity of remaining fragmented forests, represented by sacred forests, few protected dense semideciduous upland forests and riparian forests, in this dry wedge cannot be underestimated. All over the world, and particularly in tropical savanna, the natural vegetation associated with waterways and mostly represented by riparian forests (or RFs) is credited to be among the most species-rich ecosystems (Meave & Kellman 1994, PGRN/IUCN 1994, Roggeri 1995, Nilsson et al. 1997). Riparian forests are important areas for global biodiversity (Sala et al. 2000), because they protect key resources for mankind, such as water sources and quality, and stream environment (Vought et al. 1994, Lowrance et al. 1997, Montgomery 1997, Trimble 1999), and harbour a diversified flora and physical structure (Gibbs & Leitão 1978, Gregory et al. 1991, Bersier & Meyer 1994, Kellman et al. 1994, Meave & Kellman 1994, Woinarski et al. 2000, Kokou et al. 2002). Ecologists have also examined the value of riparian forests as habitats for many animals, and recognised them as a priority area for conservation of terrestrial mammals (Doyle 1990, Darveau et al. 1998, de Lima & Gascon 1999, Darveau et al. 2001), as well as birdlife (Stauffer & Best 1980, Gates & Giffen 1991, McGarigal & McComb 1992, Larue et al. 1995, Murray & Stauffer 1995, Darveau et al. 1995, Whitaker & Montevecchi 1997, Saab 1999, Woinarski et al. 2000). As tropical forests become more fragmented due to deforestation, riparian forests play a crucial role in providing habitat corridors between forest patches to increase landscape connectivity (Forman & Godron 1986, Forman 1997, Machtans et al. 1996). In proportion to their area within a watershed, they perform more ecological and productive functions than do adjacent uplands (NRC 2002). The vital ecological, hydrological and biogeochemical vital functions of these forests bordering waterways are now largely acknowledged. In recognition of these roles, they have been increasingly protected by policies and legislation (DNRE 1996, de Lima & Gascon 1999, NRC 2002, Natta et al. 2002). Unfortunately, they are under severe 1 Chapter 1: General Introduction threat worldwide (Sparovek et al. 2002), and current management strategies, particularly in the tropics, seem to have limited effects. A large body of work has demonstrated that riparian forests play a critical role regulating interactions between terrestrial and aquatic components of temperate zones landscapes (Gregory et al. 1991, Gilliam 1994, Naiman & Décamps 1997), however there have been relatively few studies dedicated to them in the tropics (Bowden et al. 1992, McDowell et al. 1992, McClain et al. 1994, Chestnut & McDowell 2000, Groffman et al. 2001) in general, and Benin in particular. Flora and fauna reserves represent 11% of Benin surface area, and the flora is estimated to have 3,000 species (PFB 1997), which is relatively poor in comparison with neighbouring humid countries. The natural wooded vegetation represented by savanna and open forests occupy 72.2 % of the country, but dense semi-deciduous forests and riparian forests cover only 3.42 % of the country (CENATEL 2001). During the preparation of Benin’s Environmental Management Plan, it was noted that the lack of scientific knowledge of the country’s environment constituted a major obstacle to the realisation of the plan and measurement of the effectiveness of the proposed actions. In 1994, following the spirit of the Rio conference, the sustainable development agreement between Benin and the Netherlands emphasised on the knowledge and valorisation of the biological diversity. In 1995 this agreement has retained, as emergency programs, scientific surveys of plant species and the compilation of Benin’s flora. The flora is known to be a major component of a region’s or a country’s biodiversity. Scientific research on species diversity is of great importance in providing accurate information and can enhance the management and sustainable use of phytogenetic resources, especially for endangered ecosystems (PFB 1997), such as riparian forests, which are among the most important forest ecosystems in the majority of degraded woodlands and mosaic of savanna landscapes (Natta 2000). In Benin the rapid changes in land use have led to the progressive destruction and fragmentation of riparian forests, which provide fertile soil for cultivation, give an opportunity for irrigation, and shelter a wide range of valuable and scarce plants and animals. Therefore, they are systematically targeted for illegal selective tree cutting, hunting and conversion to agriculture. Also, when the riparian forests degrade, forest species become sparse, leaving the vegetation open for the savanna species to invade. Their rich biological resources, especially plant species, are disappearing before they can be inventoried and assessed (Natta 2000). We see that riparian forests can be classified among the endangered ecosystems of Benin because they are marginal among the wooded vegetation and because of their advanced state of degradation. In spite of the fact that there is growing recognition of the ecological, hydrologic, biogeochemical, socio-cultural, economic and aesthetic importance of RFs in Benin, they have remained insufficiently studied (Mondjannagni 1969, Paradis 1988, Sokpon et al. 2001) and managed as key vegetation formation for biodiversity protection (Natta et al. 2002). Riparian forests have often been ignored, or excluded from vegetation studies in favour to upland forests. RFs flora is poorly known, and no published account concerning the diversity, ecology and spatial distribution of species in riparian forests is available in Benin. We often don’t even know which species are present, which one are abundant, rare and what are the underlying ecological factors that govern their presence along waterways. 1.2. OBJECTIVES AND BIODIVERSITY IN BENIN APPROACH FOR THE ASSESSMENT OF RIPARIAN FORESTS The overall objective of this research is to contribute to a better knowledge of the flora, diversity and ecology of riparian forests in Benin. The specific objectives are to (a) compile a 2 Chapter 1: General Introduction preliminary riparian forests plant species list, (b) assess plant species and ecosystem diversities, (c) investigate plant communities, (d) clarify the structural and floristic relationship of riparian forests with adjacent plant communities, and (e) assess the ecology of certain endangered tree species in riparian forests. ‘Biological diversity’, in short ‘Biodiversity’ is defined as “the variability among organisms from all sources including, inter alia, terrestrial, marine and other aquatic ecosystems and the ecological complexes of which they are part. This includes diversity within species, between species and of ecosystems”. Also Biological resources include genetic resources, organisms or parts thereof, populations of species or any other component of ecosystems with actual or potential use or value for humanity (UNCEB 1992). In this thesis, the term ‘Riparian Forests (or RFs)’ does not refer exclusively to dense vegetation dominated by woody species in which grasses are virtually absent (i.e. moist evergreen or dense semi-deciduous tropical forests). Instead, it stands for any topographic (lowest parts of a landscape occupied by waterways) and edaphic (moist soil dependent) woodland at stream/river edges, which is made of a mosaic of plant communities, dominated by woody species in open or closed canopy. Under Benin conditions, open canopies and degraded woodlands at streamside are dominant. Also, light coming from aside allows many heliophytic and pioneering species to grow in riparian forests. The compilation of the RF plant species list is made through the collection of flowering or fruiting plant growing inside RFs. Field observations coupled with empirical knowledge about the ecology of each RF species allow to find out which species are most frequent, rare, valuable, endangered, endemic, or show a specific adaptability to riparian habitat. Plant species diversity (or phytodiversity) is assessed using several parameters: species richness and abundance, diversity index (Shannon), Equitability index (Pielou), species abundance models, differences in site diversities (comparison of Shannon index), basal area, stem density, life forms and geographical affinity. The diversity of plant communities was assessed through the Braun-Blanquet phytosociological approach combined with multivariate analysis of floristic data. 1.3. ORGANISATION OF THE THESIS Chapter 2 introduces the study area that covers about 70% of the total area of Benin. Chapters 3 to 10 are self-contained and cover introduction, methods used for data collection and analysis, results, discussion and conclusions. Chapter 3 presents an overview of riparian forests in Benin, as unique but endangered ecosystem. Chapter 4 and 5 highlight the floristic diversity and variability of structural parameters within riparian forests. Chapter 6 assesses plant communities’ diversity using the Braun-Blanquet phytosociological approach, and ordination and classification of floristic relevés. Chapter 7 deals with the ecology of Pentadesma butyracea Sabine (Clusiaceae), a tree species rare in Benin. Chapter 8 and 9 focus on the variability of species composition and structural parameters across riparian forests. In chapter 10, sampling methods are compared to estimate the density of two Khaya species. A general discussion and conclusions are presented in chapters 11 and 12, respectively. 3 Chapter 2 STUDY AREA A.K. Natta Chapter 2: Study Area Chapter 2 STUDY AREA 2.1. LOCATION OF THE STUDY AREA IN BENIN Benin is located at the so-called Dahomey Gap, in which Sudan-type savanna vegetation extends as far as the sea, through a hiatus in the West African rain forest over some 200 km from South East Ghana to South East Benin (Guillaumet 1967, Schnell 1976, Onochie 1979, Whitmore 1990, Martin 1991, Maley 1996). In the entire Dahomey Gap, the present day pattern of vegetation is very much obscured by the omnipresent impact of dense human settlements and ever developing agriculture and fishing (Jenik 1994, Kokou & Caballé 2000). Nevertheless forest islands and riparian forests stretching along the rivers contain many humid forest elements (Adjanohoun 1968, Kokou 1998, Kokou et al. 1999, Sokpon et al. 2001). The conservation of forest patches and forest species is this dry wedge is of great ecological and economic importance (Natta et al. 2002). Forest patches of varying size, including riparian forests, and large trees of Antiaris toxicaria, Diospyros mespiliformis, Milicia excelsa, on firm ground indicate the potential climax of the semi-deciduous closedcanopy forest (Hall & Swaine 1981, Sokpon 1995). Figures 2.1 and 2.2 present Benin republic in West Africa and in the Gulf of Guinea, respectively. An extensive preliminary field reconnaissance of riparian forests in all the ecological regions of Benin was made. It appears that South of 7º N (i.e. Littoral, Atlantic, Ouémé, South-Plateau, Mono, and South-Couffo provinces) riparian forests are very degraded and sites at least 1 ha large (i.e. in which we can installed > 20 plots each of 500 m2) are uncommon. Therefore to fit with the present-day distribution pattern of riparian forests, and for practical reasons, several large and least degraded sites North of 7º N were selected for intensive data collection. This study area stretches from 7º to 12º20 N, and covers about 70% of the total Benin area. It is subdivided into 3 ecological zones: 1 - The Guinean region: from 7º N (South of Abomey latitude) to about 8º N (at Savè latitude); 2 - The Sudano-Guinean zone: from about 8º N to about 9º10 N (at PénéssoulouBétérou latitude); 3 - The Sudanian region: from about 9º10 N to the extreme North of the country (12º20 N at Karimama latitude). In comparison with the Sudano-Guinean and Sudanian regions, the vegetation in the Guinean region has been surveyed many more times: - Coastal zone (Adjanohoun 1966, 1968, de Souza 1979, Paradis 1975a, 1976, 1979, 1981, 1983, FAO 1980, Akoègninou 1984, Adjanohoun et al. 1989, Djego 2000, etc.); - The complex of lakes, lagoons and inundation valleys (Paradis & Adjanohoun 1974, Paradis 1975a, 1975b, 1979, 1980, 1983, FAO 1980, Paradis & Rabier 1979, Texier et al. 1980, Adjanohoun et al. 1989, Guinko 1974, Essou 1991, Sinadouwirou 1997, etc.); - The Niaouli forest (Mondjannagni 1969, Akoègninou 1984, Hountondji 1998, etc.); - The Lama forest (Paradis & Houngnon 1977, FAO 1980, Adjanohoun et al. 1989, Agbani 2002, etc.); - The Pobè forest (Akoègninou 1984, Adjanohoun et al. 1989, Sokpon 1995, etc.); - The Ouémé, Dogo and Kétou regions (apart from the Pobè forest) (Paradis et al. 1978, FAO 1980, Paradis 1983, Akoègninou 1984, Adjakidjè 1984, Adjanohoun et al. 1989, Bossou 2001, etc.); - The plateau of Allada (Paradis 1983, Adjakidjè 1984, Adjanohoun et al. 1989, etc.); 7 Chapter 2: Study Area - The Kouffo zone (FAO 1980, Paradis 1983, Akoègninou 1984, Adjakidjè 1984, Adjanohoun et al. 1989, Amétépé 1997, etc.). Figure 2.1: Benin in West Africa Figure 2.2: Benin with its rivers 8 Chapter 2: Study Area Until recently, riparian forests in the three zones of the study area have rarely been studied as compared to other vegetation types. This study is concerned with edaphic and hygrophile forests along fresh water bodies. Therefore the study area does not include the coastal complex, which consists of several littoral cordons, marshy shoals, mangroves, lagoons (e.g. Porto-Novo and Ouidah), and lakes (e.g. the Nokoué, Ahémé, Toho, Togbadji, Azili, Sélé, etc.). Likewise, the ‘barre’ region that adjoins the coastal plain through an uneven embankment where there is a depression referred to as the ‘Lama forest’ is excluded from the study area. 2.2. GEOMORPHOLOGY Geologically, the study area can be divided in three parts: the Northern plateau of South Benin, the crystalline plateau in the Centre and North, and the Niger River basin in the extreme North. The Southern part of the study area corresponds to the Northern plateau (Aplahoué, Abomey and Kétou) of Southern Benin, which corresponds to the West African continental terminal. It reaches 100-150 m in altitude, with red sandy argillaceous soil and lateritic crusts (Paradis 1983). The major part of the study area, from the North of Bohicon to Kandi, is occupied by the crystalline plateau, which has tropical ferruginous soils. Within this vast peneplain there are many residual inselberg landscapes (300 - 400 m high with convex sides) including the hills of Dassa-Zoumé, Savalou, Savè, Yaoui, Ouari-Maro, Bembèrèkè and Sinendé. The North West part corresponds to the Atacora region dominated by a series of mountain chains that reaches an elevation of 600 m. The Northern part of the study area corresponds to the Niger River basin, which ranges from the Goungoun village latitude to the Niger River. It is a sandy plateau with an average altitude of 200 to 250 m. Sedimentary formations predominate, with here and there cuirass hills of about 270 m high. There are also important valleys of variable lengths with alluvial soils along the Sota, Alibori and Niger Rivers. 2.3. CLIMATE Benin is a hot tropical country under two influences: the humid maritime wind that blows from April to November in South-North direction and the dry continental trade wind, the Harmattan, that blows from the Sahara desert in the North-South direction. The Harmattan is the main cause of drought. The mean temperatures are constantly high (25° C) with daily amplitude below 5° C in the South and 10° C in the North. The largest temperature variations occur during the days of Harmattan. Nowadays, the country is characterised by a great irregularity of annual rainfall from one year to another and within each year. In the SouthNorth direction annual rainfall decreases (1300 to 800 mm) while evapotranspiration and temperatures increase. The climate is of two types (Guinean in the South and Sudanian in the North) separated by a transitional Sudano-Guinean zone (from about 8° to 9° N). The South of the study area, (7° to 8° N), has a sub-equatorial (or tropical humid) climate subdivided into four seasons of unequal length: two rainy seasons (from April to July, and September to October), and two dry seasons (from November to March, and end of July to August). The North, from about 9° to 12° N, is the domain of Sudanian (i.e. dry tropical) climate with two seasons: a rainy season from May to September-October, and a dry season from October to April. The annual rainfall varies from 1100 to 800 mm. The rainy season is shortened as latitude increases toward the Niger River basin in the extreme North. In the 9 Chapter 2: Study Area Sudanian region higher rainfalls are linked to relief (e.g. Atacora region, Bembèrèkè and Nikki with 1200 to 1300 mm). The extreme North (Malanville - Karimama) is the domain of the North-Sudanian climate, with an average annual rainfall of 800 mm, and up to 2000 mm of evapotranspiration, aggravated by the Harmattan winds. Between the Guinean and Sudanian regions (about 8° to 9° N), there is a gradual climatic change characterised by the disappearance of the short rainy season and the fusion of the two peaks of rainfall that are typical for the sub-equatorial climate of Southern Benin. Annual rainfall varies from 1100 to 1300 mm and evapotranspiration between 1400 to 1500 mm. 2.4. HYDROGRAPHY The study area includes three hydrographic basins: 1) the medium and upper parts of the Ouémé basin in South and Central Benin; 2) the Niger River and its tributaries (i.e. the Mékrou, Alibori and Sota) in North Benin; and 3) the upper part of the Pendjari basin in North West Benin. The river network is presented in Figure 2.2. Every year the rivers are at their highest level and cause inundation from August to December in the South, and August to November in the Centre and North; while streams are characterised by daily submersion following storms or heavy rains. Streams in Benin are generally temporary; meanwhile some have a permanent regime linked to the relief (e.g. headwaters in the Atacora mountains region). In valleys, depending on the topography, and volume of water in rivers and streams, inundation and emersion alternate. Along the Pendjari and Niger rivers there are vast flood plains. 2.5. HUMAN INFLUENCE From 1979 to 2002, the population of Benin has doubled from 3.3 to 6.7 millions of inhabitants. The average annual population growth is 3.5 %. The South of the country has the highest population density with 150 inhabitants per km2, while the Centre and the North have only 20 inhabitants per km2 (INSAE 1979, 2002). Farming, tree cutting, as well as a high incidence of fire have reduced the original woodland to their current state. In recent years the deterioration of vegetation formations within Benin has increased at an alarming rate (CENATEL 2000). Each year about 100,000 ha of natural vegetation are degraded and the tendency is not decreasing (CENATEL 1995). In the whole country, farming on marginal lands, deforestation of riparian forests and savanna woodlands for cotton and yam cultivation, excessive pruning of valuable trees and uncontrolled bush fires are on the increase. In the South, the present-day pattern of vegetation is very much obscured by the omnipresent impact of dense human populations. There are many small relict groves (i.e. sacred forests) and remnant semi-deciduous forests (e.g. Lama, Pobè, Itchédé, Dangbo, etc.) of varying size scattered all over the area (Sokpon 1995). In Central Benin, original forests have been cleared and replaced by a mosaic of savanna and dry forests, while the North with a much drier climate, is unfavourable for forest establishment, and is therefore dominated by Sudanian savanna. The process of natural resources degradation leads to the impoverishment of biodiversity due to the loss of wooded vegetation and their replacement by shrub savanna, which is known to have much less ecological potential. 10 Chapter 2: Study Area 2.6. VEGETATIONS As for the whole country, the study area vegetation is characterised by a great variety and a fragmentation of phytocenoses caused, on the one hand, by climatic, topographic and edaphic factors and, on the other hand, by human influence on the environment. The South of the study area prolongs the derived savannah (Jones 1945) and Southern Guinean zone of Nigeria (Keay 1959). It is the domain of: - Humid semi-deciduous forests with Antiaris toxicaria, Strombosia glaucescens, Triplochiton scleroxylon and Terminalia superba (Northern part of the Guinean sub-zone IIA of Adjanohoun et al. 1989); - Dry semi-deciduous forests and derived savanna with Anogeissus leiocarpus, Afzelia africana and Lonchocarpus sericeus (Northern part of the Guinean sub-zone IIB & C of Adjanohoun et al. 1989); - Dry forests with Anogeissus leiocarpus and Daniellia oliveri (Guineo-Sudanian transitional sub-zone IIIA of Adjanohoun et al. 1989); - Woodland savanna (with Ceiba pentandra, Milicia excelsa and Daniellia oliveri), and tree and shrub savanna (with Daniellia oliveri, Elaeis guineensis and Lophira lanceolata) (FAO 1980); - Savanna on hills (with Afrotrilepis pilosa, Ficus abutifolia, Hildegardia barteri and Aeollanthus pubescens) (Yédomonhan 2002). - Tree and shrub savanna (with Anogeissus leiocarpus, Vitellaria paradoxa, Daniellia oliveri, Isoberlinia doka and Parkia biglobosa) with high incidence of cultivation (FAO 1980). The Centre of the study area is a prolongation of the Northern Guinean zone of Nigeria (Keay 1960) and a small part of the dry semi-deciduous forest, the fire sub-type of Hall & Swaine (1981). It is the domain of: - Patches of semi-deciduous forest with a few big stems of semi-deciduous trees (e.g. Antiaris toxicaria, Milicia excelsa, Cola spp., Khaya senegalensis and Celtis spp.) in the Bantè, Ouèssè and Bassila districts (FAO 1980). Particularly, a facies of dry semi-deciduous forest is seen in the Sudano-Guinean region of Bassila with Khaya grandifoliola, K. senegalensis, Albizia spp., Cola gigantea, Antiaria toxicaria, Milicia excelsa, Erytrophleum guineensis, Anogeissus leiocarpus, etc. (variant of sub-zone IIB of Adjanohoun et al. 1989); - Open forests with Isoberlinia spp., Pterocarpus erinaceus, Khaya senegalensis and Monotes kerstingii (Centre and North of the sub-zone IIIB of Adjanohoun et al. 1989); - Woodland, tree and shrub savanna (with Anogeissus leiocarpus, Vitellaria paradoxa, Daniellia oliveri, Isoberlinia doka and Parkia biglobosa) (FAO 1980). - Tree and shrub savanna with high incidence of cultivation (FAO 1980). - Savanna on hills South of Alafiarou, and in the Kouffé Mountains reserve forest (FAO 1980). The North of the study area is a prolongation of the Sudanian zone of Nigeria (Keay 1949) and the plains of Northern Togo (Ern 1979). The Sudanian region (or dry continental zone, FAO 1980) is the domain of typical savannas, or dry forests on deep or rich soils: - Open forests with Isoberlinia spp., Pterocarpus erinaceus, Khaya senegalensis and Monotes kerstingii (Northern part of the sub-zone IIIB of Adjanohoun et al. 1989); - Sudanian savannas with Afzelia africana, Burkea africana, Anogeissus leiocarpus, Parkia bigblobosa, Vitellaria paradoxa, Terminalia spp., Daniellia oliveri, Combretum spp., Crossopteryx febrifuga, Acacia spp., Balanites aegyptiaca, Sclerocarya birrea, etc. and numerous grasses (sub-zone IVA and IVB of Adjanohoun et al. 1989, FAO 1980); The subzone (IVB) includes the Atacora mountains chain; - Savannas and dry forests with Combretaceae of the Gourma (or Pendjari) plain in North West Benin, along the Togo and Burkina-Faso borders (sub-zone IVC for Adjanohoun et al. 11 Chapter 2: Study Area 1989). In the West of the Pendjari complex, there is a facies of tree and shrub savanna with Anogeissus leiocarpus, Combretum spp., Acacia spp., Balanites aegyptiaca and Ziziphus mauritiana (FAO 1980); - Tree and shrub savanna with high incidence of cultivation (FAO 1980). The extreme North of the study area (very dry continental zone with 6-7 dry months, FAO 1980) is dominated by: - Woodland, tree and shrub savanna (with Anogeissus leiocarpus, Combretum spp., Acacia spp., Balanites aegyptiaca and Ziziphus mucronata). - Tree and shrub savanna with high incidence of cultivation. - Saxicolous tree and shrub savanna (with Adansonia digitata, Combretum spp. and Acacia spp.). - Shrub savanna and prairie (with Echinochloa stagnina, Vetiveria nigritana, etc.) in the Niger River inundation plain. The annual crops (maize, cassava, yam, millet, sorghum, cotton), the perennial plantations (oil-palm from the last two centuries onwards in the South), the gathering of firewood, production of charcoal, selective cutting of valuable trees and uncontrolled bush fire, have brought about the destruction of forests and woodlands and replaced these by tickets and grassy phytocenoses. The result is a mosaic of savanna, bush fallow and cultivated areas in which remnant forest and savanna species are frequently seen. Meanwhile, the important hydrographic network still allows the presence of many riparian forests, widely distributed all over the country. 2.7. SITES OF DATA COLLECTION Representative sites of the largest and least degraded riparian forests were chosen along several rivers and streams in the Guinean region (Samiondji and Bétécoucou), the SudanoGuinean zone (Toui-Kilibo, Idadjo, and Pénéssoulou) and the Sudanian region (Bétérou, Onklou, Daringa, Ouaké, Affon, Yarpao, Boukombé, Toukountouna, Pendjari Biosphere Reserve, Kandi, Gbèssè and Malanville; see Figure 4.1). 12 Chapter 3 RIPARIAN FORESTS, A UNIQUE BUT ENDANGERED ECOSYSTEM IN BENIN Published as: NATTA A.K., SINSIN B. & VAN DER MAESEN L.J.G.: Riparian forests, a unique but endangered ecosystem in Benin. Notulae Florae Beninensis 4. – Bot. Jahrb. Syst. 124: 55-69. 2002. ISSN 0006-8152. Chapter 3: Riparian forests, a unique but endangered ecosystem in Benin Chapter 3 RIPARIAN FORESTS, A UNIQUE BUT ENDANGERED ECOSYSTEM IN BENIN Notulae Florae Beninensis 4 NATTA A.K., SINSIN B. & VAN DER MAESEN L.J.G.: Riparian forests, a unique but endangered ecosystem in Benin. – Bot. Jahrb. Syst. 124: 55-69. 2002. ISSN 0006-8152. ABSTRACT Riparian forests are often small in area, but are of extreme ecological and economic value for local people. The interest of riparian forests lies in their resources: basically fertile and moist soils, water, wood and non-timber forest products that are utilised by neighbouring populations to satisfy their basic needs and as source of income. Their abusive overuse is widespread and growing in Benin. As a result rich biological resources, especially plant species, are disappearing before they can be inventoried and assessed. Most recent management plans for protected areas barely consider the conservation of riparian forests. The forest law in Benin recognises their uniqueness, but several problems arise in the implementation of the rules in these particular areas. The most common and widespread tree species in riparian forests in Benin are Pterocarpus santalinoides, Cola laurifolia, Syzygium guineense, Berlinia grandiflora, Elaeis guineensis, Manilkara multinervis, Xylopia parviflora, Dialium guineense, Diospyros mespiliformis and Parinari congensis. Waterways and their forested banks are rich in birdlife, and serves as a focal point for primates and animals of many kinds. Site-specific studies at ground level are essential to assess riparian systems because their narrow linear shapes generally require data on plant community structure, floristic composition and animal presence. Details of vegetation layers under the dominant trees cannot yet be detected by remotely sensed data. This paper presents an overview of the biological diversity of riparian forests in Benin and discusses several issues associated with their protection and conservation. Key words: riparian forests, flora and fauna diversity, biodiversity protection, Benin. 3.1. INTRODUCTION The current state of knowledge about species and ecosystems in the tropics is far from complete. Detailed knowledge about the ecology of plants, floristic composition and diversity of plant communities, that would be the obvious first step in understanding and conserving them, is still incomplete or lacking. It is generally agreed that species extinction is largely related to the reduction and fragmentation of their habitats (SWAMINATHAN 1990). Therefore the protection of species can best be done through protecting habitats. In order to formulate policies or take measures to tackle the decline of biodiversity it is important to identify the ecosystems, which have a high diversity value, and are endangered. It is needed to investigate their ecology, assess their current diversity and understand the factors and processes of their impoverishment. In Benin, savanna is the dominant wooded vegetation. The narrow strips of often fragmented woodland that border streams and rivers, - the riparian forest - are good examples of such shrinking and endangered ecosystems. The rapid changes in land use in Benin have led to the destruction and fragmentation of riparian forests. They are classified as endangered ecosystems because they are marginal among the wooded vegetation and of their high degradation state. Their importance in the functioning of river/streams ecosystems and biodiversity protection is recognised by several 15 Chapter 3: Riparian forests, a unique but endangered ecosystem in Benin authors (ADJANOHOUN 1965; FORMAN & GODRON 1986; BAKER 1990; TABACCHI et al. 1990; POLANSKY 1994). Although small in size, riparian forests are important in the conservation of numerous plants and animals. Usually situated on fertile and moist soils near water, their economical interest lies in the supply of wood and non-timber forest products. Despite their wide distribution in Benin and large economic and ecological value, riparian forest ecosystems have, until recently, remained largely ignored and unmanaged as key ecosystems in biodiversity protection. This is partially due to the insufficient scientific knowledge about species occurring in riparian forests, which hinders appropriate planning, and conservation initiatives dedicated to them. Although the government of the Republic of Benin decreed a new forest law in 1993, the implementation of the law related to the protection of riparian vegetation, especially in non-protected areas, leaves to be desired. 3.2. FORESTS IN BENIN Centuries of intense human activity accompanied by a drying climate resulted in the loss of most of Benin’s closed forests long ago. Small patches of moist forest remain today only on moist soils, or as numerous sacred forests and riparian forests. FAO estimated in 1980 that only 470 km2 (only 0.4% of Benin’s land area) remained under natural cover of closed broadleaved forest. Of this, only 140 km2 was considered to be undisturbed. Flora and fauna reserves occupy 21,440 km2 (PGRN/IUCN 1994). Benin’s flora is estimated to have close to 3,000 species (PFB 1997), which is relatively poor in comparison with more humid neighbouring countries. Benin is said to be a non-forestry country because tropical rain forests are absent and dense semi-deciduous forests are rare among wooded vegetations (SOKPON 1995). In fact, Benin is located at the discontinuity of the tropical forests zone in West Africa, ‘the Dahomey gap’, which is the product of topographic, oceanographic and climatic interactions (JENIK 1994). For ERN (1988), this gap includes essentially the drier types of the Guineo-Congolian forest belt in South-Eastern Ghana, Southern Togo and parts of Southern Benin. As a result there are no evergreen tropical forests or rain forests in Benin. At the same time there is an increasing demand of wood material to satisfy the need for fuel and construction. In the South of Benin, increasing population pressure on natural ecosystems has led to rapid degradation of woodlands. The remaining forest patches include the Lama and Pobè relict forests (ERN 1988; SOKPON 1995). In Central Benin, original forests have been cleared and replaced by a mosaic of savannas and dry forests. The North, with a much drier climate, is unfavourable for forest establishment, and is therefore dominated by more thinly wooded Sudanian savanna. Meanwhile, the important hydrologic network allows the presence of many riparian forests widely distributed all over the country. The hydrographic network includes the Atlantic Ocean watershed with the Ouémé, Couffo and Mono Rivers and their tributaries, while the Niger River, takes in the waters of the Sota, Alibori and Mékrou Rivers. In the North West of Benin, the Pendjari stream starts in the Atacora mountains and ends in the Volta River in Ghana. In the South there is an important complex of coastal lagoons, lakes and swamps areas. Most streams are seasonal. 3.3. DEFINITION AND DESCRIPTION OF RIPARIAN FORESTS Riparian (from the Latin ripus, bank; as of a river) points to the banks and other terrestrial areas adjacent to watercourses, fresh-water bodies, and surface-emergent aquifers whose transported waters provide soil moisture in excess of that otherwise available locally, sufficient to support a mesic (moist soil dependent) vegetation distinct in structure and/or floristics from that of the contiguous and more xeric uplands (WARNER 1979). The regional 16 Chapter 3: Riparian forests, a unique but endangered ecosystem in Benin and temporal dimensions of climate, geomorphology, topography and biogeography mould both structure and floristics of the riparian systems (WARNER & KATIBATH 1980). Different authors have used several terms to designate riparian forests: ripicole forests (GREEN 1979a, DEVINEAU 1984), gallery forests (ADJANOHOUN 1965; DEVINEAU 1975, 1976, 1984, GREEN 1979b; KELLMAN et al. 1994; NATTA 2000; SOKPON et al. 2001); riverbank forests (GREEN 1979a), riverine forests (ADJANOHOUN 1965; DEVINEAU 1984; VARTY 1990; MONNIER 1990; MEDLEY 1992; SOKPON 1995; NATTA 2000; SOKPON et al. 2001), riverine woodlands (VAN ETTEN 1999). Other authors (e.g. BAKER 1990; TABBACHI et al. 1990; LEINARD et al. 1999) have used riparian vegetation in the sense of any plant community occurring along rivers or streams. The inter-African agreement on the definition of tropical vegetation types at Yangambi (ref. 1956) classified riparian forests among edaphic (soil dependent) forest formations (TROCHAIN 1957). Riverine forests are dependent on river processes including inundation and transport of sediments. They occur in narrow strips along river courses, and are created and maintained by groundwater seepage from the river and by periodic flooding. In the Jubba Valley (Southern Somalia) riverine forests occur in small isolated patches rarely more than 300 m wide (VARTY 1990). In the same region, MEDLEY (1992) described the Tana riverine forests (in Kenya) as fragmented patch-mosaic forests within a narrow corridor defined by the groundwater hydrologic regime, floods and meandering pattern of the stream. In the West-African region, MONNIER (1990) described the riverine forests as deciduous forest characterised by a linear shape, a structural originality and a microclimatic specificity. In the forest-savanna ecosystems of Côte d’Ivoire, ADJANOHOUN (1965) qualified riverine forests at both riversides as strips dominated by big trees and liana species, belonging almost exclusively to dense forest. They appear to be one of the most important ecosystems in the majority of savanna landscapes where they underline the course of rivers. In the Benin context, where savanna is the most dominant ecosystem, we suggest to use the term ‘riparian forest’ for any forest type occurring at river banks and along streams. This may include semi-deciduous forest, dry deciduous forest and woodland savanna located at river or stream banks with more or less sharp limits in respect to adjacent plant communities. Although these plant communities, dominated by trees, are often fragmented and degraded, there are still good examples of undisturbed riparian forests (e.g. at Yarpao, Pénéssoulou and the Bondjagou forest in the Pendjari National Park). In many regions high human pressure has reduced the width of riparian forests up to a single tree, so they are often extremely narrow and quite linear (strip-like) in configuration. Floristic data and the distinction between core and edge species may be important to describe the limit of riparian forests when the structure is uniform beyond the riparian forest. The main characteristics of typical riparian forest ecosystems can be summarised as: - Location along rivers and streams in close contact or not with the water flow; - Combination of mesic terrestrial vegetation, dependent animal life, and local microclimate and dynamic water and soil processes; - Multi-layered vegetation structure dominated by moist-dependent trees, which form a distinctive physiognomy in the landscape, particularly in the savanna region. Associated with these mesic trees is a host of mesic understorey, shrub and ground cover species; - Floristic composition often includes the plant species of the region that require most moisture. We do not consider wetland vegetations occurring in swamp areas (shrubs, grass savanna, marsh, mangrove vegetation, etc.) as riparian forests. The distinction between several riparian forest stands can be expressed in plant communities on the basis of their physiognomic, floristic, and structural characteristics through phytosociological surveys. 17 Chapter 3: Riparian forests, a unique but endangered ecosystem in Benin These are compiled on the basis of abundance-dominance by allocating a coefficient to each plant species according to the ratio of species density to the area surveyed. 3.4. ECOLOGICAL AND ECONOMIC IMPORTANCE OF RIPARIAN FOREST ECOSYSTEMS Biodiversity is often seen as an indicator of the well-being of ecological systems, as well as a useful tool in environmental monitoring and assessment. The role played by biodiversity is strongly related to natural ecosystems, which are often the only viable locations for the conservation of certain natural variability (SALLEY & NIENG 1997). Water and vegetation are central components of tropical ecosystems (OTTO 1993), which in turn are key repositories of global biodiversity. In the sub-Sahara savanna region, where small patches and fragmented woodlands often represent remnant forests, riparian forest stands are sites of important ecological and economic values. Also, resource managers understand the interrelationships between the four components of an ecological site: landscape position (physiography), climate, soils and vegetation. These basic relationships must still be considered for riparian areas, as must landform, fluvial geomorphology and hydrology. Thus soils, fluvial geomorphology and climate are the most reliable indicators of riparian vegetation potential (LEINARD et al. 1999). More than other vegetation types, riparian forests are well-known for their role in controlling water sources, watersheds, water runoff and quality, mineral nutrient flows (PIEGAY 1997), stream bank erosion (VAN ETTEN 1999). They influence sedimentation, bank stabilisation, flood and hydrologic regimes, pollution, water shading and cooling, sediment filtering, aquifer recharge, and transport of sediments (MEDLEY 1992). According to THOMAS (1996), typical riparian forest tree species are dependent on river flows, a shallow aquifer, and the community and population structure of riparian forests are related to spatio-temporal patterns of flooding. They are under the influence of fluvial geomorphology processes (BAKER 1990). Also they usually act as routes for movement of terrestrial plants and animals across the landscape (FORMAN & GODRON 1986) and serves as the most suitable area for plant species adapted to a moist climatic regime (MEDLEY 1992). On riparian systems in general, the United States Council on Environmental Quality stated ‘no ecosystem is more essential to the survival of fishes and wildlife’ (WARNER & KATIBATH 1980). Although small in size (VARTY 1990; MONNIER 1990), riparian forests are important in the conservation of a large range of plants and animals. Not only do they constitute a natural habitat or the last refuge for many species, but also they contain many endemic species and extinction-menaced species (ROGGERI 1995). Some authors have discussed the necessity to preserve plant diversity in riparian forests as a means to protect certain animals. In the Tana River National Primate Reserve of Kenya, Medley (1992) showed that the preservation of key resources in riverine forest, such as the endangered Tana River red colobus (Colobus badius rufomitratus) and crested mangabey (Cercocebus galeritus galeritus), should be coupled with protection of forest heterogeneity that characterises this dynamic landscape. 3.5. DESTRUCTION AND LOSS Misuses of riparian systems are global, chronic and accelerating (WARNER & KATIBATH 1980). Decades or centuries of human influence are usually seen as the main causes that have reduce riparian forests in both size and structural complexity. The result is a poorer biodiversity and landscape impoverishment. Their rich biological resources, especially plant species, are disappearing before they can be inventoried and assessed (NATTA 2000). As a consequence, this affects ecological equilibrium especially at local level (MONNIER 1990). 18 Chapter 3: Riparian forests, a unique but endangered ecosystem in Benin The interest of riparian forests lies in their resources: basically fertile and moist soils, water, wood and non-timber forest products that are utilised by neighbouring populations to satisfy their basic needs and as source of income. The fertile soil, the presence of water and moist soils, the supply of sediments from inundation, allow certain human activities such as shifting cultivation, irrigation and cattle ranching. Also the presence of a variety of plants and animals related to water regime, such as valuable and rare tree and animal species, enables selective tree cutting (for timber, traditional fuel gathering for sale, honey collection) and hunting. The collection of all the seeds of valuable tree species (e.g. Pentadesma butyracea, SINADOUWIROU 2000) is an additional threat to the survival of their population. Also disruption of watershed vegetation, damming of watercourses and/or diversion of stream flow, excessive lowering of aquifer levels through pumping, channelisation and levee construction on watercourses often cause significant damage to riparian systems. As a result indigenous multilayered plant communities have been completely removed in many parts and replaced by weeds, open field, shrubs or grass savanna with much less ecological value and economic potential. 3.6. PLANT COMMUNITIES AND FLORISTIC COMPOSITION OF RIPARIAN FORESTS IN BENIN Following the work of ADJANOHOUN et al. (1989), HOUINATO et al. (2000) divided Benin in ten phytogeographic districts with distinctive geomorphologic, plant communities and flora. For each district an overview of riparian forest flora is presented. The coastal region is the domain of marshlands, lakes, lagoons, coastal thicket or scrub and a number of small relict groves. The present-day pattern of vegetation is very much obscured by the omnipresent impacts of dense human populations. This flat region lacks typical riparian forests, as defined above. The Pobè district is the domain of dense semideciduous forests and is a continuation of the lowland rain forest zone of Nigeria. SOKPON (1995) studied the structure and floristic composition of plant communities in the protected forest of Pobè, typical of this South-Eastern part of the country. He identified a riverine forest dominated by Cleistopholis patens, Ficus mucuso, Elaeis guineensis and Cercestis mirabilis. The Ouémé and Kouffo districts prolong respectively the derived savanna and the Southern Guinea zones of Nigeria. Here Pterocarpus santalinoides, Cola laurifolia, Parinari congensis, Antidesma venosum, Napoleonaea vogelii, Syzygium guineense, Dialium guineense and Cynometra megalophylla are the dominant tree species in riparian forests. The Zou district is a transition zone between the Guinean and Sudanian climates. Riparian forest physiognomy is similar to the dense semi-deciduous forests. The most frequent species are Berlinia grandifolia, Elaeis guineensis, Hexalobus crispiflorus, Pouteria alnifolia, Cola gigantea, C. millenii, Lecaniodiscus cupanioides, Napoleonaea vogelii, Pterocarpus santalinoides and Uvaria chamae. The South Borgou district corresponds to the North Guinean zone of Nigeria and the riparian forest flora here contains numerous tree species of lower and much wetter latitudes such as: Berlinia grandifolia, Parinari congensis, Detarium senegalense, Diospyros mespiliformis, Dialium guineense, Khaya grandifoliola, K. senegalensis, Millettia thonningii, Albizia zygia, Albizia glaberrima, Trilepiseum madagascariense. The North Borgou district, which prolongs the Sudanian zone of Nigeria, is characterised by the omnipresence of riparian forests similar in physiognomy with those of South Borgou but with a different flora. Dominant species are Syzygium guineense, Uapaca togoensis, Berlinia grandifolia, Brenadia salicina, Khaya senegalensis, Elaeis guineensis, Manilkara multinervis, Vitex doniana, Mimusops andongensis, Diospyros mespiliformis, Synsepalum passargei, Fadogia agrestis, Ficus spp., Celtis toka, Borassus aethiopum and Raphia sudanica. 19 Chapter 3: Riparian forests, a unique but endangered ecosystem in Benin The Atacora chain oriented North-North-East/South-South-West is characterised by hygrophile riparian forests at the foot of the hills. Their structure and composition present some similarities with the ones of North Borgou. Common species are Syzygium guineense, Uapaca togoensis, Berlinia grandifolia, Brenadia salicina, Pentadesma butyracea, Chrysobalanus atacoriensis, Eriocoelum kerstingii, Khaya senegalensis, Ficus spp., Diospyros mespiliformis, Anogeissus leiocarpus and Vitex doniana. The Pendjari district corresponds to the North West of the Atacora province. It includes the major part of the Pendjari National Park and the Western plains up to the Togo Border. In the Pendjari National Park, GREEN (1979b) identified several plant associations among riparian forests. The most important are Cola laurifolia, Morelia senegalensis and Syzygium guineense association at its maximum development in the unique and species rich riparian forest of Bondjagou; Khaya senegalensis, Diospyros mespiliformis and Anogeissus leiocarpus along the Pendjari stream in the Atacora chain; Celtis integrifolia, Kigelia africana and Diospyros mespiliformis, Garcinia livingstonei and Mitragyna inermis along the Pendjari stream North of the Bondjagou forest; Anogeissus leiocarpus, Combretum nigricans sometimes with Terminalia glaucescens along several streams in the park; Anogeissus leiocarpus and Crossopteryx febrifuga along small streams around the Bondjagou forest, in the centre of the park and at the foot of the hills. In the Western plains towards the Togo border, riparian forests are dominated by Anogeissus leiocarpus, Mitragyna inermis, Elaeis guineensis, Ficus spp., Sarcocephalus latifolius, Vitex doniana and Terminalia macroptera (NATTA 1997). In the Centre and North of the W National Park in Northern Benin, riparian forests along streams are the most important ecosystems in the majority of shrub savanna dominated by Mimosoideae and Combretaceae. Here, riparian forests have the structure of woodland savanna less than 15m high along the Mékrou, Alibori and Pako streams. They are extremely narrow and quite linear with often only a single tree at stream banks. The most frequent species are: Cola laurifolia, Syzygium guineense, Morelia senegalensis, Pterocarpus santalinoides, Combretum lecardii, Crateva adansonii, Borassus aethiopum, and Vitex chrysocarpa. Oxytenanthera abyssinica is sometimes seen in the Mékrou and Alibori beds (PGRN/IUCN 1994). GREEN (1979a, 1979b) also identified several plant associations along the Niger river and streams of the W National Park such as Khaya senegalensis, Diospyros mespiliformis and Borassus aethiopum; Anogeissus leiocarpus, Combretum nigricans with sometimes Sclerocarya birrea in the North of the Park along the Mékrou stream; Anogeissus leiocarpus and Crossopteryx febrifuga along the Niger river, the Mékrou stream and small streams; Cola laurifolia, Morelia senegalensis, Syzygium guineense, Anogeissus leiocarpus, Feretia apodanthera, Anogeissus leiocarpus and Balanites aegyptiaca. Trilepiseum madagascariense, Cleistopholis patens, Albizia glaberrima, Antidesma venosum, Sterculia tragacantha, Cynometra megalophylla, Drypetes floribunda and Syzygium guineense are common species of semi-deciduous forests on ferralitic soils or vertisols in the Pobè and the Ouémé districts South of 7º10 N, (AKOÈGNINOU 1984; SOKPON 1995). These trees are also found almost exclusively along rivers and streams at higher latitudes. Such is the case in the Toui-Kilibo, Pénéssoulou, Ouari-Maro, and Ouémé Supérieur protected areas. Dialium guineense, a typical species of the plateau and tops of slopes in dense semi-deciduous forests (SOKPON 1995), is found exclusively in riparian forests between 7º20 and 11º N. Also Elaeis guineensis is widespread in the South, but is found in gallery forests in the Centre and North of the country. Syzygium guineense is always found in gallery forests as far as in the W National Park. These examples tend to confirm the hypothesis of FORMAN & GODRON (1986) that riparian forests usually act as routes for movement of plants across the landscape, especially at less moist latitudes. 20 Chapter 3: Riparian forests, a unique but endangered ecosystem in Benin Despite decades of neglect, the Mounts Kouffé reserve still contains dozens of small patches of riparian forests and dense moist forests. These remnants appear undisturbed internally and therefore have a very high biodiversity value (PGRN/IUCN 1994). GREEN (1979a) and PGRN/IUCN (1994) have recommended to extend the South-Eastern limit of Pendjari National Park beyond the Atacora chain to protect the riverine forest and adjacent mountains, thereby increasing biodiversity by adding rich habitats and beautiful landscapes to the park. Also a buffer zone should be created along the Ouémé River, including its forested banks to conserve the riverine forest and to maintain a physical corridor between the Ouémé Supérieur and Ouari-Maro forest reserves. Summing up, the most common and widespread tree species in riparian forests in Benin are Pterocarpus santalinoides, Cola laurifolia, Syzygium guineense, Berlinia grandiflora, Elaeis guineensis, Manilkara multinervis, Xylopia parviflora, Dialium guineense Diospyros mespiliformis and Parinari congensis. 3.7. WILDLIFE IN RIPARIAN FORESTS The Ouémé river and its forested banks are rich in birdlife, and serves as a focal point for primates and animals of many kinds. The Mounts Kouffé with its contiguous Ouari-Maro and Ouémé Supérieur protected forests are said to have the highest ranking of biodiversity as compared with all the other protected areas in Benin (PGRN/IUCN 1994). It is very likely that the 215 bird species and the potential 150 additional ones, reported by the PGRN/IUCN mission of 1994, have riparian forests as their main habitat. In the North of Benin and the South of Burkina Faso, flood plains, shallow lakes and riparian forests are the principal habitats of the 278 bird species found in the Pendjari and Arli National Parks (ROGGERI 1995). Riparian forests are reportedly the main habitat of certain animals such as the bushbuck (Tragelaphus scriptus), red-flanked duiker (Cephalophus rufilatus) and vervet monkey (Cercopithecus aethiops tantalus) (LEVAUX 1990). Also in the protected areas, certain animals depend on the dispersion of fruits of riparian forest trees. It is the case e.g. of elephants and baboons that use Borassus aethiopum seeds in the Pendjari National Park. 3.8. THE USE OF REMOTE SENSING IN DETECTING AND MAPPING RIPARIAN FORESTS Remotely sensed data and geographical information systems (GIS), which deal with spatial data, have considerable utility in acquiring vital information especially on vegetation. They have a great potential in providing up-to-date and accurate information that can be used for plant communities management and monitoring activities. We have conducted a study in the Toui-Kilibo protected area (Central Benin) to assess qualitatively and quantitatively the usefulness of satellite imagery (Spot XS image) for identification and classification of land cover types in riparian forests and neighbouring vegetation types (see NATTA 2000). It was possible to differentiate between several land cover types from the aerial photographs and satellite images along the river, but not along streams. Although results seem promising in terms of spectral separability of gallery (along streams) and riverine (along rivers) forests, differences in tree species structure and floristic composition, as observed in the field, could not always be made out from the remotely sensed data. The coarse spatial resolution of SPOT XS image (20 m) and the unknown spectral signatures of species or group of species have lowered the accuracy records. The great spatial resolution of predictive vegetation maps obtained from digital image classification exclude small patches and narrow strips of vegetation such as riparian forests, and therefore limit the 21 Chapter 3: Riparian forests, a unique but endangered ecosystem in Benin use of the current civilian satellite imagery in riparian forests study. WARNER & KATIBATH (1980) have already found that satellite imagery allowing vegetation mapping with a scale range between 1:80,000 to 1:35,000 was only of limited use; meanwhile they found that an aerial photograph at 1:6,000 is the most useful for detailed riparian system inventory. In our case study we found that even aerial photographs with a scale of 1/20,000 or larger are useful to identify and map broad stands of riparian forests in a savanna landscape. Because data on plant communities’ structure (dominant trees, understorey, shrub cover and ground cover), floristic composition, as well as associated animal life, are generally required ground level surveys are an essential component of any detailed riparian system assessment. Site-specific studies in the field are required to assess riparian forests because their narrow linear shapes generally require data on plant community structure, floristic composition and animal presence. Details of vegetation layers under the dominant trees cannot yet be detected by remotely sensed data. 3.9. PROTECTION, AND RESTORATION OF RIPARIAN ECOSYSTEMS RESOURCES In the new Benin forest law (no. 93-009 of July the 2nd of 1993, Ref. RB 1993), clearance of wood and shrubs is not allowed within 25 m at both sides of any water course and stretch of smooth water (article 28). This limit seems to be too short, in reality stands of riparian forests can be wider than 50 m. A reasonable limit, without any clearance or selective tree cutting from watercourses, of 100 m will provide sufficient guarantee for the protection of watercourses and riparian resources. In the management plans of Goun-Goun, Sota and Goroubi, Toui-Kilibo, Ouémé Supérieur and Pénéssoulou forest reserves, gallery forests will be left uncut, and rare species outside the gallery forests will not be cut either. In the meantime, foresters and rural extensions workers must give the good example and promote local people’s rights and duties specified by the forest law, especially to help local people respect the restrictions concerning protected or valuable trees (e.g. Milicia excelsa, Khaya senegalensis, and K. grandifoliola at stream sides). Still several questions remain with regard to the implementation of the forest law: Do local people respect the restrictions on riparian forest utilisation? Are alternatives provided to neighbouring people, particularly farmers, not to overuse or destroy riparian forests (e.g. when seeking for fertile soils or water for irrigation)? Are private forestry developers aware of protection of riparian forests in unprotected areas? How far can foresters enforce the law in applying the punishments to those who illegally destroy riparian forest resources? Can any suggestions be given to improve the efficiency of the law in protecting riparian forests resources in the terrain? To protect, improve, and restore in a sustainable way the riparian vegetation resources, several objectives have to be addressed: - identify the major riparian forests in the country and determine their historical extent; - make inventories of their floristic and faunistic resources; - identify causes of riparian forests destruction and threats to their resources maintenance, with special attention to local people’s views; - implement activities for the sustainable conservation of riparian forests in partnership with neighbouring inhabitants. Information needed To achieve these objectives, data should be collected on different topics related to riparian forest such as: 22 Chapter 3: Riparian forests, a unique but endangered ecosystem in Benin - location, extent, climate, soils types, river/stream bank conditions; - vegetation structure, floristic composition, and vegetation condition (species distribution, abundance, size, reproduction strategies, evolution trends, etc.); - fauna with their characteristics and ecological requirements; - river and /stream water quality, hydrologic characteristics; - adjacent land uses and their impacts on riparian resources (effects of farming, tree cutting, fire, hunting, channelisation, hydrologic manipulations and other discernible threats); - measures to be taken for riparian resource protection, improvement, and restoration (possible solutions to the major threats at local, district, provincial and national levels, restoration potential for the riparian systems). Stakeholders Several stakeholders have to be involved in this task: the forest and natural resource management authorities, foresters, projects, research institutes and Non-Governmental Organisations (NGO’s) dealing with natural resources protection, local population and forest resources users (e.g. timber gatherers, hunters), etc. Effective conservation actions can best be achieved if the local residents are either collaborating or in control. Therefore initiatives have to be taken to rise local people’s awareness concerning their interests/benefits in protecting riparian forests (landscape beauty, watershed protection, water sources protection, conservation of valuable or rare floristic and faunistic species). This can be done through the empowerment of local forest management committees through participatory approaches. The implementation and improvement of the forest law are primarily the task of foresters in the terrain. The application of the law and good collaboration with local natural resources management committees, if effective, will surely provide positive effects on forested areas including riparian systems. Role of the Forest and Natural Resource Management (NRM) Department and Research Institutes in the implementation of the forest law concerning riparian forests Their attributions could be: - Definition of thematic objectives (e.g. riparian forests fauna, flora, river and stream hydrology, reforestation, socio-economic aspects, etc.); - Setting up of criteria for the assessment and monitoring of riparian forests resources; - Creation of a data-base on riparian forests in the framework of any NRM program; - Monographs of specific species (e.g. Chrysobalanus atacoriensis, Pentadesma butyracea, etc.) or resource. Role of research projects and NGOs dealing with natural resources protection Projects and NGOs dealing with natural resources protection are certainly very important for the protection, improvement, and restoration of riparian vegetation resources. Their actions could focus on: - Inventory of natural resources in riparian forests; - Vulgarisation of the forest law regulations on riparian forests and effective actions to fight against the depletion of riparian forest resources in collaboration with local people; - Empowerment of local forests management committees through participatory approach; 23 Chapter 3: Riparian forests, a unique but endangered ecosystem in Benin - Alternatives to reduce the depletion of riparian forest resources (e.g. cheap and easy to use irrigation systems for farmers, agroforestry in uplands, appropriate techniques to improve soil fertility, beekeeping, etc.); - Experimentation on husbandry of rare or native riparian forest species; - Participatory reforestation of degraded riparian forests and revegetation of stream banks with rare or native riparian forest species (e.g. Khaya grandifoliola, K. senegalensis, Pentadesma butyracea). Exotic plants (e.g. teak, Gmelina arborea, Anacardium occidentale and Mangifera indica) should not be avoided; - Valorisation of riparian forests in National Parks and forest reserves for ecotourism, recreation, research and sport fishing (e.g. construction of a viewing tower near the Bondjagou forest); - Protection and valorisation of rich riparian forests outside protected areas, especially around human settlement. Indigenous riparian forests around cities should not to be converted to settlement, instead be managed (delineated and protected) as natural parks for recreation, research, education, etc.; - Improvement to the forest law. Some improvements can be given to the forest law. It is urgent to put under protection or extend the protected areas of Ouémé Supérieur, Mounts Kouffé, Ouémé-Boukou, Dogo, Kétou to the second side of the Ouémé River that is not currently included in the reserves. This disposition can be extended to any other stream or river harbouring important riparian vegetation (e.g. Yarpao). Also the principle of regulated tree harvest in riparian forests might be accepted (e.g. up-rooted trees, old or dying trees at river/stream edges) followed by special care to rare or useful tree species regeneration. 3.10. CONCLUSIONS In line with the Convention on Biological Diversity (UNCED 1992), which emphasised the need for every country to take responsibility for conserving its own biodiversity, a prudent approach to biodiversity conservation necessitates preserving all potential rich or endangered ecosystems. Therefore Benin’s relict forests, especially riparian forests, are important conservation sites that need more care than currently available. In the majority of savanna ecosystems, efforts toward the conservation of riparian forests will increasingly become a priority in land use planning (NATTA 2000), this will allow a wide range of plants and animals to benefit from the protection of forests along rivers and streams. Riparian systems are at the same time so important as a collection of natural resources and so limited in individual size that they both merit and require higher resolution inventory procedures than those deployed for other larger ecosystems (savanna and upland forests). The responsibility of different actors, (e.g. forest division, research institutes, projects and NGO’s dealing with the conservation of biodiversity in Benin in collaboration with local residents), should be drawn to seek for appropriate initiatives to preserve the remaining rich and diverse riparian forests, especially in non-protected areas. 24 Chapter 4 ASSESSMENT OF RIPARIAN FOREST FRAGMENTS PLANT DIVERSITY IN WEST AFRICAN SAVANNA REGIONS: AN OVERVIEW FROM BENIN Submitted to the Journal of Biogeography Natta, A.K. and van der Maesen, L.J.G. Chapter 4: Assessment of riparian forest plant diversity in Benin Chapter 4 ASSESSMENT OF RIPARIAN FOREST FRAGMENTS PLANT DIVERSITY IN WEST-AFRICAN SAVANNA REGIONS: AN OVERVIEW FROM BENIN (*) Natta, A.K. (1) and van der Maesen, L.J.G. (2) (1) Faculty of Agronomic Sciences, University of Abomey-Calavi, Benin 01 BP 526 Cotonou, aknatta@yahoo.com; (2) Biosystematics Group, Wageningen University, Gen. Foulkesweg 37, 6703 BL Wageningen, The Netherlands, jos.vandermaesen@wur.nl. (*) Submitted to the Journal of Biogeography ABSTRACT: Aim: To assess the flora and plant diversity of riparian forests. To compare the diversity of several riparian forest sites. Location: From 7º10 to 12º20 N in Benin (West Africa). Methods: The largest and least disturbed forest fragments along rivers and streams were surveyed in the three ecological (i.e. Guinean, Sudano-Guinean and Sudanian) regions of Benin. In total 373 phytosociological relevés (each of 500 m2) were made from 1999 to 2002, using the Braun-Blanquet method. A plant species list was made from these relevés. Biodiversity was assessed through species richness, Shannon index (H’), Equitability index of Pielou (E), and species abundance models. Results: In all 1003 species (about 1/3 of the estimated number of species in the flora of Benin), 120 families, 513 genera, and 224 tree species (i.e. dbh ≥ 10 cm) were identified. The richest (sub) families were the Papilionioideae, Poaceae, Rubiaceae and Euphorbiaceae. Riparian forests are characterised, on the one hand, by a flood-dependent flora, and on the other hand by many species from semi-deciduous forests and savannas. Shannon index varied from 2.4 to 5.8 bits and Pielou’s equitability from 0.51 to 0.86. Species richness varied from 129 to 358 while tree species richness ranged from 27 to 99 per ha. Main conclusions: This study showed that in fire-prone environments, such as Benin, relatively large numbers of species still are maintained in small forest fragments along waterways. Riparian forests are characterised by a diversified flora, and specific plant species that should be used for the restoration of degraded stands. Key words: riparian forests, flora, diversity, species abundance model, Benin. 4.1. INTRODUCTION Benin is located in ‘the Dahomey gap’, which is a break of the West African rain forest belt between Nigeria and Ghana. This interruption permits the incursion of savanna towards the coast as a result of the dry conditions prevailing in Southern Benin, Togo and South-Eastern Ghana (Onochie, 1979; Ern, 1988; Jenik, 1994; Kokou et al., 1999; Kokou & Caballé, 2000). As a result there are no evergreen tropical forests in Benin (Sokpon, 1995; Tossou, 2002). High incidences of farming as well as tree cutting and fire have reduced the original woodland to their current state. Flora is a major component of a region or a country’s biodiversity. It is historically the result of a long process of natural selection (Braun-Blanquet, 1972) in relation with climatic and edaphic conditions as well as human interference. In the majority of degraded woodlands and mosaic of savanna landscapes, riparian forests (hereafter RFs) belong to the most 27 Chapter 4: Assessment of riparian forest plant diversity in Benin important forest ecosystems, which follow the outline of rivers and streams (Natta, 2000). The flora of Benin is incompletely known (PFB, 1997), and the diversity, relative rarity and threats of many hot-spots in vegetation, such as riparian forests, have remained insufficiently studied (Mondjannagni, 1969; Paradis, 1988; Sokpon et al., 2001). This lack of information to the scientific community and managers hinders appropriate planning and conservation initiatives dedicated to them (Natta et al., 2002). The aim of this paper is to investigate the flora of riparian forests through several parameters: plant species richness and abundance, endemic, valuable and threatened plants, adaptability to a specific habitat, Diversity Index (Shannon), Equitability Index (Pielou), differences in sites diversities (comparison of Shannon indices), and species abundance models. This research also intends to provide accurate floristic data for riparian forests biodiversity protection and restoration in different ecological regions. 4.2. METHODS Surveys were made in all ecological regions of Benin excluding the coasts, beaches, coastal lagoons, lakes and marshlands. The study area (Figure 4.1) covered the latitudes 7º10 to 12º20 N. The largest and least disturbed forest fragments along rivers and streams were surveyed in the Guinean region (Samiondji and Bètèkoukou), Sudano-Guinean zone (TouiKilibo, Idadjo, Bétérou and Pénéssoulou) and Sudanian region (Yarpao, Pendjari Biosphere Reserve, Gbèssè). In total 373 plots (each of 500 m2) were installed in several RF stands using the phytosociological approach of Braun-Blanquet, but tree height and diameter at breast height (i.e. dbh) were measured in 304 plots (Table 4.1). About 19 ha of RFs were surveyed for the compilation of a plant species list. Rectangular plots, with variable length and width, were preferred to circular ones to fit with the shape of waterways and width of riparian forests. The identity of each plant was recorded in the field and specimens taken as vouchers for the National Herbarium of Benin and the National Herbarium of the Netherlands, Wageningen University Branch. Species names mainly follow Keay & Hepper (1954-1972), Brunel et al. (1984), Berhaut (1967), and Lebrun & Stork (1991-1997). Stem diameter at breast height (dbh) was measured at 1.3 m above the ground, or above buttresses when present for trees (dbh ≥ 10 cm). In the present paper the assessment of plant species diversity in RFs encompasses species richness, Shannon diversity index (H’), Equitability index of Pielou (E), and species abundance models (also called Dominance-diversity curves). Here species abundance models test the statistical difference between observed pattern within the floristic data (tree species abundance and families richness) and expected pattern as described by each model. Visual inspection of rank-abundance figures suggested two predictors of species abundance models: the geometric and log series. The χ2 test compared the two distributions, the expected and observed number of each species in each abundance class. We compared the diversity of RF sites (1 ha per site), using a t-test based on the Shannon Index and its variance (cf. Hutcheson, 1970; Magurran, 1988; Kent & Coker, 1992). This method was already successfully used on RFs in Benin and Ivory Coast (cf. Natta, 2000; Porembski, 2001; Natta & Porembski, in press, see chapter 9). The usefulness of multivariate methods (e.g. classification and indirect ordination), in detecting trends and subdivisions within riparian forests floristic data, was assessed in another paper (cf. Natta et al., in press, see chapter 6). 28 Chapter 4: Assessment of riparian forest plant diversity in Benin N Gbèssè * Pendjari Reserve * Towns Yarpao * Pénéssoulou Bétérou * * * TouiKilibo Idadjo * Bétékoukou * Samiondji 1 : Grand-Popo 2 : Cotonou 3 : Ouidah 4 : Porto-Novo 5 : Niaouli 6 : Aplahoué 7 : Pobè 8 : Bohicon 9 : Kétou 10 : Savalou 11 : Savè 12 : Bantè 13 : Tchaourou 14 : Bassila 15 : Djougou 16 : Bembèrèkè 17 : Natitingou 18 : Parakou 19 : Kouandé 20 : Tanguiéta 21 : Toui 22 : Kandi 23 : Banikoara * Sites of data collection (with > 20 plots) States limits Major waterways * Isohyets Figure 4.1: Study sites where data were collected in riparian forests in Benin 29 Chapter 4: Assessment of riparian forest plant diversity in Benin 4.3. RESULTS 4.3.1. Floristics and stand characteristics of riparian forests throughout Benin In Benin, RFs vary in their width along the watercourses, which they fringe. Some are no more than a few meters wide and have an open canopy, while others are much more extensive and have a closed canopy and a well-developed structure. The floristic and stand characteristics of riparian forest are summarised in Table 4.1. Species richness (SR) per ha varied from 129 to 358. Tree density and basal area per ha varied from 253 to 785, and from 23 to 59 m2, respectively. Table 4.1: Floristic and stand characteristics of riparian forests throughout Benin Regions Sites Relevés Species Richness Species Richness Trees (SR) per site (SR) per ha density/ha Basal area (BA/ha) Guinean Samiondji 36 306 249 726 41.7±17.4 Sudano Guinean Pénéssoulou Idadjo-Bétérou Okpara Streams at Toui- Kilibo 40 40 56 62 450 183-195 - 343 129-195 - 544 748-785 315 257 38.9±15.3 45.6±15.8 - All Sudano-Guinean sites 198 556 129-343 257-785 42.6±15.8 RFs on plateau RFs at hill foot 43 27 435 467 182-278 358 538-627 732 35.6±16.5 42.7±9.9 All Sudanian sites 70 591 182-358 538-732 40±20.6 304 1003 129-358 253-785 41.5±17.9 Sudanian All RFs sites in Benin SR = Species richness per site (SR/site) or per ha (SR/ha); BA/ha = Basal area per ha in m2; Trees = stems with dbh ≥ 10 cm; RFs = Riparian forests. 4.3.2. Diversity of riparian forest flora in Benin From 373 floristic relevés altogether 1003 species (about 1/3 of the estimated number of species in the Benin flora) from 120 families and 513 genera were recorded. The ten (sub) families richest in species are the Fabaceae-Papilionoideae, Poaceae, Rubiaceae, Euphorbiaceae, Cyperaceae, Asteraceae, Acanthaceae, Fabaceae-Caesalpinioideae, Moraceae and Malvaceae. In all, 224 tree species (i.e. dbh ≥ 10 cm) were identified. Tree species with the highest abundance are Pterocarpus santalinoides, Cola laurifolia, Syzygium guineense, Dialium guineense, Berlinia grandiflora, Cynometra megalophylla, Elaeis guineensis, Diospyros mespiliformis, Uapaca togoensis (see Table 4.2). The list of plant species collected in several riparian forests sites is presented at the end of this dissertation. This flora contains species from various forest types (dense semi-deciduous, dry, open), savannas and various degraded woodlands and fallow communities. Species such as Pterocarpus santalinoides, Cola laurifolia, Syzygium guineense, Culcasia scandens and Morelia senegalensis are seen at all latitudes along waterways. On the other hand, some species are confined to certain regions showing that RFs present niches to conserve dense forest species in various landscapes. In the humid Guinean region of Southern Benin and along larger waterways (rivers), typical species are Cynometra megalophylla, Salacia pallescens, Parinari congensis, Drypetes floribunda, Cassipourea congoensis, and Garcinia livingstonei. Along small waterways (streams), mainly in the Sudanian region of North Benin, typical RF species are Raphia sudanica, Khaya 30 Chapter 4: Assessment of riparian forest plant diversity in Benin senegalensis, Uapaca togoensis, Vitex doniana, Oxytenanthera abyssinica, Berlinia grandiflora, Brenadia salicina, Garcinia ovalifolia, Pentadesma butyracea and Chrysobalanus atacoriensis. Table 4.2: Riparian forests tree species dominance: number of individuals and basal area no. 1 2 3 4 5 6 7 8 9 Species Pterocarpus santalinoides Cola laurifolia Syzygium guineense Dialium guineense Berlinia grandiflora Cynometra megalophylla Elaeis guineensis Diospyros mespiliformis Uapaca togoensis Total of the 9 most abundant tree species Total of the remaining (215) tree species Overall total for 224 tree species Absolute abundance 1194 631 499 271 249 248 222 204 201 Abundance (%) 14.86 7.86 6.21 3.37 3.1 3.08 2.76 2.54 2.50 3719 4311 8030 46.3 53.6 100 Total BA BA(%) (m²) 1038 12.8 707 8.7 585.2 7.2 264.6 3.25 502.7 6.18 460.7 5.66 240.7 2.96 463.2 5.7 237.4 2.92 4499.52 3630.82 8130.34 55.3 44.6 100 BA= Basal Area in m² Many species from forest, regrowth vegetation or pioneer species in the Guinean region and Sudano-Guinean zone of Southern and Central Benin (e.g. Elaeis guineensis, Dracaena arborea, Dialium guineense, Holarrhena floribunda) are found in streamside vegetations in the savanna region of North Benin. Also, certain pioneer or secondary forest species (e.g. Leea guineensis, Rauvolfia vomitoria, Phyllanthus muellerianus, Anthocleista nobilis, Cleistopholis patens, Spathodea campanulata, Ceiba pentandra, Milicia excelsa and Ricinodendron heudelotii) of the tropics (cf. Whitmore 1990) are usually seen in RFs in Benin. Other species (e.g. Khaya senegalensis, Afzelia africana, Milicia excelsa and Antiaris toxicaria) that have completely disappeared on plateaus in densely populated areas are found in RFs at higher latitudes or far from settlements. We also find some either endemic or valuable species almost exclusively in RFs. Thunbergia atacorensis (Acanthaceae) is one of the rare endemic species in Benin mainly found in RFs at hills feet. Among valuable species, Dissotis anthennima is known for its importance in medicine, Pentadesma butyracea and Xylopia aethiopica as major Non Timber Forest Products in West Africa. Summing up, RFs in Benin are characterised, on the one hand by a flood-dependent flora widely distributed all over the country, on the other hand by many forest (pioneer and typical semi-deciduous) and savanna species. 4.3.3. Riparian forest sites diversity The Shannon index varies from 2.43 to 5.4 bits respectively along the Ouémé river at Samiondji (South Benin) and along streams in Toui-Kilibo reserve forest (Central Benin). The overall RFs diversity is 5.8 bits, and the Equitability index of Pielou is 0.74 (Table 4.3). RFs along streams (Toui-Kilibo, Pénéssoulou, Sudanian RFs on plateau and at hill feet) have higher diversities than those along rivers (Okpara, Ouémé, Pendjari and Sota): t = 31.57; df = 7502; p < 0.001. 31 Chapter 4: Assessment of riparian forest plant diversity in Benin Table 4.3: Plant species diversity in several riparian forests sites in Benin REGION Sites Plots Tree species Abundance Shannon index Evenness (n) richness (SRt) (N) (H') (E) Guinean Samiondji (Sa) 36 27 1307 2.43 0.51 Sudano Guinean Pénéssoulou (Pe) Idadjo-Bétérou (ID-BE) Okpara River (Okp) Streams at Toui-Kilibo (Str TK) 40 40 56 62 99 64 59 78 1086 1537 883 1003 5.17 3.91 4.91 5.4 0.78 0.65 0.84 0.86 Sudanian RF on plateau (Sud Plat) Sudanian RF at hill foot (Sud Hill) 43 27 84 65 1226 988 4.97 4.27 0.78 0.71 Rivers (large waterways) Streams (small waterways) 156 148 111 144 4291 3739 4.44 5.87 0.65 0.82 304 224 8030 5.81 0.74 Sudanian Riparian forests of BENIN n = Plots; SRt = Tree species richness per site; N = Abundance; H’ = Shannon index; E = Equitability index of Pielou. The comparison of RF sites diversities (Table 4.4), based on the t-test adapted to Shannon index, gave the ranking: Streams at Toui-Kilibo (Str TK) > Pénéssoulou (PE) > Sudanian RF on plateau (Sud plat) > Okpara River (Okp) > Sudanian RF at Hills feet (Sud Hills) > Idadjo-Bétérou (ID-BE) > Samiondji (SA). Table 4.4: Comparison of the diversity of riparian forest sites SITES Pénéssoulou Idadjo/Bétérou Okpara River Streams at TouiKilibo forest Samiondji (SA) PE>SA (***) Pénéssoulou (PE) Idadjo-Bétérou (IDBE) Okpara River (Okp) Streams at TouiKilibo (Str TK) Sudanian RF on Plateau (Sud plat) ID-BE>SA (***) Okp>SA (***) Str TK>SA (***) PE>ID-BE (***) PE>Okp (***) Str TK>PE (***) Okp>ID-BE Str TK>ID-BE (***) (***) Str TK>Okp (***) - - - Sudanian RF on plateau Sudanian RF at Hill feet Sud plat>SA (***) PE>Sud plat (**) Sud plat>ID-BE (***) Sud Hills>SA (***) PE>Sud Hills (***) Sud Hills>ID-BE (***) Sud plat : Okp (NS) Okp>Sud Hills (***) Str TK>Sud plat (***) Str TK>Sud Hills (***) Sud plat>Sud Hills (***) NS = Not statistically significant at p=0.05, i.e. similar in terms of tree species diversity; (>) More diverse than; ** Significance at 0.01; *** Significance at < 0.001; Details of the calculations are shown in Table 4.5 in the Annex. 4.3.4. Species and families abundance models Species abundance model, which describes the distribution of species richness frequency, shows that not all species are equally abundant (Figure 4.2). 45 % of the tree species have less than 4 individuals. Tree species with 1 or 2 individuals are very common (left part of the curve) while those with larger number of individuals (>10) are uncommon (right part forming a tail). The best abundance distribution model for RF tree species data set is the log series: χ2calculated = 12.15 < χ2table = 15.5 at p = 0.05 df = 8. There is no significant difference, at p = 0.05, between the observed and expected distributions of tree species abundance as predicted by the log series model. 32 Chapter 4: Assessment of riparian forest plant diversity in Benin Number of tree species 40 36 32 28 24 20 16 12 8 4 0 0 100 200 300 400 500 600 700 800 900 1000 1100 1200 Number of individuals per tree species Figure 4.2: Tree species abundance distribution of riparian forests tree species in Benin Also, the distribution of family richness frequency shows that obviously not all families have equal numbers of species (Figure 4.3). Out of 120 families, 33 (i.e. 27.5%) are represented by 616 species (about 62% of the total RF flora) and 55% families have each less than 3 species. Families with 1 or 2 species are very common (left part of the curve), while families with large number of species (i.e. > 10) are fairly uncommon (right part forming a tail). The fitting of models shows that a log series is the best fit for the family richness: χ2calculated = 5.7 < χ2table = 11.07 at p = 0.05 df = 5. We conclude that there is no significant difference, at p = 0.05, between the observed and expected distributions of family richness as predicted by the log series model. 40 36 Number of families 32 28 24 20 16 12 8 4 0 0 10 20 30 40 50 60 70 80 90 Number of species per family Figure 4.3: Family richness abundance distribution of riparian forests in Benin 33 100 Chapter 4: Assessment of riparian forest plant diversity in Benin 4.4. DISCUSSION 4.4.1. Flora of riparian forests In our large-sized data set, we enumerated more species than for the RFs in South Togo (West of Benin Republic), where Kokou et al. (2002) found 499 species grouped into 41 families and 108 genera. Many species (e.g. Pterocarpus santalinoides, Cola laurifolia, Cynometra megalophylla, Manilkara multinervis and Parinari congensis) typical of RFs in Benin are dominant and characteristics of RFs in Southern Côte d’Ivoire (see Devineau, 1975; 1976). Many common species in RFs are found to be encroachments of upland forests. Elaeis guineensis, for example is a heliophytic and pioneer species abundant in all West and central African forests (Sowunmi, 1999; Maley & Chepstow-Lusty, 2001). In the Benin context, as RFs combine plants from various ecosystems, it is likely that they are more diverse than vegetation formations consisting of only one ecosystem. Meave & Kellman (1994) already found that most riparian forest flora in a Neotropical savanna environment of Belize is not specialised to the riparian environment and there is a significant supplement of species having other habitat affinities (i.e. rain forest and non-rain forest taxa). A similar result was found in Northern Australia, where riparian strips contain vegetation dominated by diverse floristic elements unrelated to monsoon rain forests (Woinarski et al., 2000). For Bersier & Meyer (1994), RFs are mosaics composed of patches of different vegetation types, which cover the entire range of forest succession. Thus RFs potentially provide more different categories of food items for a variety of animals than other forest types. Little is known about the relative rarity and threats facing most RF species in Benin. More is known from Khaya senegalensis, Afzelia africana, Milicia excelsa and Antiaris toxicaria (Eyog Matig et al., 2001; Sinsin et al., 2002a; Eyog Matig et al., 2002; Sokpon & Ouinsavi, 2002). The absence of numerous rare, endemic or disjunct plant species in the RFs of Benin, may suggest that these forests have not geographically been isolated for a very long time, instead the historical patterns of migration resulted in an accumulation of various taxa from surrounding vegetations in, and at the edge of RF corridors. Yarpao (Atacora hilly region) and Pénéssoulou reserve forest (semi-deciduous forest region) have the highest plant richness of riparian forests, 358 and 343 species per ha respectively (see Table 4.1). Therefore we might say that protection of the dry semideciduous forests and the difficult accessibility of genuine RFs at hills feet have a positive impact on the conservation of plant diversity. Species richness of RFs (per ha) in Benin varied from 129 to 358. This is within common limits of certain tropical dense forests: 156 in the Taï National Park in Côte d’Ivoire (Dengueadhe, 1999), and 100-300 in South America (Loizeau, 1992). The total vascular plant list for a hectare of forests at Kade (Ghana) is nearly 300 species (Swaine et al., 1987). In a 2.5 ha survey of primary forest in Dja Fauna Reserve (Cameroon) Sonké & Lejoly (1998) found 125 to 138 species. On the contrary, SR per ha of RFs in Benin is far lower than values from several neotropical rain forests (cf. Whitmore, 1990). 4.4.2. Family richness in riparian forests There are some similarities, with regards to dominant families, between the RFs of Benin and various forest types in Benin, Africa and more generally in the tropics. Akoègninou (1984) found 93 families with the Rubiaceae, Papilionoideae, Poaceae, Euphorbiaceae and Apocynaceae as the most pluriform in several dense semi-deciduous forests in Southern Benin. In the humid zones of Southern Benin, Sokpon & Adjakidjè (1999) found 100 families with 364 species, the richest families being the Poaceae, Rubiaceae, Cyperaceae, Fabaceae, 34 Chapter 4: Assessment of riparian forest plant diversity in Benin Euphorbiaceae, Moraceae and Apocynaceae. For these authors, the Leguminosae (Papilionoideae, Mimosoideae and Caesalpinioideae) are often the most species-rich family in forest formations. Likewise, the Cyperaceae are usually abundant in humid zones, while the Poaceae, the second richest family, show the effects of degradation or drier Sudanian climate. In Burkina-Faso (North of the Republic of Benin), the Papilionaceae, Rubiaceae, Caesalpiniaceae and Poaceae were the most species-rich families in gallery forests of the Hippopotamus Pond Biosphere Reserve (Bélem & Guinko, 1998). In the Dja Fauna Reserve (Cameroon), the Euphorbiaceae, Rubiaceae, Annonaceae, Meliaceae, Caesalpiniaceae, Sapindaceae and Sapotaceae have the highest tree species richness (Lejoly et al., 1996; Sonké & Lejoly, 1998). On a 2-ha dense forest at Kade (Ghana) the Leguminosae was the best represented family (Swaine et al., 1987). Meave & Kellman (1994) found the same result in a 1.6 ha of RF in the Mountain Pine Ridge (Belize). According to Whitmore (1990) the three tropical rain forest blocks have abundant Leguminosae (subfamily Caesalpinioideae), Annonaceae, Euphorbiaceae, Lauraceae, Myristicaceae, Rubiaceae and Sapotaceae. For Gentry (1988) three families (Rubiaceae, Annonaceae and Euphorbiaceae) are always among the ten most species-rich families on all continents. Summing up, the Leguminosae, Euphorbiaceae, Rubiaceae and Annonaceae are the most species-rich families always found in Benin RFs and in most tropical forest formations. Meanwhile the Cyperaceae show the influence of humid sites, while the high richness of the Poaceae is indicative of open canopy, degradation of RFs as well as edge effects. 4.4.3. Tree species richness in riparian forest sites Tree species richness per ha varied from 27 to 99 (see Table 4.1), and these values are quite similar to the mean value (52) found by Meave & Kellman (1994) in a RF of Belize. In view of the large interregional variability in tree species richness per ha, figures of RFs in Benin are comparable to values (65-77) of some dense forests in Côte d’Ivoire at Taï National Park and Yapo (Corthay, 1996; Dengueadhe, 1999). Lejoly (1994) and Lejoly et al. (1996) found 92 and 84 tree species per 0.5 ha respectively in Dja Fauna Reserve (Cameroon) and Ngotto forest (Central Africa Republic). Swaine et al. (1987) enumerated 120 tree species in a two 1ha samples of moist semi-deciduous forest at Kade (Ghana). On the other hand, the species richness values found in Benin are lower than those (86-122) of Dogbo, Guiroutou and Djapadji forests in Côte d’Ivoire (Dengueadhe, 1999); 69131 for Gabon forests (Reitsma, 1988); and in some neotropical wet forests (e.g. Bajo Calima, Western Columbia), which are among the most species-rich in the world, with over 250 tree species (Faber-Langendoen & Gentry, 1991). For Whitmore (1990), tree species richness per ha of forests vary from 23 in Nigeria to 283 at Yanamomo (Peruvian Amazon), the richest forest found so far. 4.4.4. Comparison of riparian forest sites diversities A species diversity index, which takes into account the total number of species and their relative abundance offers a good description of communities and allows comparison between them. The overall RFs diversity in Benin (5.8 bits) represents quite a high value among forest formations. This figure is similar to the one (5.4 bits) of Dja Fauna Reserve in Cameroon (Sonké & Lejoly, 1998). Meanwhile it is higher than the values given by (Sokpon et al., 2001) and (Kokou et al., 2002) for certain RFs in Benin and Southern Togo, respectively. Sites along streams have higher diversity values than sites along rivers, and the reason is the lower species richness and uneven distribution of the dominant trees along rivers where 35 Chapter 4: Assessment of riparian forest plant diversity in Benin 4.5% of trees are represented by 58% of total individuals, while along streams 4.8% of trees are represented by only 35.4% of the total abundance. Tree species along streams are more evenly distributed (E = 0.82) than along rivers (E = 0.65). This difference in diversities between streams (smaller waterways) and rivers (larger waterways) is confirmed by the multivariate analysis (DCA and TWINSPAN) of the RFs data set (cf. Natta et al., in press, see Chapter 6). The comparison of tree species diversity between the two RF types (along streams and rivers) is an indirect way of comparing the effect of flood frequency, duration or intensity, and importance of waterways on species richness and abundance. The absence or short periods of inundation along streams allow plants with wider ecological range (forest pioneers and savanna species) to grow. At the same time a specialised (thus less diversified) flora, which supports 4 to 6 months of submersion per year, can only grow along rivers. Therefore at this stage of the investigations on RFs, the duration of submersion or flood seems negatively correlated with tree species diversity. This result substantiates the hypothesis that frequent or long environmental stresses, such as floods along rivers, will lead to a decrease in the diversity of species (at least for tree species), and a loss of diversity will lead to the increased abundance and dominance of those species that can successfully maintain themselves on isolated or harsh sites. Hanson et al. (1990) described a similar phenomenon in RF patches along the Upper mid-West of Iowa river (USA). Meanwhile, as species diversity is sometimes indicative of the diversity, stability and well-being of ecological systems, and used for environmental monitoring (Magurran, 1988; Withmore, 1990; Huston, 1994), we might say that any dense and least disturbed RF is in equilibrium with the prevailing ecological conditions. 4.4.5. Species and families abundance models Log series adequately described tree species and family richness patterns of the RFs data set. In an assessment of forest plant diversity in Nepal, Acharya (1999) found that the log series describes species abundance pattern better than the geometric series, meanwhile the fit was not so good for plots with low disturbance. Also a gradual decrease in disturbance was well explained with the progression of species abundance models from the geometric series to a log series. From a 50-ha forest plot on Barro Colorado Island (Panama), the pattern of total species abundance deviates from the log-normal in having too many rare species (Hubbell & Foster, 1986). In the Peruvian Amazon, only 15% of species had more than 2 individuals and 63 % had only one individual (Whitmore, 1990). Log series, that are intermediate between geometric series and log normal series, often characterise communities in which species of intermediate abundance become more common. Theoretically, the geometric series provides the best fit to the observed species abundance in species-poor communities under a harsh environment, while in large species-rich or mature communities the distribution of species abundance is usually log normal. As succession proceeds, or as conditions ameliorate, species abundance patterns grade from geometric series into those of the log series (Whittaker, 1972; Magurran, 1988). Therefore RFs, under the Benin conditions, are made of moderately species-rich or mature communities in which families with intermediate species richness and tree species with intermediate abundance are fairly common. In large assemblages (as it is for our RFs data set), empirical patterns predict log normal series. However, new insights raised by Magurran & Henderson (2003), suggest that an ecological community can be separated into two components: 1 - core (i.e. typical) species, which are persistent, abundant, and grow naturally under good/optimal ecological conditions. They are log normally distributed. 36 Chapter 4: Assessment of riparian forest plant diversity in Benin 2 – occasional (i.e. non-typical) species, which occur infrequently, have low abundance and different habitat requirements. They follow a log series distribution. 4.5. CONCLUSIONS This paper has made an attempt to assess the flora, plant species diversity, as well as species abundance models that best fit RFs in Benin. The flora of RF contains about 1/3 of all species in Benin, 120 families and 513 genera. The Leguminosae, Euphorbiaceae, Rubiaceae and Annonaceae are the most species-rich families found time and again in Benin’s RFs as well as in tropical rain forests. RFs can be recognised and characterised by exclusive, frequent and species typical to certain regions. Endemism is low compared to rain forests, yet in fire-prone environment the RFs of Benin harbour plant species that prefer the wettest conditions. They are a refugee ecosystem for forest plants, and act as route or corridor for movement of certain species across the landscape. As RFs along streams have higher tree species diversity than those along rivers, we conclude that flood frequency, duration and intensity are somehow negatively correlated with tree flora diversity. A challenge is to assess the spatio-temporal effects of flood variables on plant species diversity. This study showed that relatively large numbers of species still are maintained in small forest fragments along waterways. Although RFs are often strip-like and fragmented ecosystems, their rich and varied flora invites protection and restoration. Also this phytodiversity, should be used in selecting the most suitable plant species for the rehabilitation of degraded RF corridors, in each ecological region of the country. This will allows them to fulfil their vital ecological, socio-economic and cultural functions. 37 Chapter 4: Assessment of riparian forest plant diversity in Benin ANNEX Table 4.5: Details of the comparison of tree species diversity between seven riparian forests sites Comparison of riparian forest sites H'1 H'2 df t calculated t-table Significance Hypothesis accepted Pénéssoulou (40 relevés) / Samiondji (36 relevés) Idadjo-Bétérou (40 relevés) / Samiondji (36 relevés) Okpara (56 relevés) / Samiondji (36 relevés) Streams Toui-Kilibo (62 relevés) / Samiondji (36 relevés) Sudanian RF on plateau (43 relevés) / Samiondji (36 relevés) Sudanian RF at foot of hills (27 relevés) / Samiondji (36 relevés) Idadjo-Bétérou / Pénéssoulou Okpara/Pénéssoulou 5.17 3.91 4.91 5.4 2.43 2.43 2.43 2.43 2392 2679 2186 2211 34.08 18.82 32.36 40.48 p=0.001; t=3.29 p=0.001; t=3.29 p=0.001; t=3.29 p=0.001; t=3.29 *** *** *** *** H1 H1 H1 H1 4.97 2.43 2460 32.76 p=0.001; t=3.29 *** H1 4.27 2.43 2210 21.96 p=0.001; t=3.29 *** H1 3.91 4.91 5.17 5.17 2498 1969 17.09 3.68 p=0.001; t=3.29 p=0.001; t=3.29 *** *** H1 H1 Streams Toui-Kilibo(TK streams) / Pénéssoulou 5.4 5.17 2015 3.39 p=0.001; t=3.29 *** H1 Sudanian RF on plateau / Pénéssoulou 4.97 5.17 2263 2.79 p=0.01; t=2.57 ** H1 Sudanian RF at foot of hills / Pénéssoulou 4.27 5.17 1980 10.85 p=0.001; t=3.29 *** H1 Okpara / Idadjo-Bétérou 4.91 3.91 2318 14.34 p=0.001; t=3.29 *** H1 TK streams / Idadjo-Bétérou 5.4 3.91 2538 22.58 p=0.001; t=3.29 *** H1 Sudanian RF on plateau / Idadjo-Bétérou 4.97 3.91 2747 15 p=0.001; t=3.29 *** H1 Sudanian RF at foot of hills / Idadjo-Bétérou 4.27 3.91 2105 4.43 p=0.001; t=3.29 *** H1 TK Streams / Okpara 5.4 4.91 1826 7.78 p=0.001; t=3.29 *** H1 Sudanian RF on plateau / Okpara 4.97 4.91 2071 0.88 p=0.2; NS H0 t=1.28 Sudanian RF at foot of Hills / Okpara 4.27 4.91 1778 8.03 p=0.001; t=3.29 *** H1 Sudanian RF on plateau / TK streams 4.97 5.4 2224 6.72 p=0.001; t=3.29 *** H1 Sudanian RF at foot of Hills / TK streams 4.27 5.4 1722 14.84 p=0.001; t=3.29 *** H1 4.27 1954 8.7 p=0.001; t=3.29 *** H1 Sudanian RF on plateau / Sudanian RF at foot of Hills 4.97 Conclusion PE>SA: Pénéssoulou is more diverse than Samiondji Id-Be>SA: Idadjo-Bétérou is more diverse than Samiondji Okp>SA: Okpara is more diverse than Samiondji Tk>SA: Streams of Toui-Kilibo forest are more diverse than Samiondji Sud Plat>SA: Sudanian RF on plateau are more diverse than Samiondji Sud Hill>SA: Sudanian RF at foot of hills are more diverse than Samiondji PE>Id-Be: Pénéssoulou is more diverse than Idadjo-Bétérou PE>Okpara: Pénéssoulou is more diverse than Okpara TK Streams>PE: Streams at Toui-Kilibo are more diverse than Pénéssoulou PE>Sud Plat: RF of Pénéssoulou are more diverse than Sudanian RF on plateau PE>Sud Hill: Pénéssoulou are more diverse than Sudanian RF at foot of hills Idadjo-Bétérou>Okpara: Idadjo-Bétérou is more diverse than Okpara TK>Idadjo-Bétérou : Streams of Toui-Kilibo forest are more diverse than Idadjo-Bétérou Sud RF on plateau : Sud RF on plateau are more diverse than Idadjo-Bétérou TK>Idadjo-Bétérou : Streams of Toui-Kilibo forest are more diverse than Idadjo-Bétérou TK>Okpara : Streams of Toui-Kilibo forest are more diverse than Okpara Sudanian RF on plateau and RF along Okpara are similar in terms of tree species diversity Okpara>Sudanian RF at foot of Hills: Okpara is more diverse than Sudanian RF at foot of hills Streams TK>Sudanian RF on plateau: TK streams are more diverse than Sudanian RF on plateau Streams TK>Sudanian RF on plateau: TK streams are more diverse than Sudanian RF on plateau Sud RF plateau>Sud RF Hills: Sud RF on plateau are more diverse than Sud RF foot of Hills NS = Non Statistically significant at p=0.05 ; ** Significant at 0.01 ; *** Significant at < 0.001 ; H’1& H’2 = Shannon indices. 38 Chapter 5 STRUCTURE AND ECOLOGICAL SPECTRA OF RIPARIAN FORESTS IN BENIN A.K. Natta Chapter 5: Structure and ecological spectra of riparian forests in Benin Chapter 5 STRUCTURE AND ECOLOGICAL SPECTRA OF RIPARIAN FORESTS IN BENIN Natta A.K. (1) (1) Department of Environment Management, Faculty of Agronomic Sciences. FSA/UAC 01 BP 526 Cotonou, Benin; aknatta@yahoo.com. Abstract This paper is a preliminary investigation of the structural characteristics of riparian forests in Benin. So far such information is not available in Benin. The structure was analysed through different methods: life forms, phytogeographic types, diameter class distribution, basal area, and stem density. Riparian forests displayed a physiognomy that is highly variable in terms of vertical stratification. Riparian forests in Benin were similar to many dense tropical forests, and to the West African ones in particular, in terms of phytogeographical types (large contribution of Guineo-Congolian basin species), life forms (high abundance of therophytes and medium size trees, low percentage of mega-phanerophytes and woody lianas), diameter class distribution (reverse J type), basal area (23-59 m2/ha) and stem density (253 to 785 trees • 10 cm dbh /ha). Species dominance, in terms of abundance and basal area, was a major characteristic of riparian forests in Benin. Many stands of riparian forests are facing various levels of structural and floristic simplification, which include fundamental transformations in vegetation physiognomy, from dense pristine stands to riparian scrub, or bare land. Key words: riparian forests, structure, life forms, phytogeographic types, diameter, basal area, stem density, species dominance, Benin. 5.1. INTRODUCTION Pattern in vegetation structure is a four-dimensional phenomenon. It can be studied at different scales, in space and time. Generally, vegetation structure is concerned with distribution of individuals along the vertical and horizontal axes, as well as spatial arrangement of physiognomical, taxonomic, morphological and functional characteristics of the elements building the vegetation (Popma et al. 1988, Sterck & Bongers 2001). Various methods have been used to assess the structure of tropical forests (Whitmore 1990), among which riparian forests are a major component. In savanna regions, several authors have documented the structural characteristics of riparian forests (abbreviated as RFs), which contrast strongly with the open forest and savanna types which otherwise dominate the landscape (Adjanohoun 1965, Monnier 1990, Woinarski et al. 2000). However, in the pre-forest zone of Côte d’Ivoire (West Africa), the physiognomy of riparian forests is very much similar to the one of dense semi-deciduous forest, with which they share similar features such as presence of buttresses, cauliflory, presence of numerous epiphytes and lianas (Devineau 1975, 1976, 1984). The floristic composition and plant community diversities of riparian forests in Benin have been studied (cf. Natta & van der Maesen, see Chapter 04), but an account of the general structure is lacking. The scale of the study (19 ha of riparian forests) and sites conditions (various climatic conditions, floristic compositions and successional states), imply 41 Chapter 5: Structure and ecological spectra of riparian forests in Benin that the selected structural parameters should give a general picture of riparian forests in Benin. Life forms are frequently used as a means of ecological characterisation of vegetation formations or plant communities, while the geographic affinity of a flora contains information about evolution patterns of species within the region of occurrence. Also basal area, stem density and diameter class distribution are intrinsic characteristics of a given plant community, and are often used to compare vegetation types. This paper investigates the structural characteristics of riparian forests in Benin, in terms of life forms, phytogeographic types, stem density and basal area. Two research questions will be answered: 1 - What are the most prominent life forms and phytogeographic types of riparian forests, and what are the implications of their occurrence in Benin? 2 - What are the average basal area and stem density of riparian forest sites in Benin, and are these figures comparable to those from other tropical forests? 5.2. METHODS The study was conducted from 7º10 to 12º20 N in Benin (West Africa). About 19 ha of the least disturbed forest stands along rivers and streams were surveyed in the Guinean region (Samiondji and Bètèkoukou), Sudano-Guinean zone (Toui-Kilibo, Idadjo, Bétérou and Pénéssoulou) and Sudanian region (Yarpao, Gbèssè, Pendjari Biosphere Reserve), see Figure 4.1. In total 373 rectangular plots (each of 500 m2), were installed in several RF sites using the Braun-Blanquet phytosociological method of floristic data collection. Tree height and dbh were measured in 304 plots (i.e. dendrometric relevés, see Table 5.1). Species names follow Keay & Hepper (1954-1972), Brunel et al. (1984), Berhaut (1967) and Lebrun & Stork (1991-1997). Life forms followed Raunkaier (1934), Schnell (1971) and Keay & Hepper (19541972). They were: Phanerophytes (Ph) subdivided into mega-phanerophyte (MPh > 30 m), meso-phanerophyte (mPh: 8 to 29 m), micro-phanerophyte (mph: 2 to 7 m) and nanophanerophyte (nph < 2m); Therophytes (Th); Hemicrytophytes (He); Chamaephytes (Ch); Lianas (L); Geophytes (Ge); Epiphytes (Ep) and Parasites (Par). The phytogeographic types were named after White (1986) and Keay & Hepper (1954-1972). They were: - Species widely distributed in the tropics (Cosmopolitan-Cosmo, Pantropical-Pan, AfroAmerican-AA and Paleotropical-Paleo). - Species widely distributed in Africa (Tropical Africa-TA, Pluri Regional in Africa-PRA) - Regional species in Africa (Sudanian-S, Guinean-G, Sudano-Guinean-SG, SudanoZambesian-SZ, Guineo-Congolian-GC). The family, life form and phytogeographic affinity of each plant species were added to the species list. Stem diameter at breast height (dbh) was measured for trees (i.e. dbh ≥ 10 cm). Several variables were measured in each plot and on individual trees. Plot variables were: centre co-ordinates, riparian forest width, species richness and abundance. We plotted frequency distribution curves for the life forms, phytogeographic types, diameter class, and calculated basal areas and tree densities of riparian forest sites per ha. 5.3. RESULTS Floristic and stand characteristics of riparian forests throughout Benin In Benin, riparian forests vary in their width along the watercourses, which they fringe. Some are no more than a few meters wide and have an open canopy, while others are much more 42 Chapter 5: Structure and ecological spectra of riparian forests in Benin extensive and have a closed canopy and a well-developed structure. The floristic and stand characteristics of riparian forest in Benin are presented in Table 5.1. Table 5.1: Floristic and stand characteristics of riparian forests throughout Benin Regions Sites Floristic Dendrometric SR/site relevés (n1) relevés (n2) Guinean Samiondji 38 36 Sudano Guinean Pénéssoulou Idadjo-Bétérou Okpara Streams at Toui- Kilibo 41 48 56 78 40 40 56 62 All Sudano-Guinean sites 223 198 556 RFs on plateau RFs at hill foot 76 36 43 27 All Sudanian sites 112 All riparian forest sites in Benin 373 Sudanian 306 SR/ha SR of trees Trees Basal area / site density/ha (BA/ha) 249 27 726 41.7±17.4* 99 34-54 59 78 544 748-785 315 257 38.9±15.3 45.6±15.8 - 129-343 34-99 257-785 42.6±15.8 435 467 182-278 358 36-44 65 538-627 732 35.6±16.5 42.7±9.9 70 591 182-358 36-65 538-732 40±20.6 304 1003 129-358 224 253-785 41.5±17.9 450 343 183-195 129-195 - n1 = number of plots where floristic data were collected using the phytosociological approach; n2 = number of plots where data were collected on individual trees; SR = Species richness per site (SR/site) or per ha (SR/ha); BA/ha = Basal area per ha in m2; Trees = stems with dbh ≥ 10 cm; (*) ± Standard Deviation; (-) data not available. Life forms variability and frequency distribution Phanerophytes (Ph), lianas (L) and therophytes (Th) are the richest life forms (Figure 5.1a). Field data confirm that trees are on average 14 to 18 m tall, with occasional heights of 20 to 25 m. If present, layers (i.e. vertical distribution of tree crowns) occur at variable heights, and are stand dependent. In the least disturbed stands, the top layer of uneven and discontinuous crowns is composed of a few emergent trees such as Ceiba pentandra, Parinari congensis and Cynometra megalophylla. Within phanerophytes, medium size trees (meso and microphanerophytes) are more frequent (75.6 %) than mega-phanerophytes (3 %) (Figure 5.1b). Also, liana richness in the Benin RFs is high (236 species, i.e. 23.6% of the overall RF species). Meanwhile, woody lianas (LMPh and LmPh, see Figure 5.1c) are only represented by 21 species (i.e. 8.9 % of the liana richness and 2 % of the total RF flora). Among the six epiphyte species (0.6% of RF richness), the most frequent are the stem-tuber epiphytes Calyptrochilum christyanum and C. emarginatum. Less frequent is the humus-collecting epiphytic fern with clustered roots, Platycerium angolense. 43 Chapter 5: Structure and ecological spectra of riparian forests in Benin Species richness 35 0 30 0 25 0 20 0 15 0 10 0 50 0 Ph L Th Ch He Ge Ep Pa r Life fo r m s Species Richness (a): Overall life forms frequency distribution. Ph = Phanerophytes; L = Lianas; Th = Therophytes; Ch = Chamephytes; He = Hemicryptophytes; Ge = Geophytes; Ep = Epiphytes; Par = Parasites. 150 120 90 60 30 0 MPh m Ph m ph nph Ph an e r o p h yte life fo rms (b): Phanerophytes life forms frequency distribution. MPh = Mega-phanerophytes; mPh = mesophanerophytes; mph = micro-phanerophytes; nph = nano-phanerophytes. Species richness 100 80 60 40 20 0 LMPh Lm P h Lm p h Ln ph Lge Lth Lian a life fo r ms (c): Liana life forms frequency distribution in riparian forests. LMPh = Megaphanerophyte lianas; LmPh=mesophanerophyte lianas; Lmph=microphanerophyte lianas; Lnph=nanophanerophyte lianas. Figure 5.1: Life form frequency distribution in riparian forests 44 Chapter 5: Structure and ecological spectra of riparian forests in Benin Phytogeographical affinity of riparian forest species Beside the large contribution of species from the Guineo-Congolian basin (33 % for GuineoCongolian and Guinean types), species of wide distribution all over the tropics (Pantropical 16.7 %), and in tropical Africa (13.6 %) in particular, are the richest phytogeographic types of riparian forests in Benin (Figure 5.2). 250 Species richness 200 150 100 Cosmopolitan Sudanian Afro-American Paleo-Tropical Sudano-Zambesian Sudano-Guinean Guinean Tropical Africa Pantropical Guineo-Congolian 0 Pluri Regional in Africa 50 Phytogeographic types Figure 5.2: Phytogeographic types frequency distribution in riparian forests Diameter class distribution The size-class distribution (or stand table) of trees with a dbh ≥10 cm is drawn from 7765 stems. Stems with dbh < 50 cm are on the left part while bigger tree with dbh > 70 cm form the tail at the right part of the curve (Figure 5.3). Here higher abundance is found in the smaller size classes with a more or less logarithmic decline in numbers with increasing size. 76.4% of stems have less than 30 cm dbh. 45 Chapter 5: Structure and ecological spectra of riparian forests in Benin 2500 2000 1500 1000 180 to 189 160 to 169 140 to 149 120 to 129 100 to 109 90 to 94 80 to 84 70 to 74 60 to 64 50 to 54 40 to 44 10 to 14 0 30 to 34 500 20 to 24 Abundance per DBH class 3000 DBH class (cm) Figure 5.3: Diameter class distribution for riparian forest trees (dbh ≥ 10 cm) in Benin Species dominance in riparian forests Nine (i.e. 4%) tree species out of 224 contribute 46.3 % of the total abundance and 55.3 % of the total basal area. They are Pterocarpus santalinoides, Cola laurifolia, Syzygium guineense, Dialium guineense, Berlinia grandiflora, Cynometra megalophylla, Elaeis guineensis, Diospyros mespiliformis and Uapaca togoensis. 5.4. DISCUSSION Life forms variability The life forms found in riparian forests are similar to many riparian forest sites in tropical Africa, in general and the savanna zones in particular. Akoègninou (1984) found similar ranking of life forms types in several dense semi-deciduous forests of Southern Benin. In gallery forests of the Hippopotamus Pond Biosphere Reserve (Burkina-Faso), phanerophytes and therophytes accounted for 64.8 % and 13.6 % of the total species richness respectively (Bélem & Guinko 1998). Phanerophytes are known to be the major life form in forest ecosystems, but in Benin’s RFs the overall percentage is lower (56 %) than in other forest formations (e.g. 80-90 % given by Mangenot (1955) for dense tropical forests). Also at Lamto (South Côte d’Ivoire), Devineau (1984) found that RFs have more medium-size trees (meso- and micro-phanerophytes) than large ones (i.e. mega-phanerophytes). Along the Tana river (Kenya), riparian forests have a disturbed physiognomy characterised by a low mean height (14.4 m) (Medley 1992). In contrast, in tropical rain forests canopy tree heights fluctuate between 30 to 35 m and emergent trees may reach heights above 40 m (Popma et al. 1988, Whitmore 1990). In Southern Togo, Kokou et al. (2002) found that gallery forests supported the greatest liana richness with 150 species (i.e. 30.8 % of the overall gallery forest richness), and the highest Shannon diversity index. In riparian forests along the Tana river (Kenya), 27% of the woody species are lianas (Medley 1992). Akoègninou (1984) found 150 liana species (i.e. 26.1% of the plant richness) in dense semi-deciduous forests in Southern Benin. In several rain forests in Côte d’Ivoire, the density of lianas was negatively correlated with the age of 46 Chapter 5: Structure and ecological spectra of riparian forests in Benin logged areas, and it ultimately reaches the maximum value of undisturbed patches (Kouamé & Traoré 2001). Lianas (woody climbers) and vines (herbaceous climbers) are an intrinsic and important component of tropical forests (Hall & Swaine 1981, Hegarty & Caballé 1991, Kokou et al. 2002) and are considered a major component for forest reconstitution (Kouamé & Traoré 2001, Parren 2003). Degradation (e.g. selective tree cutting or pruning, tree fall), lateral lighting as well as the multiplication of edge effects and ecotones through fragmentation, are favourable conditions for high abundance of lianas in RF. The high abundance of vines and climbing shrubs (Lmph, Lnph, Lge and Lth, see Figure 5.1c) over woody lianas (LMPh) in RFs is indicative of relatively low forests having an irregular canopy (Medley 1992). On the other hand, the high species richness of therophytes (Figure 5.1a) indicates that many RFs in Benin are either degraded or influenced by neighbouring open plant communities. Like woody lianas, epiphytes reach their fullest development in humid tropical forests (Braun-Blanquet 1972). Phytogeographical affinity of riparian forest species Guinean species (Guineo-Congolian and typical Guinean) are abundant in dense semideciduous forests in Southern Benin (Akoègninou 1984), and in riparian forests at Bantè (Central Benin) (Akoègninou et al. 2001). In several gallery forests at Lamto (South Côte d’Ivoire), Devineau (1975) found 70 to 75% of Guineo-Congolian species against 15% of Sudano-Zambesian distribution. In gallery forests of the Hippopotamus Pond Biosphere Reserve (Burkina-Faso), Guineo-Congolian and Sudano-Zambesian species accounted for 61.7% and 38.3% of the total plant richness respectively (Bélem & Guinko 1998). This justifies the classification of RFs among humid vegetations. The floristic affinity towards the Guineo-Congolian rain forests is indicative for supply of water. The continuous and adequate availability of ground water and shallow water table, which originate from waterways, allow for the establishment and long-term existence of these species. The geographical affinities (Guineo-Congolian, Pantropical and Sudanian species) of RFs from the West-African savanna regions support three conclusions: (a) The presence of numerous pan-african species with heavy seeds and fruits (e.g. Diospyros mespiliformis, Oncoba spinosa) suggest that the dense or evergreen forests in Africa were once continuous (Medley 1992); (b) the high contribution of Guineo-Congolian basin species substantiates the thesis of an early period of continuous forest block in West and Central Africa (see also Akoègninou 1984, Tossou 2002); (c) on the other hand, the incursion of savanna taxa and Sudano-Guinean transitional zone species (11.3 and 9.4 %, respectively) in RFs indicates the lower water balance along waterways in the dry season, particularly in North Benin. Diameter class distribution Similar results in diameter class distribution were found in RFs of Belize (Meave & Kellman 1994), where about 78% of trees were in the smallest dbh class (10-20 cm). Following these authors we might characterise RFs in Benin as low-biomass community with many smallstemmed trees, compared to the high-biomass continuous upland tropical rain forests. The decreasing curve has a reverse-J (or J-shaped, see Figure 5.3) distribution typical of natural forest regenerating from seed. It suggests a stable size and age class distribution (Swaine et al. 1987). This shape is characteristic of climax species (Whitmore 1990). For Faber-Langendoen & Gentry (1991) this distribution type describes a mature stand, with many small individuals and few large ones. We hypothesise that the most abundant species among the flood-tolerant species of RFs are certainly at a stage of equilibrium with the climatic and edaphic conditions. 47 Chapter 5: Structure and ecological spectra of riparian forests in Benin Basal area and stem density of riparian forests The overall mean basal area (± S.D.) of RFs in Benin, obtained from 7765 stems (dbh ≥10 cm), was 41.5 ± 17.9 m2 ha-1 and the mean density 586 ± 192 stems ha-1 (see Table 5.1). Sokpon et al. (2001) found a similar range for stem density ha-1 (312 to 665) for some edaphic forests in Benin with basal area ranging from 24.8 to 41 m2 ha-1. Goudiaby (1998) found a density of 507 trees/ha in a gallery forest of South East Senegal. Along the Tana river (Kenya), riparian forests have a low tree density and coverage, 409 stems ha-1 and 23.1 m2/ha respectively (Medley 1992). The mean tree density of a 1.6 ha RF at the Mountain Pine Ridge (Belize) was 766 ± 241 stems ha-1 (Meave & Kellman 1994) that exceeds the values from Benin, obviously the mean basal area (21.9 ± 8.8 m2 ha-1) in Belize is far lower. Values of basal area of RFs in Benin (23-59 m2/ha) are similar to those of some West African upland dense forests (Taï, Dogbo, Guiroutou, Djapadji) in Côte d’Ivoire with 25-59 m2/ha (Dengueadhe 1999) and humid forests in Ghana (25.5 to 33 m2/ha) (Hall & Swaine 1976, Swaine et al. 1987). In Gabonese dense forests, Hladik (1982) and Reitsma (1988) found 35 and 35.7-42.9 m2/ha respectively. In Cameroon, Sonké & Lejoly (1998) found a mean value of 30.5 m2/ha in Dja fauna reserve and Gartlan et al. (1986) found 27.6 m2/ha in the Korup forest reserve. At Parana high forest (Paraguay) the mean basal area is 39.6 m2/ha (Stutz De Ortega 1987), while trees in the Los Tuxtlas rain forest plot (Mexico) had an average basal area of 34.9 m2/ha (Bongers et al. 1988). For Gentry (1988) values range from 35 to 45 m2/ha in Neotropical and Asian forests. Basal area figures found in Benin fit well with those given by Mori & Boom (1987) for tropical forests (21.3 - 53 m2/ha). Tree densities of RFs in Benin are similar to those in Taï, Dogbo, Guiroutou, Djapadji and Yapo upland forests in Côte d’Ivoire with 376–649 stems/ha (Corthay 1996, Dengueadhe 1999); while Swaine et al. (1987) obtained 552 ± 13 trees/ha in a 2-ha moist semi-deciduous forest at Kade (Ghana). In the Dja fauna reserve (Cameroon), Lejoly et al. (1996) and Sonké & Lejoly (1998) found respectively 603 and 461 trees/ha. The mean tree density was 471 ha-1 in the Korup reserve forest, Cameroon (Gartlan et al. 1986). Forest structure at the Los Tuxtlas rain forest (Mexico) was characterised by a low density, i.e. 346 trees/ha (Bongers et al. 1988). Whitmore (1990) found a density of 580 stems/ha in Peru. For Neotropical forests the values ranged from 167 to 1947 trees/ha (Gentry 1982). Studies of many riparian forest fragments from Belize and Venezuela have shown that they contain tree species densities comparable to continuous forests, and a non-specialised tree flora comprised of species characteristic of continuous forest (Meave et al. 1991). Species dominance in riparian forests The phenomenon of species dominance, in which one or a few species contribute much to the total abundance and basal area, is not new in riparian systems as well as tropical forests in general. Along the Tana river (Kenya), riparian forests were characterised by a high species importance attributable to a few trees (Medley 1992). In a RF of Belize, Meave & Kellman (1994) found that the seven most abundant species (2.4% of the total species richness) account for 1/3 of the total number of stems. According to Wolter (1993) cited by Sonké & Lejoly (1998), dominant species can comprise up to 58 % of the total individuals and 75 % of basal area. Tropical forests dominated by a few species have been considered to result from edaphic conditions unfavourable for a large majority of species (Richards 1952, Hartshorn 1988) such as poor nutrient status, presence of some minerals in toxic concentrations, and temporary or permanent flooding (Martijena & Bullock 1994, Martijena 1998). The latter 48 Chapter 5: Structure and ecological spectra of riparian forests in Benin factor seems more important in the present case, and this indicates the superior adaptation of such dominant tree species to a regime with vertical and lateral harsh disturbance. Therefore these species can be termed ‘edaphic and climatic climax species’ of RFs in the prevailing ecological conditions of Benin. They are keystone species and intrinsic components in the functioning of complex and self-organising RF systems. Moreover, they are essential for ecosystem resilience (i.e. resistance to disturbance and speed to return to a stable equilibrium) (Folke et al. 1996). Process of riparian forests degradation The impact of natural events and human activities on the structure and functioning of riparian areas in Benin include changes in the hydrology of waterways, alteration of the geomorphic structure of riparian areas, and the removal of riparian vegetation. From field observations, we have recognised three major non-successive phases of riparian forests degradation: a) Structural simplification of dense and undisturbed stands, which is caused either by selective removal of big stems, destruction of the understorey (e.g. due to domestic livestock grazing or fire in the savanna region), or localised landslides that induce canopy opening and gaps at lower heights. b) Floristic richness depletion, which involves constant and abusive harvest (e.g. systematic cutting, collection of all fruits or seeds of certain valuable plants), reduction in stands size (too frequent fires, shifting cultivation, dam or bridge construction), etc. c) Complete disappearance consecutive to severe or irreversible transformation of the vegetation to other land and water-use (e.g. farm, dam, road, bridge construction, housing, etc.). 5.5. CONCLUSIONS Riparian forests in Benin display a physiognomy (i.e. canopy height and vertical stratification) that is highly variable, though the understorey is generally dense. Tree size (stem diameter and height) is generally small. Most RF sites in Benin are degraded, highly influenced by adjacent more open ecosystems (edge effect) and characterised by open canopies. This is shown by a high amount of grasses (particularly annual species), a relatively low percentage of phanerophytes (particularly mega-phanerophytes and woody lianas as compared to tropical dense forests), and a high abundance of medium size and small trees. Therefore they can be termed as relatively low hygrophile forests with irregular canopy. The most prominent phytogeographic types of riparian flora range from the GuineoCongolian ones to species of wide distribution all over the tropics, and in tropical Africa. This substantiates the hypotheses of an early period of continuous dense forest block in West and Central Africa, and the influence of edge effects. In terms of diameter class distribution, basal area and stem density RFs in Benin are very similar to many tropical dense forests and in particular to the West African ones. The phenomenon of species dominance, whereby few species contribute much to the total abundance and basal area, is also a major characteristic of RFs in Benin. Many riparian forests stands are facing various levels of structural and floristic simplification. This includes fundamental transformations in vegetation physiognomy, from dense pristine riparian forests to riparian scrub, or bare land. Action should be taken to overcome these threats to biodiversity by maintaining the existing legal protection and/or durable exploitation. 49 Chapter 6 A PHYTOSOCIOLOGICAL STUDY OF RIPARIAN FORESTS IN BENIN (WEST AFRICA) In press, Belgian Journal of Botany NATTA A.K., SINSIN B. and VAN DER MAESEN L.J.G. Chapter 6: A phytosociological study of riparian forests in Benin (West Africa) Chapter 6 A PHYTOSOCIOLOGICAL STUDY OF RIPARIAN FORESTS IN BENIN (WEST AFRICA)* NATTA A.K. (1), SINSIN B. (2) and VAN DER MAESEN L.J.G. (3) (1) (2) (3) Department of Environment Management, Faculty of Agronomic Sciences. University of Abomey-Calavi FSA/DAGE/UAC 01 BP 526 Cotonou Benin; aknatta@yahoo.com; natta@bj.refer.org Department of Environment Management, Faculty of Agronomic Sciences. University of Abomey-Calavi FSA/DAGE/UAC 01 BP 526 Cotonou Benin; bsinsin@bj.refer.org Biosystematics group, Building no. 351, Generaal Foulkesweg 37 6703 BL, Wageningen University, the Netherlands; jos.vandermaesen@wur.nl. (*) In press, Belgian Journal of Botany Abstract. - Floristic ordination and classification of riparian forests in Benin were derived from a comprehensive floristic inventory. TWINSPAN classification and DCA analysis of a data set of 818 plant species and 180 relevés yielded 12 plant communities. Importance of waterways, relief, topography, latitude and longitude were the five major environmental gradients that best differentiated riparian plant communities. A syntaxonomic classification of the identified riparian forests plant communities is presented. Riparian forests in Benin belong to the Mitragynetea Schmitz 1963, which is the phytosociological class of hygrophile fresh water forests of tropical Africa. Based on similarities of ecological conditions and floristic composition, we classified the 12 plant communities into 3 orders: Alchornetalia cordifoliae Lebrun 1947, Lanneo-Pseudospondietalia Lebrun & Gilbert 1954 and Pterygotetalia Lebrun & Gilbert 1954. Key words: Riparian forests, classification, ordination, syntaxonomy, Benin. 6.1. INTRODUCTION The recognition and definition of natural groupings of plant species in vegetation formations have always been a challenging field of study for many researchers. Phytosociology is concerned with the detection and characterisation of distinct plant assemblages as social units (groups of similar relevés), which repeat themselves over space (BRAUN-BLANQUET 1972, KENT & COKER 1992). Some authors have shown the applicability of the phytosociological approach in forests of the humid tropics (HOMMEL 1990). Plant communities provide a far better means of analysing the relation between vegetation and climate, edaphic, topographic and human factors (LAWSON et al. 1970, GARTLAN et al. 1986, RUSSEL-SMITH 1991) than the physiognomic approach does (HOMMEL 1990). Multivariate methods for vegetation classification and ordination have been applied to tropical West African forest types (SWAINE & HALL 1976, GARTLAN et al. 1986, SOKPON 1995, KOKOU 1998). These methods have shown to be useful in extracting meaningful gradients and vegetation types or plant communities in tropical humid forests, which seems too complex for satisfactory analysis by more traditional methods (SWAINE & HALL 1976). In Benin, apart from the work of SOKPON et al. (2001), no attempt was made so far in the field 53 Chapter 6: A phytosociological study of riparian forests in Benin (West Africa) of the phytosociology of riparian vegetation in general, and in particular riparian forests (hereafter abbreviated as RFs). Control over internal structure of riparian ecosystems, in particular vegetation, has been attributed to fluvial dynamics, frequent perturbations (e.g. flooding) and soil moisture (MEDLEY 1992, THOMAS 1996, PIEGAY 1997, LEINARD et al. 1999), all of which are processes related to topography (HANSON et al. 1990). In this paper we report the results of a multivariate analysis of RF based on comprehensive floristic data collected throughout Benin, a country located at the discontinuity of the tropical West African rain forest block. The aim of the paper is to assess whether meaningful ecological factors and environmental gradients can be derived from numerical analysis of RF samples and to provide a floristic classification (i.e. a typology of RF plant communities). A related aim is to document the environmental relations of the identified plant communities and their characteristic species. 6.2. MATERIAL & METHODS 6.2.1. Study area All types of ecological regions of Benin, excluding the coast, beaches, coastal lagoons, lakes and marshlands, were surveyed. The study area (Figure 6.1) covered the latitudes 7º10 to 12º20 N. Representative sites along rivers and streams were surveyed in the Guinean region (Samiondji and Bétékoukou), the Sudano-Guinean zone (Toui-Kilibo, Ouèssè, Bétérou, Onklou-Daringa and Bassila-Pénéssoulou) and the Sudanian region (Ouaké, Affon, Natitingou, Koussoukouingou, Toukountouna, Pendjari Biosphere Reserve, Kandi, Ségbana and Malanville). 6.2.2. Phytosociological relevés in riparian forests Floristic data were collected in 12 riparian forests sites (Table 6.1) using the BRAUNBLANQUET (1972) method for vegetation analysis. Rectangular plots, each of 500 m2, with variable length and width, were preferred for practical reasons to fit with any shape of the waterway and the structural uniformity (physiognomy of the RF). When all terrestrial plants in all stages are included, many authors found smaller plot sizes (500 to 1000 m2) acceptable to allow the classification of forest plot data by means of tabular comparison (HOMMEL 1990). Field plots (50m by 10m) were established in fragmented forests in Southern Togo (KOKOU et al. 2002). Rectangular plots of 500 or 1000 m2 were used to collect floristic data in several edaphic forests in Benin (SOKPON et al. 2001). In the strip-like forest fragments that are RF, homogeneous plots of 500 m2, with varying length and width, seem to be of a reasonable size for the numerical ordination and classification of phytosociological relevés (see also Chapter 8). At each site plots were established, in homogeneous stands, at about 100 m from each other, from a random starting point. In each plot all vascular plants were recorded and specimens taken for the herbarium. As we are dealing with a particular vegetation formation defined by certain hydrographic, soil and topographic characters, relevés taken into account in the present paper were selected from a larger data set with the greatest possible uniformity or homogeneity in regard to their physiognomy, species composition, and dominant species. Plots highly disturbed or less than 10 m wide, and outlier relevés were removed from the original 232 plots. This yielded 180 relevés and 818 plant species. These relevés are taken as the most representative of the ecological conditions occurring in the studied sites. Plant nomenclature follows KEAY & Hepper (1954-1972), BRUNEL et al. (1984), BERHAUT (1967) and LEBRUN & STORK (1991-1997). The conventional six-part abundancedominance scale of BRAUN-BLANQUET (1972) is adopted in the present work. This scale was 54 Chapter 6: A phytosociological study of riparian forests in Benin (West Africa) applied for each species within each of the three strata: Dominant (height > 10 m), Medium (height between 3 and 10m) and Understorey (< 3 m high) species. N Gbèssè * Pendjari Reserve * Towns Yarpao * Pénéssoulou Bétérou * * * Idadjo * Bétékoukou * Samiondji 1 : Grand-Popo 2 : Cotonou 3 : Ouidah 4 : Porto-Novo 5 : Niaouli 6 : Aplahoué 7 : Pobè 8 : Bohicon 9 : Kétou 10 : Savalou 11 : Savè 12 : Bantè 13 : Tchaourou 14 : Bassila 15 : Djougou 16 : Bembèrèkè 17 : Natitingou 18 : Parakou 19 : Kouandé 20 : Tanguiéta 21 : Toui 22 : Kandi 23 : Banikoara *Sites of data collection (with > 20 plots) States limits Major waterways * Isohyets FIG. 6.1.- Study sites where data were collected in riparian forests in Benin 55 Chapter 6: A phytosociological study of riparian forests in Benin (West Africa) Table 6.1: Floristic and stand characteristics of riparian forests throughout Benin Climatic region Sites Type of riparian forests Relevés (n) * Species Trees Basal area richness/ha density/ha (BA/ha) Guinean Samiondji (Sa) Ouémé river 31 249 726 41.7±17.4 Sudano Streams in Pénéssoulou forest Ouémé river Ouémé river 34 343 544 38.9±15.3 17 30 129 195 748 785 35.8±10.7 41.6±15.8 8 22 358 540 732 33.1±7.9 42.7±9.9 Pénéssoulou (Pe) Idadjo (Id) Guinean Bétérou (Be) Sota (So) Yarpao (Ya) Sudanian Bétérou (Be), Batia (Ba) Sota river Streams at hills feet Streams on plateau with inundations Bétérou (Be), Toucountouna Streams on plateau (Tc), Gbèssè (Gb), Péhunco (Ph) seldom inundated 10 - 602 40.5±8.9 16 - 531 25.9±4.6 Porga (Po), Konkombri (Ko) 12 - 571 33.1±17.5 180 129-358 253-785 41.5±17.9 Pendjari river Total * n = number of phytosociological relevés; Trees = stems with dbh ≥ 10 cm. 6.2.3. Data analysis For data analysis of the 818 plant species and the 180 representative RF samples, both ordination and numerical classification methods were used. No species of the representative relevés was omitted in the analysis. Detrended Correspondence Analysis (DCA) (HILL & GAUCH 1980, TER BRAAK 1995, TER BRAAK & SMILAUER 1998) was used as indirect gradient method to detect the underlying ecological gradients/factors within the RF relevés set. The definition of plant communities was achieved through Two-Way Indicator Species (TWINSPAN) (HILL 1979). Dendrograms were obtained with STATISTICA® 5.1 (1998) using Euclidean distance as distance measure and Ward’s minimum variance as aggregation method. The classification of the individualised plant communities followed the work of LEBRUN (1947), LEONARD (1952), SCHNELL (1952), LEBRUN & GILBERT (1954), SCHMITZ (1963, 1971, 1988), and SOKPON et al. (2001). 6.3. RESULTS Environmental gradients/factors of riparian forests samples partition The stand characteristics of the surveyed riparian forests are presented in Table 6.1. Detrended Canonical Analysis (DCA) of 180 relevés and 818 plant species (Figure 6.2) showed a major indirect floristic factor (axis 1) correlated with the type or importance of waterways: small (streams) or large (rivers). There are three groups of plant communities: RF plant communities of the Pénéssoulou forest (group 1), along streams all over the country (group 2) and along rivers (group 3). This result is in accordance with the TWINSPAN classification output. The TWINSPAN output table (three A0 format sheets) is available on request from the authors. Each one of the 3 groups is further analysed in a partial analysis. Group 1: Riparian Forests (RF) of Pénéssoulou protected forest The partial DCA analysis of 34 samples and 282 plant species showed a floristic gradient (axis 1) of two plant communities correlated with topography (Figure 6.3). This 56 Chapter 6: A phytosociological study of riparian forests in Benin (West Africa) result is in accordance with the dendrogram (Figure 6.4). The denomination of each plant community is derived from two characteristic species, which are also the most indicative of the ecological conditions, topography in this case. They are: 1 - Plant community of Isolona thonneri and Callichilia barteri (10 relevés) along streams in the centre of Pénéssoulou protected forest. This plant community is located at the lowest parts of the forest along streams with frequent inundation. 2 - Plant community of Motandra guineensis and Pararistolochia goldieana (24 relevés) along streams at the East and West parts of Pénéssoulou reserve forest. This community is located in drained sites (i.e. seldom inundated). Group 2: Riparian forests along streams all over the country DCA of 48 relevés and 284 plant species (Figure 6.5) showed a major floristic factor (axis 1) correlated with land form variation, namely relief (RF at hill foot versus RF on plateau) and topography (RF on plateau regularly inundated versus RF on plateau seldom inundated). In general, this result is in accordance with the TWINSPAN output and the dendrogram (Figure 6.6). Four plant communities are discriminated from the ordination and classification analysis: 3 - Plant community of Chrysobalanus icaco subsp. atacoriensis and Pentadesma butyracea along streams at hill feet in the Atacora mountain chain (22 relevés). 4 - Plant community of Alchornea cordifolia and Ficus trichopoda along streams on regularly inundated plateaus all over the country (10 relevés). 5 - Plant community of Berlinia grandiflora and Khaya senegalensis along streams on drained plateaus (i.e. seldom inundated), mainly in the Sudanian region of the country (8 relevés). 6 - Plant community of Raphia sudanica and Oxytenanthera abyssinica along streams on drained plateaus, mainly in the Sudanian region (8 relevés). Group 3: Riparian forests along rivers all over the country DCA of 98 relevés and 291 plant species (Figure 6.7) showed a major gradient correlated with latitude (Axis 1: Guinean region of South versus Sudano-Guinean and Sudanian regions in the Central and North Benin) and longitude (Axis 2: extreme West versus Centre and East Benin). 3 sub-groups of relevés are shown: (a) - Bétérou, Idadjo and Sota in the Centre; (b) - Porga and Konkombri in the North-West and (c) - Samiondji in the South of the country. Each subgroup is further analysed in a partial DCA ordination and a classification. Only the DCA of sub-group (a), (Figure 6.8) showed a floristic gradient correlated with latitude (Axis 1): Centre of the Sudano-Guinean zone (Idadjo), North of the SudanoGuinean zone (Bétérou), and Sudanian region (Sota) at Ségbana latitude. These results are in accordance with the dendrogram (Figure 6.9) and TWINSPAN classification (Figure 6.10). Therefore five plant communities are discriminated in RF along rivers: 7 - Plant community of Cynometra megalophylla and Parinari congensis along the Ouémé river in the Guinean region of Southern Benin (31 relevés). 8 - Plant community of Capparis thonningii and Crateva adansonii along the Ouémé river in the Sudano-Guinean region of Central Benin (30 relevés). 9 - Plant community of Lepisanthes senegalensis and Drypetes floribunda along the Ouémé river in the Sudano-Guinean region of Central Benin (17 relevés). 10 - Plant community of Uapaca heudelotii and Irvingia smithii along the Sota river in the North-East of the country (8 relevés). 57 Chapter 6: A phytosociological study of riparian forests in Benin (West Africa) 11 - Plant community of Garcinia livingstonei and Combretum acutum along the Pendjari river in the North-West of the country (12 relevés). 6.4. DISCUSSION 6.4.1. Environmental gradients and factors partitioning riparian forest plant communities In total, 5 environmental gradients/factors were identified from the ordination process: 1 - Type or importance of waterways: RF along streams / RF along rivers; 2 - Relief: RF at hill feet / RF on plateau; 3 - Topography: RF on regularly inundated plateau / RF on seldom inundated plateau; 4 - Latitude: RF in the Guinean region of South-Benin / RF in the Centre of the SudanoGuinean region (Idadjo) / RF in the North of the Sudano-Guinean region (Bétérou) / RF in the Sudanian region (Sota) at Ségbana latitude; 5 - Longitude: RF in the extreme West / RF in the Centre and East of the country. The three sets of the RF data: Pénéssoulou, rivers and streams (see Figures 6.2, 6.10 and 6.11), somehow match the vegetation formations and phytogeographical districts of Benin. The flora of RF along rivers is much more dependent on frequent floods than the flora along streams, where the surrounding vegetation greatly influences stream side vegetation. On the other hand the Pénéssoulou region in West-Central Benin is the easternmost point of the dry semi-deciduous forest fire subtype (HALL & SWAINE 1981). This vegetation also covers Central Togo, and Central and South Ghana. In this hygrophile enclave surrounded by savannas, RFs occupy thalweg bottoms and share numerous species (e.g. Cleistopholis patens, Pierrodendron kerstingii, etc.) with the Pobè region, which is the wettest region of Benin in the South-East. The first 2 DCA axes, which have the highest contribution to the total variation in the data set, were linked to environmental gradients/factors: importance of waterway, relief, topography, latitude and longitude. Depending on forest types and sites characteristics, various authors have also documented the major environmental gradients/factors that explain the grouping of floristic relevés. In an application of ordination and classification to closedcanopy forest in Ghana, rainfall was not the only factor with a large effect on forest composition. Also rock type and occasional ground fires were important in drier forests. Altitude was important in wetter forest (HALL & SWAINE 1976, SWAINE & HALL 1976). RUSSEL-SMITH (1991) identified a primary latitude-moisture gradient and a subsidiary topographic-drainage gradient as the two major environmental gradients underlying the grouping of 1219 sites and 559 rain forest taxa in Northern Australia. According to HANSON et al. (1990), the most important processes related to topography are flooding, soil moisture and fluvial dynamics. For various tropical forests, altitudinal zonation, rainfall variability, as well as the influence of rock type, occasional ground fires, play an important role in the typology of forest communities (NAKASHIZUKA et al. 1992). The interpretation of the DCA gradients is broadly in accordance with valuable work on the phytosociology of edaphic and hygrophile vegetations of fresh water (e.g. LEBRUN & GILBERT 1954, SCHMITZ 1988, SOKPON et al. 2001). LEBRUN & GILBERT (1954) enumerated the ecological factors that play a role in the discrimination of edaphic forests linked to hydromorphic soils: a) variation of the water table (i.e. level and periodicity) above the soil surface and in the edaphic profile; b) the degree of alluviation; and c) the intensity of soil drainage. They further mentioned that physiographic characters of the site, the nature of water and substrate might play a role in the typology of hydromorphic vegetations, but marginal as compared to the precedent factors. Generally, the influence of flood and variation of the water table is considerable (MANDANGO & NDJELE 1986). SOKPON et al. (2001) found 58 Chapter 6: A phytosociological study of riparian forests in Benin (West Africa) that latitude, drainage and duration of flooding were the best discriminating factors for 51 wetland forest relevés in Benin. Meanwhile, their typology did not go beyond the formation level (i.e. gallery, swamp, periodically inundated and riverine forests). A limitation of that work is that no plant communities were identified, as in the present study. Also the differentiation in such broad vegetation types, based on the floristic composition, does not always conform to the reality of the terrain. The amount of variation explained by the DCA analysis is fairly low (14 to 21%). A reason might be the high number of relevés (180), species (818) and uncontrolled variables. GANGLO (2000), and HOUINATO (2001) found almost the same range of values for the explained variance, respectively 16 to 22% for plantations understorey in Central and Southern Benin, and 18 to 25% for savanna in Central Benin. BOCARD et al. (1992) pointed out that it may not be feasible to measure all the environmental variables (in the broad sense: biological interactions and external environmental factors) that are relevant in an ecological study. Given these constraints, and at this stage of the study of RFs in Benin, the underlying factors found through the interpretation of the DCA axes can nevertheless be considered as important in the structuring of these plant communities. 6.4.2. Classification of RF relevés The apparently ‘homogeneous’ site characteristics of the narrow RF on continually moist, sometimes flooded and with sandy shores, hides in fact more complex fluvial processes moulded by the regional climate, relief, and human activities. This allows a variety of groupings of plant species. The floristic classification of RF relevés allows us to identify 11 constant plot assemblages, each one referring to a plant community (i.e. plant association). An additional riparian forest plant community (Mimosa pigra and Ficus asperifolia) regularly cited in the literature (LEBRUN & GILBERT 1954, MANDANGO 1982, MANDANGO & NDJELE 1986, SCHMITZ 1988) is widely distributed on sandy banks of rivers in tropical Africa. However it could not be clearly distinguished and classified as the 11 previous types through the DCA and TWINSPAN analysis. Therefore we can say that there are at least 12 plant communities, each having its distinct and well-developed stands, in the Benin riparian forests. This classification, based on comprehensive phytosociological relevés, is both readily interpretable and repeatable. The two species forming the denomination of the plant community were carefully selected among the characteristic (exclusive, selective, preferential) species according to the degrees of fidelity defined by BRAUN-BLANQUET (1972). These species are among characteristic species having a high ecological indicator value. Each plant community may occur in many stations (or sites) but they all exist in only one well-defined and ecologically characteristic habitat, e.g. RF along rivers in the Guinean region of Southern Benin, RF along streams at hill feet in the North West of the country, and RF along streams on plateau in the Sudanian region. In general, the results from DCA, TWINSPAN classification and dendrograms are similar and they appear to be complementary techniques in the partition of groups of relevés based on their floristic composition affinity. 6.4.3. Relation plant communities/environmental factors Although some environmental factors were found to induce a certain grouping of RF samples based on presence/absence of species, the interpretation of causal relationships between the identified plant communities and these gradients should be taken with care (KENT & COKER 1992). From detailed understanding of the relationship between riparian vegetation and ecological factors, we might say that the importance of waterways, latitude, longitude, relief and topography act indirectly through climatic, geomorphologic and edaphic 59 Chapter 6: A phytosociological study of riparian forests in Benin (West Africa) variables. HOMMEL (1990) already found that both physiognomy and floristic composition clearly reflect the relation between climate and vegetation. On the one hand, the most important climatic conditions in Benin are rainfall, temperature, hygrometry, Potential Evapotranspiration (PET) and seasons duration; on the other hand, soil texture, structure and nutrient status, and bedrock type are the most important edaphic variables. The combination of these variables modulated by the relief and topography not only control the volume, velocity of water and the duration of inundation but also specific requirements of individual species in these edaphic and hygrophyle plant formations. Generally for plant communities, the effective external factors are numerous and variable, and the possible combinations so manifold. Also the overlapping is frequent and the relation of habitat to plant community is not a simple and reversible function. In consequence, a clear and unequivocal delimitation of the identified plant communities’ habitats according to operative external factors appears quite unattainable (BRAUN-BLANQUET 1972). As plant species experience the conditions provided by many environmental variables (TER BRAAK 1987), from the field data we can draw preliminary conclusions concerning the response of certain species to the combination of the ecological variables cited above. Species that seem to be influenced by climate and edaphic conditions along waterways in: - the Guinean region and Sudano-Guinean zone: Cynometra megalophylla, Lepisanthes senegalensis, Drypetes floribunda, Cassipourea congoensis, Cleistopholis patens, Napoleonaea vogelii, Pierreodendron kerstingii; - the Sudano-Guinean zone: Pararistolochia goldieana, Callichilia barteri, Detarium senegalense, Motandra guineensis, Pandanus candelabrum, Leea guineensis, Khaya grandifoliola, Pseudospondias microcarpa; - the Sudano-Guinean and Southern Sudanian zones: Pentadesma butyracea, Isolona thonneri, Raphia sudanica, Celtis toka, Garcinia ovalifolia, Ixora brachypoda, Uapaca togoensis, Ficus trichopoda; - the Sudanian region: Chrysobalanus icaco subsp. atacoriensis, Brenadia salicina, Chionanthus niloticus, Eriocoelum kerstingii, Thunbergia atacoriensis, Oxytenanthera abyssinica, Synsepalum passargei; - the Northern Sudanian zone: Irvingia smithii, Garcinia livingstonei, Combretum acutum; - the whole country: Pterocarpus santalinoides, Mimosa pigra, Ficus asperifolia, Dialium guineense, Elaeis guineensis, Cola laurifolia, Berlinia grandifolia, Parinari congensis, Kigelia africana, Crateva adansonii, Morelia senegalensis, Syzygium guineensis var. guineensis, Alchornea cordifolia, Khaya senegalensis, Uapaca heudelotii. Biotic factors such as intra and inter species relationships and human induced disturbance, which play a role in the grouping and layering plant species are not explored in the present study. This preliminary information about the occurrence of individual species might be useful in drawing distribution maps of these species and the plant communities in which they reach their optimal development in Benin. 6.4.4. Syntaxonomic relationships of the 12 plant communities We present a syntaxonomic classification of the identified plant communities under the hierarchical structure (orders and suitable alliances) of edaphic forest formations of fresh water occurring in West and Central Africa. Thus far, in Benin such a general system of RF plant communities classification is not available. All tropical African RF belongs to the Mitragynetea SCHMITZ 1963, the class of hygrophile forests of fresh water (SCHMITZ 1971, 1988). As for gallery forests of the Congo basin (see LEBRUN & GILBERT 1954), RF of Benin have either a Guinean or SudanianZambesian tendency. In Benin the most frequent plants of the Mitragynetea SCHMITZ 1963 60 Chapter 6: A phytosociological study of riparian forests in Benin (West Africa) are Pterocarpus santalinoides, Cola laurifolia, Syzygium guineense, Dialium guineense, Morelia senegalensis, Elaeis guineensis, Parinari congensis, Manilkara multinervis, Phaulopsis barteri, P. ciliata, Taccazea apiculata, Achyranthes aspera, Afzelia africana, Xylopia parviflora, and Antidesma venosum. Based on field data, similarities of ecological conditions and floristic composition, we classified the 12 RF plant communities into 3 orders that are Alchornetalia cordifoliae LEBRUN 1947, Lanneo-Pseudospondietalia LEBRUN & GILBERT 1954 and Pterygotetalia LEBRUN & GILBERT 1954 (Figure 6.11). * Alchornetalia cordifoliae LEBRUN 1947: It is the order of shrubby and pre-forest plant communities following herbaceous semi-aquatic formations along larger waterways (Ouémé, Sota rivers), with relatively important alluvial deposits. This order also occurs in larger valleys along streams with a long period of inundation. The most frequent species are Mimosa pigra, Ficus asperifolia, Alchornea cordifolia, Ficus spp., etc. Within this order we can distinguish 3 alliances each containing one plant community: 1 – Mimosion pigrae MANDANGO 1982, alliance with a Guinean tendency although it can be found in the Sudano-Zambesian region (LEBRUN & GILBERT 1954). The Mimosa pigra and Ficus asperifolia plant community is classified in this alliance. It occurs as dense shrubby but isolated stands always found on sandy bars close to the lowest level of the water along rivers. Therefore, every year it is flooded for 3 to 4 months. MANDANGO & NDJELE (1986) defined it as an association in the Congo basin. 2 – Alchorneion cordatae LEBRUN 1947, alliance of shrubby and pioneer plant communities of fresh water. The Alchornea cordifolia and Ficus trichopoda plant community is the only that can be classified in this alliance. In fact this community is widely distributed over the country in sites where the substrate is either constantly humid or periodically inundated or muddy with clayey soil. Several authors (e.g. LÉONARD 1952, KALANDA 1981, LEJOLY & MANDANGO 1982, MANDANGO & NDJELE 1986) have described a variety of plant communities with Alchornea cordifolia as characteristic species. 3 – Uapacion heudelotii LEBRUN & GILBERT 1954. This alliance harbours the plant community of Uapaca heudelotii and Irvingia smithii that was already defined as association by LÉONARD (1947) and KALANDA (1981). It is widely distributed throughout tropical Africa from Congo to Senegal (SCHMITZ 1988). Meanwhile in Benin the two characteristic species are only seen associated in the North, showing the association’s Northern Sudanian tendency. * Lanneo-Pseudospondietalia LEBRUN & GILBERT 1954: This order groups all riverine forest with a seasonally alternation of heavy inundation and short period of drainage. It is typical to the Guinean region (e.g. Samiondji latitude), meanwhile penetrates the Sudano-Guinean zone (Ouèssè, Bétérou and Pénéssoulou latitudes). The most frequent species of this order are Napoleonaea vogelii, Cassipourea congoensis, Salacia pallescens, Synsepalum brevipes, Cleistopholis patens, Pseudospondias microcarpa, Pierreodendron kerstingii, Spathodea campanulata, Trilepiseum madagascariense, Parquetina nigrescens. The associated plant communities are: a) - In the Guinean region of Southern Benin at Samiondji latitude: Cynometra megalophylla and Parinari congensis along the Ouémé river. b) - In the Sudano-Guinean region of Central Benin at Ouèssè and Bétérou latitudes: Lepisanthes senegalensis and Drypetes floribunda; Capparis thonningii and Crateva adansonii along the Ouémé river. c) - In the Pénéssoulou protected forest: Isolona thonneri and Callichilia barteri; Motandra guineensis and Pararistolochia goldieana. 61 Chapter 6: A phytosociological study of riparian forests in Benin (West Africa) * Pterygotetalia LEBRUN & GILBERT 1954: This order groups all riparian forests with a Sudanian-Zambesian tendency occurring in valleys or on drained plateau with short periods of inundation (few days during the rainy season) and long periods of dryness. For LEBRUN & GILBERT (1954), this order is not only edaphic but also physiographic as steep banks of hills can hem in streams. The most frequent species are Berlinia grandiflora, Khaya senegalensis, Raphia sudanica, Oxytenanthera abyssinica, Uapaca togoensis, Diospyros mespiliformis, Vitex doniana, Strychnos nigritana, Oxyanthus unilocularis, Chionanthus niloticus, Garcinia ovalifolia, etc. This order includes: a) - gallery forests of Central and North Benin on drained plateau (Berlinia grandiflora and Khaya senegalensis; Raphia sudanica and Oxytenanthera abyssinica plant communities), b) - gallery forests at hill feet (Chrysobalanus icaco subsp. atacoriensis and Pentadesma butyracea plant community), and c) - gallery forests along the Pendjari streams and its tributaries in the North-West of the country (Garcinia livingstonei and Combretum acutum plant community). Sites descriptions, floristic composition, physiognomy, ecological spectra and structural characteristics and phytosociological tables for each plant community will be discussed in separate papers (Natta et al., in preparation, for association of Garcinia livingstonei and Combretum acutum). 62 +5.0 Chapter 6: A phytosociological study of riparian forests in Benin (West Africa) Pe11 Pe7 Pe12 Pe9 Pe19 Pe8 Pe13 Pe5 Pe2 Pe25 Pe21Pe6 Pe15 Pe20 Pe16 Pe1 Pe22 Pe14 Pe17 Pe27Pe49 Pe18 Pe54 Pe47 Pe24 Pe55 Pe4 Pe53 Pe28Pe23Pe51 Pe48 Pe52 Pe26 Pénéssoulou (Group I) Sa29Sa28 Sa26 Sa25 Id7 Sa8 Id8 Id9 Id4 Id1 Id15 Id3 Be7 Id16 Id2 Id14 Be30 Streams all over the country (Group II) Ya1 Ya10 Ya5 Ya17 Ya8 Ya3 Ya14 Di4 Di3 Ya11 Ya15 Ya6 Ya13 Ya18 Ya7 Ba1 Di1 Di2 Ya19 Ou1 Be27Be20 Ph2 tk2 Gb5 Gb3 So8 Be29 Tc4 Tc7 Tc1Tc2 Tc5 Tc6 Gb4 Tc3 Gb2 Wa2 Ba7 Id12 Be23 Be26 Be1 Be22Be18 So7 Be13 Be14 So4So6 So5 So3 So1 Sa35 Sa19 Sa37 Sa30 Sa31 Sa9 Sa11 Sa23 Sa15 Be24 Be3 Be21 Be5 Sa36 Sa14 Sa33 Sa13 Sa7 Sa24Sa4 Sa22 Sa12 Id5 Id10 Be2 Be4 Id13 So2 Ph1 Sa34 Id11 Id6 tk1 Ya4 Ya9 Sa27 Sa38 Sa18 Sa32 Sa6 Be28 Sa16 Sa17 Sa10 Sa20 Sa3 Sa21 Rivers (Ouémé, Sota, Alibori, Pendjari) Group III Be11Be10 Be15 Be16 Be9 Be17 Be12 Be25Be6 Be19 Pj9 Tc8 Be8 Pj8 Po2 Ko5 Ba5 Ba4 Pj10 Ko7 Wa1 Ba3 Po3 Ko6 Po1 Po4 Pj11 Ba2 Ba8 Relevés along streams Relevés along rivers Axis 1 symbolises the type or importance of water ways: Relevés of RF along Streams versus those along Rivers FIG. 6.2. - Detrended Correspondence Analysis ordination of 180 riparian forests plots in Benin, showing 3 groups of plant communities. This result is in accordance with the TWINSPAN analysis. Relevés of RFs along streams versus those along rivers. -2.5 63 -1.5 +5.0 Chapter 6: A phytosociological study of riparian forests in Benin (West Africa) Axis 2 Plant community 1: west & east of Pénéssoulou forest (higher and medium parts) Plant community 2: centre of Pénéssoulou forest (lowest part of the forest with frequent Inundation) Pe53 Pe4 Pe2 Pe51 Pe47 Pe52 Pe55 Pe48 Pe15 Pe14 Pe49 Pe23 Pe20 Pe22 Pe24Pe27 Pe25 Pe13 Pe21 Pe26 Pe16 Pe12Pe54 Pe6 Pe18 Pe17 Pe8 Pe1Pe5 Pe28 Pe19 Pe7 Pe11 Pe9 Axis 1 Axis 1 symbolises the topography: bottom versus medium & centre of valleys FIG. 6.3. - DCA ordination of 34 plots of Group I (RF of Pénéssoulou reserve forest; see Fig. 6.2.) showing 2 plant communities correlated with topography. ( ) subdivisions obtained with TWINSPAN. 64 Chapter 6: A phytosociological study of riparian forests in Benin (West Africa) Dendrogram of 34 releves of Penessoulou (WARD, Euclidian distance) 110 100 90 Dissimilarity (%) 80 70 60 50 40 30 PE28 PE25 PE27 PE23 PE22 PE55 PE47 PE4 PE54 PE17 PE19 PE14 PE12 PE9 PE13 PE5 PE2 PE26 PE49 PE24 PE21 PE20 PE48 PE53 PE52 PE51 PE16 PE15 PE18 PE11 PE7 PE6 PE8 PE1 Plant community 1 Plant community 2 FIG. 6.4. - Dendrogram of 34 plots in the Pénéssoulou reserve forest showing 2 plant communities. ( ) subdivisions obtained with TWINSPAN. 65 Chapter 6: A phytosociological study of riparian forests in Benin (West Africa) Axis 1 symbolises land form variation: Relief (hills feet/plateau) and Topography of plateau (inundation/no inundation) Axis 2 Axis 1 RFs on plateau with no inundation Riparian forests (RFs) at hill feet RFs on plateau regularly inundated FIG. 6.5. - DCA ordination of 48 plots of RF along streams (see Group II in Fig. 6.2) showing 4 plant communities. ( ) divisions obtained with TWINSPAN. Tree of 48 releves of riparian forests along streams (WARD, Euclidian distance) 110 RFs on inundated plateau RFs at hills feet 100 RFs on plateau seldom inundated RFs on plateau seldom inundated 90 80 60 50 40 30 20 FIG. 6.6. - Dendrogram of 48 plots of RF along streams showing 4 plant communities. ( ) subdivisions from TWINSPAN. 66 GB2 GB3 GB4 GB5 WA1 SO8 WA2 TK1 PH1 PH2 TK2 OU1 BE29 TC1 BE30 TC7 TC5 TC6 TC2 TC3 BA2 TC4 BA8 BA3 BA4 BA5 BA7 DI1 BA1 DI2 DI3 DI4 YA1 YA9 YA10 YA8 YA11 YA3 YA15 YA14 YA5 YA4 YA6 YA7 YA13 YA17 YA18 10 YA19 Dissimilarity (%) 70 Chapter 6: A phytosociological study of riparian forests in Benin (West Africa) Axis 2 Centre & East > 20 N (c) (a) (b) West < 10 50’ N Axis 1 RFs along rivers in the Centre and North RFs along rivers in the South FIG. 6.7. - DCA ordination of 98 plots of RF along rivers all over the country (see Group III in Fig. 6.2) showing 3 sub-groups according to Latitude (axis 1) and Longitude (axis 2). Axis 2 Axis 1 Sota river at Ségbana latitude (North) Ouémé river at Bétérou latitude (Centre) Ouémé river at Ouèssè latitude (South) FIG. 6.8. - Partial DCA ordination of sub-group (a) of Fig. 6.7. (i.e. 55 plots of RF along rivers at Ouèssè, Bétérou and Ségbana) showing 3 plant communities according to Latitude (axis 1). 67 Dissimilarity (%) 120 100 80 60 40 20 0 Centre North west South Sudanian region Chapter 6: A phytosociological study of riparian forests in Benin (West Africa) South of the country 68 North Sudanian Guinean region: Samiondji = Sa; Sudano-Guinean zone: Pénéssoulou = Pe; Bétérou = Be; Idadjo = Id; Toui-Kilibo = Tk; Sudanian region: Yarpao = Ya; Péhunco = Ph; Dimari = Di; Ouaké = Ou; Gbèssè = Gb; Sota = So; Warrana = Wa; Toucountouna = Tc; Batia = Ba; Pendjari = Pj; Porga = Po; Konkombri = Ko. Name of riparian forest sites in Figs 6.2, 6.3, 6.4, 6.5, 6.6, 6.7, 6.8 and 6.9: FIG. 6.9. - Dendrogram of 98 relevés of riparian forests along rivers in Benin, showing 5 plant communities. ( ) divisions from TWINSPAN SA34 SA26 SA25 SA38 SA18 SA32 SA28 SA29 SA27 SA6 SA30 SA19 SA9 SA8 SA37 SA36 SA35 SA24 SA23 SA14 SA16 SA15 SA17 SA22 SA13 SA12 SA11 SA4 SA33 SA31 SA7 SA21 SA10 SA20 SA3 ID13 ID6 ID12 ID5 ID3 ID2 ID14 ID9 ID11 ID7 ID8 ID10 ID4 ID16 ID15 ID1 PO4 PJ8 PJ9 PO3 PO2 PO1 TC8 KO7 PJ11 PJ10 KO6 KO5 BE17 BE19 BE25 BE8 BE15 BE13 BE11 BE16 BE9 BE12 BE6 BE23 BE22 BE20 BE27 BE18 BE14 BE28 BE24 BE21 BE10 BE3 BE7 BE2 BE26 BE5 BE4 BE1 SO7 SO5 SO4 SO3 SO6 SO2 SO1 Riparian Forests (RF) data base Twinspan cut levels Identified gradients 1 Importance of waterway 2 Latitude 3 Topography Relief Latitude 4 Topography Latitude 5 Longitude No. Plant community Number of plots (180 relevés and 818 plant species) RF along STREAMS (82 relevés) RF along RIVERS (98 relevés) Centre (inundated) Sudano-Guinean & Sudanian regions Streams all over the country Sudano-Guinean zone (Pénéssoulou) East & West (drained) Plateau With inundation Hill feet (Yarpao) South SudanoGuinean zone (Idadjo) drained Guinean region (Samiondji) North Sudano-Guinean & Sudanian regions North Sudano-Guinean North-East Sudanian North-West Sudanian (Porga) North SudanoNorth-East Guinean (Bétérou) Sudanian (Sota) 1 Is. Th. & Ca. ba 10 2 Mo. gu. & Pa. go. 24 3 Al. co. & Fi. tr. 10 4 Be. gr. & Kh. se. 8 5 6 Ra. su. & Ox. ab. Ch. at. & Pe. bu. 8 22 7 Le. se. & Dr. Fl. 17 8 Ca. th. & Cr. ad. 30 9 Ua. he. & Ir. sm. 8 10 Ga. li. & Co. ac. 12 = Plant community of Isolona thonneri and Callichilia barteri along streams in the centre of Pénéssoulou reserve forest. 1 - Is. Th. & Ca. ba. 2 - Mo. gu. & Pa. go. =Plant community of Motandra guineensis and Pararistolochia goldieana along streams at the East and West parts of Pénéssoulou reserve forest. 3 - Al. co. & Fi. tr. = Plant community of Alchornea cordifolia and Ficus trichopoda along streams on plateau regularly inundated all over the country. =Plant community of Berlinia grandiflora and Khaya senegalensis along streams on drained plateau, mainly in the Sudanian region of the country. 4 - Be. gr. & Kh. se. 5 - Ra. su. & Ox. ab. =Plant community of Raphia sudanica and Oxytenanthera abyssinica along streams on drained plateau, mainly in the Sudanian region of the country. =Plant community of Chrysobalanus icaco subsp. atacoriensis and Pentadesma butyracea along streams at hill feet in the Atacora mountain chain. 6 - Ch. at. & Pe. bu. = Plant community of Lepisanthes senegalensis and Drypetes floribunda along the Ouémé river in the Sudano-Guinean zone of Centre Benin. 7 - Le. se. & Dr. Fl. = Plant community of Capparis thonningii and Crateva adansonii along the Ouémé river in the Sudano-Guinean zone of Centre Benin. 8 - Ca. th. & Cr. ad. = Plant community of Uapaca heudelotii and Irvingia smithii along the Sota river in the North-East of the country. 9 - Ua. he. & Ir. sm. 10 - Ga. li. & Co. ac. =Plant community of Garcinia livingstonei and Combretum acutum along the Pendjari stream in the North-West of the country. 11 - Cy. me. & Pa. co. = Plant community of Cynometra megalophylla and Parinari congensis along the Ouémé river in the Guinean region of Southern Benin. 12 - An additonal plant community, Mi.pi & Fi.as. = Mimosa pigra and Ficus asperifolia, from the literature exist but could not be discriminated through the numerical classification proccess. Fig. 6.10. - TWINSPAN classification hierarchy and relationship of the initial 180 relevés to 11 plant communities 69 11 Cy. me. & Pa. co. 31 Chapter 6: A phytosociological study of riparian forests in Benin (West Africa) Riparian forests of Benin (7-12ºN) Class of Mitragynetea Schmitz 1963 Typical Sudanian and Sudanian tendency riparian forests Typical Guinean and Guinean tendency riparian forests Order: Alchorneetalia cordifoliae Lebrun 1947 Order: Pterygotetalia Lebrun & Gilbert 1954 Order: Lanneo-Pseudospondietalia Lebrun & Gilbert 1954 Order: Alchorneetalia cordifoliae Lebrun 1947 Alliance: Alchorneion Alliance: Uapacion heudelotii Lebrun cordatae Lebrun 1947 Alliance: Mimosion pigrae Mandango 1982 & Gilbert 1954 11 12 Association: Mimosa pigra & Ficus asperifolia Mandango & Ndejele 1986 7 8 Pl. com.: Lepisanthes senegalensis & Drypetes floribunda Pl. com.: Cynometra megalophylla & Parinari congensis 2 Pl. com.: Motandra guineensis & Pararistolochia goldieana Pl. com.: Capparis thonningii & Crateva adansonii 6 1 4 Pl. com.: Chrysobalanus icaco subsp atacoriensis & Pentadesma butyracea Pl. com.: Isolona thonneri & Callichilia barteri 5 10 Pl. com: Raphia sudanica & Oxytenanthera abyssinica Pl. com: Berlinia grandiflora & Khaya senegalensis 3 9 Pl. com.: Alchornea Association: Uapaca heudelotii & Irvingia cordifolia & smithii Léonard 1947 & Ficus trichopoda Kalanda 1981 Pl. com.: Garcinia livingstonei & Combretum acutum FIG. 6.11.- Synthesis of the 12 Riparian forests plant communities identified in Benin and their synsystematic relationships. Pl. com.= Plant community. Numbers 1 to 12 refer to FIG. 6.10. 70 Chapter 7 SPATIAL DISTRIBUTION AND ECOLOGICAL FACTORS DETERMINING THE OCCURRENCE OF PENTADESMA BUTYRACEA SABINE (CLUSIACEAE) IN BENIN Submitted to Acta Oecologica Notulae Florae Beninensis 6 Natta A. K., Sinadouwirou Th.A., Sinsin B. and van der Maesen L.J.G. Chapter 7: Spatial distribution of Pentadesma butyracea in Benin Chapter 7 SPATIAL DISTRIBUTION AND ECOLOGICAL FACTORS DETERMINING THE OCCURRENCE OF PENTADESMA BUTYRACEA SABINE (CLUSIACEAE) IN BENIN (*) Notulae Florae Beninensis 6 Natta A. K.(1), Sinadouwirou Th.A.(2), Sinsin B.(3) and van der Maesen L.J.G.(4) (1) Agronomic Engineer, MSc. Department of Environment Management, Faculty of Agronomic Sciences; University of Abomey-Calavi FSA/UAC, 01 BP 526 Cotonou, Benin. Tel/Fax: + 229 303084; aknatta@yahoo.com (2) Agronomic Engineer, MSc. Centre National de Gestion des Réserves de Faune (CENAGREF) - Parc du W., BP 75 Kandi, Benin. Tel/Fax: + 229 303084; tsina@yahoo.com (3) Professor of Ecology, Department of Environment Management, Faculty of Agronomic Sciences. University of Abomey-Calavi. FSA/UAC, 01 BP 526, Cotonou, Benin. Tel/Fax: + 229 303084; bsinsin@bj.refer.org (4) Professor of Plant Taxonomy, Biosystematics group, Wageningen University; National Herbarium of the Netherlands, Wageningen University Branch, Gen. Foulkesweg 37, 6703 BL Wageningen, The Netherlands. Tel: + 31(0)317 483170; Fax: + 31(0) 317 484917, jos.vandermaesen@wur.nl (*) Submitted to Acta Oecologica ABSTRACT Sample surveys were conducted in all ten ecological districts recognised in Benin (West Africa) to assess the spatial distribution and ecological factors determining the occurrence of Pentadesma butyracea, a multipurpose tree species found exclusively in some gallery forests. Natural stands of the species occur in two regions at Bassila (Pira to Bodi) and in the Atacora mountain chain (Perma to Tandafa). Isolated trees are found around Agbassa village (between Ouèssè and Alafiarou in Central Benin) and as far as the latitude of Gbèssè village (Ségbana district) in the North East of the country. The presence of Pentadesma in the Bassila region was linked to the presence of dry semi-deciduous forests in the SudanoGuinean region of Central Benin. In the Sudanian region only riparian forests, with permanent water at hills foot, provide a water balance of sufficient level: the most important causal factor for the persistence of this species in a relatively dry and fire-prone landscape. Key words: Pentadesma butyracea, Non Timber Forest Products, indicator species, gallery forest, Benin. 7.1. INTRODUCTION Every country needs to undertake monographical research on biological diversity with special care to the multipurpose values of keystone or indicator ecosystems and species (UNCED 1992). Conservation agencies and scientists have recognised the requirement for spatial knowledge of biodiversity for purposes of planning, management and conservation evaluation (Acharya 1999). 73 Chapter 7: Spatial distribution of Pentadesma butyracea in Benin Pentadesma butyracea, tallow tree (Clusiaceae), is known to be a common species of dense evergreen forest (Aubreville 1959, Vivien & Faure 1985, Hawthorne 1996), and natural stands are found in Africa from Sierra-Leone to the Democratic Republic of Congo (Bamps 1971, Ouattara 1999). Therefore its presence in Benin, located at the discontinuity (the Dahomey gap) of the belt of closed tropical rain forest from Elmina (near Cape Coast, Ghana) to Porto-Novo, is an important issue for Beninese phytogeography. In Benin savanna is the dominant ecosystem, and there is an increasing interest in the study and conservation of Pentadesma butyracea, a multipurpose species found exclusively in some gallery forests (i.e. riparian forests) (Houngbedji 1997, Sinadouwirou 2000). The species is well known for its edible butter produced from its nuts that resemble those of Vitellaria paradoxa (Adomako 1977, Guelly 1994, Baumer 1995), and the fat is used for tallow production (Whitmore 1990, Abbiw 1990, Schreckenberg 1996). The wood of Pentadesma butyracea is of good quality (Purba & Sumarua 1987, Rachman et al. 1987, Tuani et al. 1994). Young stems are cut and sold as vegetal toothbrush in the region of Natitingou (North West Benin). In Ghana, Pentadesma oil is used for candle-making and margarine, its stems serve as chewing sticks and are made into for hair combs, its roots decoction is taken as anthelmintic and its bark combats diarrhoea and dysentery (Abbiw 1990). Meanwhile, a few authors have studied (e.g. Gunasekera et al. 1977, Tuani et al. 1994) the chemical, cosmetic and pharmacological properties of its butter, leaves, bark and roots. The ecology of Pentadesma butyracea upon which it is useful to base conservation and management activities (e.g. reforestation of gallery forests, agroforestry and valorisation of its various uses) is yet to be fully investigated. So far no information has been available on the spatial distribution of Pentadesma butyracea under the less favourable climatic conditions of Benin, as compared with the two West-African rain forest blocks. This paper maps the current distribution of the species and assesses the ecological factors of its presence and development in Benin. 7.2. MATERIAL AND METHODS 7.2.1. Study area Riparian forests were surveyed in the ten ecological districts of Benin (Adjanohoun et al. 1989, Houinato et al. 2000). Benin is located at the discontinuity of the tropical rain forest zone in West Africa, the ‘Dahomey Gap’, which is the product of topographic, oceanographic, climatic and human interactions (Jenik 1994). This gap includes essentially the drier types of the Guineo-Congolian forest belt in South-Eastern Ghana, Southern Togo and parts of South-Eastern Benin (Ern 1988). As a result, there are no evergreen tropical forests or rain forests in the country. Centuries of intense human activity accompanied by a drying climate resulted in the loss of most of Benin closed forests long ago. Small patches of moist forest remain today only on moist soils or as numerous sacred groves and riparian forests. The important hydrographic network allows the presence of many riparian forests widely distributed all over the country. The hydrographic network includes the Atlantic Ocean watershed with the Ouémé, Couffo and Mono Rivers and their tributaries and the Niger River, which is the end point of the Sota, Alibori and Mékrou Rivers. In the North West of Benin, the Pendjari River starts in the Atacora Mountains and ends in the Volta River in Ghana. In the Southern part there is an important complex of coastal lagoons, lakes and swamps. Table 7.1 summarises the climatic and edaphic characteristics of the two main areas of occurrence (Bassila and Natitingou) of Pentadesma butyracea in Benin. 74 Chapter 7: Spatial distribution of Pentadesma butyracea in Benin 7.2.2. Data collection in riparian forests Several riparian forests were surveyed using the Braun-Blanquet method for vegetation analysis based on phytosociological relevés. 500 m2 plots were installed in different sites in the Guinean region (Samiondji and Bétécoucou), the Sudano-Guinean region (Toui-Kilibo, Ouèssè, Agbassa, Alafiarou, Bétérou, Daringa, Onklou, Bassila and Pénéssoulou) and the Sudanian region (Ouaké, Affon, Boukombé, Toucountouna, Natitingou, Matéri, Tanguiéta and all the Pendjari Biosphere Reserve, Kouandé, Péhonko, Kérou, Banikoara, Kandi, Malanville, Karimama, Ségbana). For other parts of the country, a review of scientific work on the flora of Benin and the National Herbarium was made to check the presence of the species. Additional information was gathered as well from local people about the presence of this species on their land. Table 7.1: Climatic and edaphic characteristics of the two main regions (Bassila and Natitingou) of occurrence of Pentadesma butyracea in Benin Major factors (1) Region of Bantè-Bassila (Sudano-Guinean) Region of Natitingou (Sudanian) Climate type Sudano-Guinean climate (transitional Typical Sudanian climate zone between the Guineo-Congolian and the Sudanian climate characterised by progressive fusion of the two peaks of rainfall typical of the GuineoCongolian climate) Seasons 2 seasons: *rainy season: mid-April to endOctober (6 months) *dry season: End-October to mid-April (6 months) 2 seasons: *rainy season: May to mid-October (5 months) *dry season: mid-October to April (7 months) Humid season (P>½PET) from April to October from May to October Period of humidity excess (P>PET) from June to September from June to September Annual rainfall (mm) 1100 – 1300 1000 –1100 Average temperature (oC) 26 – 32 28 - 38 Relative humidity (%) 15 (dry season) to 99 (rainy season) 16 (dry season) to 98 (rainy season) Mean annual insolation (hours) 2420 2660 Average annual potential evapotranspiration (mm) 1536 1510 Hydromorph soils at hill feet, or Hydromorph and deep soils of lowlands, or along streams surrounded along streams surrounded by plateaus, or stony stream banks. by plateau. silt-clayed texture Clayed-sandy or silt-clayed texture with shallow water table Often presence of sand bars (1) Climatic data observed over the last 40 years (1956-1995); P = Rainfall (mm); PET = Potential EvapoTranspiration Edaphic conditions at stream banks 75 Chapter 7: Spatial distribution of Pentadesma butyracea in Benin In total 373 plots were installed all over the country and their coordinates taken by GPS. The presence of Pentadesma butyracea was checked and measurements taken from individual trees (dbh ≥ 5 cm, height, crown width, and distance to river or stream edge). The presence of the species was also checked outside plots when walking along rivers and streams. All Pentadesma butyracea trees detected outside plots were measured and their coordinates taken. Special care was made to detect the regeneration (dbh < 5 cm). Co-ordinates of plot centres where the species is found and coordinates of trees (≥ 10 cm) allow mapping the distribution of the species all over the country using the ArcView GIS 3.2 software. 7.3. RESULTS AND DISCUSSIONS 7.3.1. Pentadesma butyracea is a core species of some gallery forests in Benin Pentadesma butyracea grows in multi-species stands and is always found along some streams with more or less permanent water all year round. It is characterised by a barochore distribution of its diaspores. This species has not yet been seen along larger waterways such as the Ouémé, Zou, Sota, Alibori and Niger Rivers. We might say that the species somehow compensates the hydric deficit of the Sudanian region and Sudano-Guinean zone while avoiding long periods of floods (4 to 6 months) that characterises most riparian forests along rivers in Benin. Pentadesma butyracea has a high indicative value of gallery forests, as it is always tied to water. Moreover the distribution of the individual trees show that they are always found close to stream beds inside the gallery forests, near trees such as Pterocarpus santalinoides, Syzygium guineense var. guineense and Berlinia grandiflora. Pentadesma only occurs at gallery forest edges when the vegetation is degraded. Thus it can be termed as a core gallery forest species in Benin. Although it is an evergreen tree, it is frequent along waterways in the pre-forest, Sudano-Guinean and savanna zones of Central Côte d’Ivoire (Ouattara 1999). It has a gregarious distribution pattern shown by clusters of individuals in several segments of the streams (Sinadouwirou 2000). In Benin typical stands are found in the Pénéssoulou protected forest and along the Yarpao stream (Natitingou district). 7.3.2. Distribution of Pentadesma butyracea in Benin Based on recorded presence of the species, there are four separate areas of occurrence of Pentadesma butyracea in Benin (Figure 7.1). The first and second areas of occurrence are located in the Sudano-Guinean zone of Central Benin (Bassila region and around Agbassa village, mid-way between Idadjo and Alafiarou villages in Cental Benin). The third area of occurrence is in the Sudanian region of North West Benin (Natitingou region), while the fourth one is a small area near Gbèssè village far in the North East (Ségbana district). The Southern limit is around 8o15’ N in the district of Bantè. The Northern limit is the axis Tanougou-Séri (10o45’ N) and Gbèssè village (11o N). The Western limit follows the Togo border (around 1o22’ E at Pénéssoulou latitude), while the Eastern limit extends to (3o16’E). So far the species has not been seen elsewhere, but isolated trees might be found along streams in potential zone of occurrence that connect the four mentioned areas. The distribution of Pentadesma butyracea in Benin, based on intensive data collection, is a subset, of its larger-scale African distribution map (Bamps 1971). The species occurs in corresponding ecological areas: the mountains and plains of North and Central Togo (Zepernick & Timler 1984). In Ghana, the species is found in wet evergreen forests (Swaine & Hall 1976). 76 Chapter 7: Spatial distribution of Pentadesma butyracea in Benin While the Atacora mountain chain and the dry semi-deciduous forest (fire sub-type) of Bassila region common to Benin, Togo and Ghana do harbour stands of Pentadesma butyracea, the absence of the species in Southern Benin (below 8o N) is questionable. 0°30 1°00 1°30 2°00 2°30 3°00 3°30 4°00 NIGER 12°00 12°00 BURKINA - FASO Benin N AFRICA 11°30 11°30 Kandi N 11°00 4 Tanougou 11°00 Gbèssè I Tandafa Guilmaro Tanguiéta G 10°30 10°30 Kouandé Natitingou E 3 R Nikki 10°00 Djougou 10°00 I 9°30 A Parakou 9°30 Pénéssoulou Legend Bassila 9°00 T 8°30 Agbassa 1 Ouèssè Bantè O 9°00 1 Bassila region 2 Agbassa 3 Natitingou region Gbèssè 8°30 State limit 8°00 2 4 Savè G 8°00 Presence of Pentadesma butyracea Potential area of occurrence of Pentadesma butyracea O 7°30 Abomey 7°00 7°30 7°00 Lokossa Porto-Novo Cotonou 6°30 0°30 1°00 1°30 2°00 2°30 6°30 Scale : 1/3,300,000 3°00 3°30 4°00 Figure 7.1: Distribution map of Pentadesma butyracea in Benin 77 Chapter 7: Spatial distribution of Pentadesma butyracea in Benin One might expect the presence of Pentadesma in the Pobè region, the wettest region of Benin in terms of rainfall (1300 to 1400 mm per year) and relative humidity. In fact this region is dominated by remnant dense semi-deciduous forests that are contiguous to the relatively ”dry” forest of South-Western Nigeria (Onochie 1979). According to Swaine & Hall (1976) Pentadesma butyracea occurs usually in wet evergreen forest with at least >1750 mm mean annual rainfall. It is obvious that Pentadesma occurs in Benin in a less suitable climate environment than elsewhere. Therefore we cannot expect large stands in fire-prone and drier landscapes. Centuries of intense land clearance, vegetation fragmentation through fire, destruction of riparian forests, selective tree cutting, over-collection of nuts, as well as a drying climate have probably contributed to the fragmentation and isolation of distribution areas of Pentadesma butyracea. It is likely that the current non-contiguous areas of occurrence of this species are remnant and refugee zones in Benin that need to be protected and managed for their conservation. 7.3.3. Causal factors of the presence and persistence of Pentadesma butyracea in Benin Several authors have documented the influence of abiotic and biotic factors on the presence and spatial repartition of plants. For natural stands, climatic and edaphic conditions are the most important factors, meanwhile intense disturbance may determine a particular distribution pattern or cause complete disappearance of a species. * Expanse of the dry semi-deciduous forest: the probable cause of the presence of Pentadesma butyracea in the region of Bassila. The region of Bassila is the easternmost point of the dry semi-deciduous forest, fire subtype (Hall & Swaine 1981). This hygrophile enclave surrounded by savannas stretches out to Togo (Southern part of Togo mountains) and to the South and Centre of Ghana, where the species is present (Bamps 1971, Zepernick & Timler 1984). From this entry point Pentadesma butyracea probably spread to the East around the Alafiarou village and South around Bantè district. Another viewpoint is that the species has always been there in the past and due to certain factors (e.g. human-induced degradation or drying climate) the current occurrence area in the Sudano-Guinean zone of Central Benin is remnant of a previously larger one. * High water balance: the most important causal factor for the persistence of Pentadesma butyracea in the Sudanian region of Natitingou. The Sudanian region of Natitingou (North West of Benin), is dominated by the Atacora mountains chain continuing into the mountains of Northern Togo; mountains at the foot of which Pentadesma butyracea is present (Zepernick & Timler 1984). The niches for Pentadesma in this region are gallery forests at the foot of hills, because they provide an ample water supply: a combination of medium rainfall and evapotranspiration, high relative humidity, lower temperatures, resurgence of water at the foot of hills, presence of water in streams all year round. Here the relief and topography counterbalance the less suitable climatic conditions. According to Ouattara (1999), in Côte d’Ivoire Pentadesma butyracea disappears when annual rainfall falls below 1000 mm. Therefore the viability of natural stands of the species depends on the existence of high to medium quality habitats, such as the gallery forests in the Bassila and Natitingou region. On the other hand the presence of a few individuals as far as at the latitude of Kandi-Ségbana (with an average annual rainfall of 900 78 Chapter 7: Spatial distribution of Pentadesma butyracea in Benin to 1000 mm) shows the adaptation of the species to sub-optimal habitats in terms of annual rainfall and relative humidity. Under natural conditions, this matches the thesis that the longterm survival of populations can be strongly affected by the spatial and temporal distribution of both suitable and unsuitable habitat patches (Pulliam et al. 1992, Carrol et al. 1996). Gallery forests of the Sudano-Guinean zone and Sudanian region Benin, therefore act as refugee ecosystem for Pentadesma butyracea. In general, little is known about the geographical patterns of genetic diversity for most tropical forests species (Aide & Rivera 1998); and about Pentadesma butyracea populations in West Africa in particular. The distribution of genetic diversity within and among the four non-contiguous populations in Benin, and those in Togo and Ghana will give insight in the origin and evolution of this species in Benin. As Pentadesma butyracea in Benin is exclusively present in gallery forests, some hypotheses on the ecological importance of these hygrophile and edaphic forest formations are substantiated. * Gallery forests act as refugee ecosystem for forest tree species in a fire prone environment In many seasonally dry regions, but also in formerly forested areas, gallery forests have the character of a refuge for plants and animals (Porembski 2001). The extent and impact of fire in a system of gallery forests were evaluated in the Mountain Pine Ridge savanna of Belize, (Kellman & Tackaberry 1993, Kellman et al. 1994) and such systems represent plausible refugia for forest species in fire-prone landscapes. In Benin, it is obvious that Pentadesma butyracea cannot survive but inside the wettest ecosystem that exists in the majority of fire-influenced landscapes. * Gallery forests harbour the tree species requiring moist habitats in the landscape Fresh-water bodies, surface-emergent aquifers and soil moisture in excess in gallery forests are the underlying factors that support mesic vegetation distinctive in structure and/or floristics from that of the contiguous more xeric uplands (Warner 1979). Due to their elevated moisture regime (compared to the surroundings), they enable the establishment and persistence of plant species, which under zonal conditions are bound to wet forest types, often evergreen rain forest (Porembski 2001). Tropical riparian forest fragments are known for their potential, albeit limited, to maintain large numbers of species and may act as safe sites for tropical rain forest species (Kellman et al. 1994, Meave & Kellman 1994). Likewise gallery forests are the most suitable vegetation for plant species adapted to a moist climatic regime (Medley 1992). Therefore, Pentadesma finds suitable ecological niches in gallery forests that border streams in savanna dominated landscape. * Gallery forests act as route for movement of certain species across the landscape Riparian forests are important as routes for movement of plants and terrestrial animals across the landscape (Forman & Godron 1986). They play an important role as migration channels, which provide opportunities for genetic exchange between geographically isolated populations (Porembski 2001). This seems to be the case with Pentadesma butyracea in gallery forest corridors at hills feet along the Atacora mountain chain and in the semideciduous forest stretching from Ghana to Benin. 79 Chapter 7: Spatial distribution of Pentadesma butyracea in Benin 7.3.4. Major causes of gallery forests degradation in Benin and their consequences on natural stands of Pentadesma butyracea Natural stands of Pentadesma butyracea are disappearing along with the destruction of gallery forests in Benin. * Destruction of gallery forests Rapid changes in land use in Benin have led to the destruction and fragmentation of riparian forests. Riparian forests provide fertile soil for cultivation and provide an opportunity for irrigation (Natta 2000). Riparian forests form a highly endangered ecosystem, and their destruction could have severe consequences for the genetic exchange between disjunct vegetation stands (Natta et al. 2002). In the Sudano-Guinean region of Côte d’Ivoire, agropastoral management is the underlying factor of the disappearance of the species from its natural habitat (Ouattara 1999). As a core species of gallery forest, destruction of gallery forest implies disappearance of Pentadesma butyracea individuals. * Harvest of Pentadesma nuts The regeneration of Pentadesma is already very difficult under natural conditions as seeds have a short germinative power (Ouattara 1999). Therefore the collection of all or most nuts of this species from the wild by local people is a major threat to its survival in Benin. * Cutting of young stems for sale as vegetal teeth brushes As in Ghana (see Abbiw 1990), young stems of Pentadesma are used as chewing sticks in the North West of Benin. Therefore, the growth of young stems is seriously affected and many cannot reach the stage of fructification, particularly in the Natitingou region. * Selective tree cutting of the largest stems for timber The good wood quality and the facility of working induce people to harvest the biggest stems for the local wood industry. * Disturbance due to fire Uncontrolled bush fires often affect badly the regeneration of Pentadesma butyracea in the dry season. Effects are more visible in isolated patches or narrow gallery forests than in fair-sized forests. 7.4. CONCLUSION Pentadesma butyracea is found in Benin in four non-contiguous areas. Its distribution is principally influenced by the presence of numerous gallery forests in the semi-deciduous forest of Bassila region and at the foot of hills in the Atacora mountain chain that continue in Togo and Ghana. These four current areas appear to be remnant and refugee zones of Pentadesma butyracea, and the assessment of their genetic diversity will provide new insight on this enigma. In the meantime, these remnant populations need to be protected and managed with and by local people. 80 Chapter 8 FOREST STRUCTURAL PARAMETERS AND FLORISTIC COMPOSITION SPATIAL VARIATION AND MODELLING ACROSS RIVERS IN BENIN A.K. Natta Chapter 8: Structural parameters and floristic composition variation across rivers in Benin Chapter 8 FOREST STRUCTURAL PARAMETERS AND FLORISTIC COMPOSITION SPATIAL VARIATION AND MODELLING ACROSS RIVERS IN BENIN Natta A.K.(1) (1) Department of Environment Management, Faculty of Agronomic Sciences. FSA/UAC 01 BP 526 Cotonou, Benin; aknatta@yahoo.com. ABSTRACT Forest structural parameters, floristic composition and spatial distribution of tree species were assessed across several rivers in Benin through multivariate analysis, distribution curves and model fitting. The variation of tree stems from riverside up to 100 m away reveals an uneven distribution of abundance, while tree height and basal area variations at riverside do not show any clear pattern. Results from the numerical analysis show differences in the floristic composition at riverside, across riparian forests and neighbouring plant communities. The gradual change in species composition and relative abundance, which corresponds to a gradual environmental change, indicates that tree species have different affinities to water and humidity from the river. Typical riparian forest tree species are Syzygium guineense, Pterocarpus santalinoides, Parinari congensis, Cola laurifolia, Napoleonaea vogelii, Cynometra megalophylla, Drypetes floribunda and Manilkara multinervis. This research not only confirms empirical knowledge about the ecology of certain riparian forest and surrounding vegetations plant species, but also makes a distinction between river front, the central portion and edge species of riparian forests. Criteria (e.g. average distance range and best fit models) are also provided for the most common tree species along river. These results suggest that riparian forests little disturbed by men can be partitioned in three different habitats among tree species, along the horizontal gradient of wetness. Meanwhile due to lack of sufficient data we cannot yet draw conclusions about the distribution behaviour of all tree species encountered along the riverside. An implication for plant species diversity assessment in riparian forest is that sampling design (i.e. plot size, shape and layout in the terrain) should take into account the three major parts of the floristic composition at riverside (river front, central portion and edge). Therefore rectangular plots with varying width and length, and covering the whole riparian forest cross section are the most suitable sampling units for savanna regions. Keys words: Riparian forest, adjacent vegetation, structure, floristic composition, spatial distribution, modelling, Benin. 8.1. INTRODUCTION The riparian zone, functionally defined as a three dimensional zone of direct interaction between aquatic and terrestrial environments, has boundaries that extend laterally from the channel to the limits of flooding and vertically into the vegetative canopy (Swanson & Lienkamper 1982, Gregory et al. 1991). Riparian forests are at the same time so important as a collection of natural resources and so limited in individual size that they both merit and require higher resolution inventory procedures than those deployed for other larger 83 Chapter 8: Structural parameters and floristic composition variation across rivers in Benin ecosystems such as savanna and upland forests (Natta et al. 2002). In both predominantly forested areas and savanna regions, they sustain a type of vegetation that is distinct from the surrounding area in species composition and vegetation structure. Their location between river and adjacent landscape contribute to their ecotone character (Brinson & Verhoeven 1999). Due to their elevated moisture regime, they enable the establishment of plant species which under zonal conditions are bound to wet forest types (Porembski 2001). The study of West-African riparian forests has hitherto been neglected and the published studies dealt mainly with their structural and floristic characteristics and dynamics (Devineau 1975, Bonkoungou 1984, Bélem & Guinko 1998, Goudiaby 1998, Lykke & Goudiaby 1999, Natta 2000, Natta et al. in preparation). Few publications have been devoted to the spatial variation of species and forest structural parameters across riparian forests (Porembski 2001, Natta & Porembski, in press, see chapter 9). Classification and ordination of riparian forests floristic data in Benin resulted in three major groups of plant communities: riparian forest along rivers, along streams and the one of Pénéssoulou protected forest (Sudano-Guinean zone of Central Benin) (Natta et al. in press, see Chapter 6). This paper deals with the spatial variation of forest structural parameters and floristic composition across rivers. The objective is to investigate the floristic composition variation and assess the spatial distribution of tree species at riverside, across riparian forests and adjacent vegetation formations. We also explore the effect of distance from water on the composition and structure of tree species, so as to detect riverside, interior and edge species of riparian forests. The research questions were: Do structural parameters (i.e. abundance, height and basal area) vary across riparian forests and neighbouring plant communities at riverside? Is there any relationship between riparian forest width and tree species richness? Does the floristic composition vary across riparian forests and neighbouring plant communities? Is there any recognisable spatial distribution of tree species, across riparian forests and neighbouring plant communities, and how can it be modelled? This contribution aims to provide insights into the ecology of certain tree species at riversides in West African riparian systems and provide recommendations for sampling designs. 8.2. Material and Methods To assess the floristic composition variation and capture distribution patterns of tree species, a cross section or crown cover intersection method was applied. This method consists of a 100 m long transect, which runs through the riparian forest and beyond its edge through the adjacent (i.e. neighbouring) plant communities. At every point, a tape was laid on the ground, perpendicular to the river direction, beyond the riparian forest edge. The position of each tree stem (dbh ≥ 10 cm) is measured from the river bed to the base of each tree having its crown intercepting the tape. This method was applied every 100 m from a random starting point, perpendicularly to river direction, so as to cover at least all the riparian forest and the ecotone (i.e. transitional) zone with adjacent vegetation. Five riparian forest sites were surveyed from 7º20 to 11º30 N in Benin, along the Ouémé, Alibori, Sota and Pendjari rivers. One riverside was surveyed at each site. In total 56 transects were done. We measured the distance from the river, dbh, height and crown width of each stem. Riparian forest width, ecotone zone and the type of adjacent vegetation formation were recorded as well. The position of trees permitted to compile floristic lists at different distance intervals (e.g. every 1 m, 2m, etc.) from the river bed up to 100 m. The floristic composition of each distance class has been taken taken as a relevé. The variability and similarity of the 84 Chapter 8: Structural parameters and floristic composition variation across rivers in Benin floristic composition from the river bed were assessed through Detrended Correspondence Analysis (DCA) (Hill & Gauch 1980, ter Braak & Smilauer 1998) and Two-Way Indicator Species Analysis (TWINSPAN) (Hill 1979). Clusters are obtained with Statistica® (1998) using Euclidean distance as distance measure and Ward’s minimum variance as aggregation method. The numerical analysis will detect any gradual change of the floristic composition in relation with the distance to river bed. Break lines, if detected, within this gradual floristic change will indicate types of natural grouping of species linked to landscape units at riverside. The upper limits of the detected distance intervals will be taken as threshold values to classify each tree species among these natural groupings. We tested the presence of riverside species (RS), species of riparian forest middle (MS), riparian forest edge species (ES), and adjacent ecosystem species (AES), through a t-test, which compared the distribution of stem positions to each class threshold value. The distribution pattern of each tree species, in the direction perpendicular to river, is obtained in plotting the relative abundance of stems (Y) versus their distance to river (X). Scatter plots are drawn and several models fitted. The best model, in terms of the coefficient of correlation (R2), significance of overall regression equation (F value) and regression coefficients (t-tests) was selected. The illustration of such typology (see Figure 8.1), which substantiates the presence of a riparian vegetation width effect (or patch width effect, Forman & Godron 1986), is summarised as follow: a - Riverside species (RS) are those close to river bed and only present in riparian forests. This group includes species that have constantly decreasing number of individuals within riparian forest, from the river. b - Species of the middle of riparian forest (MS) have a more or less bell shape abundance distribution centred at the middle (i.e. central portion) of riparian forest. c - Riparian forest edge species (ES) have their highest abundance at the external end of riparian forests. d - Adjacent ecosystem species (AES) are generally not seen inside riparian forests and belong to mature neighbouring natural vegetation formations. Under the study area conditions this group includes species typical of savanna woodland, and open, dry or dense semi-deciduous upland forests. The response of plant species expressed in individual plant growth, frequency or cover to the varying conditions along an ecological gradient often takes the form of bellshaped curves (van Groenewoud 1975). Meanwhile, the theoretical species distribution curves presented in Figure 8.1 are consistent with field evidence. In effect, along larger waterways (i.e. rivers) with more or less extensive flood plain the riparian vegetation is usually distinct from that of surrounding uplands. Moreover within riparian forests, the abundance of certain tree species decreases dramatically a short distance from the river while other shows the reverse trend (Ericsson & Schimpf 1986, MacDougall & Kellman 1992, Stromberg et al. 1996, Natta 2000, Porembski 2001, Natta & Porembski, in press see chapter 9). 85 Chapter 8: Structural parameters and floristic composition variation across rivers in Benin 10 Abundance (%) per species 9 (a ) - R ive r s ide s pe c ie s (R S ) (b) - S pe c ie s a t the c e ntra l po rtio n o f ripa ria n fo re s t (M S ) (c ) - R ipa ria n fo re s t e dge s pe c ie s (ES ) (d) - Adja c e nt e c o s ys te m s pe c ie s (AES ) 8 7 Maximum riparian forest w idth Adjacent vegetation formation 6 5 Ecotone 4 3 2 1 0 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 Distance from river bed (m) Figure 8.1: Simplified theoretical distribution curves of tree species at riverside, across riparian forests and neighbouring vegetation formations (riparian vegetation width effect) 8.3. RESULTS 8.3.1. Characteristics of the cross sections at each site We present here results of the least disturbed transects along the Ouémé, Sota, and Pendjari rivers. Table 8.1 summarises the characteristics of the cross sections at riverside at each site. Some 106 tree species were identified from 56 cross sections through riparian forests and adjacent plant communities. The average width of undisturbed riparian forest (35.31 ± 7.97 m) includes three units generally found at riverside: the riverside close to river bed, the middle and edge of riparian forest. Table 8.1: Characteristics of the studied sites Sites Rivers Samiondji (Guinean region) Idadjo (Sudano-Guinean zone) Bétérou (Sudano-Guinean zone) Sota & Porga (Sudanian region) Total of all rivers Ouémé Ouémé Ouémé Sota & Porga No. of transects 21 16 11 08 56 No. of individuals 496 426 420 228 1571 No. of tree species 44 48 55 39 106 Mean (± SD) RF width (m) 29.14 ± 9.63 41.12 ± 19.5 38.27 ± 8.96 27.25 ± 9.60 35.3 ± 7.97 SD = Standard Deviation; RF = Riparian forest 8.3.2. Variation of the number of stems, tree height and basal area at riverside The abundance of tree species stems over a width of 100 m is unequally distributed (Figure 8.2), and is best explained by an exponential function. There is a constant decrease of tree individuals in dense riparian forests (up to 50 m from the river). After 50 m, in the adjacent plant communities, the further decrease is less pronounced. Eventually a slight increase of stem abundance is seen from 90 to 100 m. On the contrary, tree height (Figure 8.3) and basal area (Figure 8.4) variations at riverside do not show any clear patterns, and the coefficients of 86 30 Chapter 8: Structural parameters and floristic composition variation across rivers in Benin Abundance of stems (%) determination (R2) were very low. This shows the weak relationships between these variables. Meanwhile the first distance class (0 - 2 m), that has the highest abundance (see figure 8.2) and lowest average height (see figure 8.3), gives the highest basal area value (see figure 8.4). 10 9 8 7 6 5 4 3 2 1 0 y = 0.0011x2 - 0.1489x + 5.8331 R2 = 0.8619 0 10 20 30 40 50 60 70 80 90 100 Distance from river bed (m) Average height of stems (m) Figure 8.2: Distribution of tree stems across the riverside 20 18 16 14 12 10 8 6 4 2 0 y = -0.0004x 2 + 0.0486x + 9.7428 R2 = 0.0963 0 10 20 30 40 50 60 Distance from river bed (m) 70 80 90 100 Figure 8.3: Variation of the average tree height across the riverside Average basal area (m2/ha) 8 y = 0.0001x 2 - 0.0185x + 2.1511 R2 = 0.0859 7 6 5 4 3 2 1 0 0 10 20 30 40 50 60 Distance from river bed (m) 70 80 90 100 Figure 8.4: Variation of the average basal area of trees across the riverside 87 Chapter 8: Structural parameters and floristic composition variation across rivers in Benin 8.3.3. Relationship between tree species richness and riparian forest width Tree species richness and riparian forest width are positively correlated (Figure 8.5), but R2 is very low (0.21 for linear and polynomial curves). Therefore we expect more tree species to accumulate more or less linearly as riparian forest width increases. 16 y = 0.0871x + 3.1726 14 R2 = 0.2166 Tree species richness 12 10 8 6 4 2 0 0 10 20 30 40 50 60 Riparian forests w idth (m) 70 80 90 100 Figure 8.5: Relationship between tree species richness and riparian forest width along rivers 8.3.4. Floristic composition variation at riverside Variation of tree species richness and floristic relevés distribution from the river bed up to 100 m are assessed. Tree richness seems to decrease with increasing distance from the river, but the fit is not good (see Figure 8.6). 25 y = -0.0005x 2 - 0.0209x + 15.437 R2 = 0.3045 Number of tree species 20 15 10 5 0 0 10 20 30 40 50 60 70 Distance from river bed (m) 80 90 100 110 Figure 8.6: Distribution of tree species richness at riverside The DCA and TWINSPAN analyses (see Figures 8.7 and 8.8, respectively) show a clear pattern in the floristic relevés distribution. The gradual change in the floristic composition is illustrated by the succession of the relevés from the river to about 60 m, this includes the whole riparian forest width, the transitional zone and parts of the neighbouring vegetations. Beyond 60 m the floristic composition seems to exhibit no variation. The first DCA axis (Figure 8.7) is correlated with distance from the river bed. It symbolises the 88 Chapter 8: Structural parameters and floristic composition variation across rivers in Benin variation of the floristic composition at riverside, across riparian forest and adjacent vegetations. The gradual change in species composition at riverside is also clearly shown by the TWINSPAN output (available on request from the authors), and is summarised in Figure 8.8. It shows the existence of an ecotone between 30 and 57 m, which separates the dense riparian forests and the adjacent plant communities. The interpretation of the DCA (Figure 8.7), TWINSPAN (Figure 8.8) outputs, the average RF width (27.34 to 43.28 m), and the knowledge about the ecology of each species occurring at river edges, suggest three break lines (or intervals) within riparian forests: - [0 - 16] m from the river bed: for species mostly found close to the river bed; - [17 - 29] m from the river bed: for species mostly found in the centre (i.e. middle portion) of the riparian forest; - [30 - 57] m from the river bed: for species in the transitional zone with adjacent plant communities. These results confirm the theoretical subdivision of species composition, at least for tree species, into riverside, middle and edge species of riparian forest, and species typical of adjacent vegetations (see Figure 8.1). Table 8.2 is derived from Figures 8.7 and 8.8. It summarises the floristic variation at riverside and the natural grouping of tree species within riparian forests. The distance intervals that are obtained from the TWINSPAN output indicate that the floristic composition of undisturbed riparian forests (i.e. on average 27.34 to 43.28 m wide), include those species always found close to the river bed, in the middle and edge of riparian forest respectively. Table 8.2: Floristic composition variation across riparian forests and adjacent vegetations Distance interval (1) 0 to 16 m 17 to 29 m 30 to 57 m 58 to 101 m Landscape units Riverside close to river bed Middle of riparian forest Edge of riparian forest Adjacent vegetation and ecotone zone formations Pattern of species RS MS ES AES Threshold distances Most faithful tree species in each landscape unit 16 m 17 and 29 m 30 and 57 m More than 57 m Syzygium guineense, Pterocarpus santalinoides, Parinari congensis, Cola laurifolia, Napoleonaea vogelii Cynometra megalophylla, Drypetes floribunda, Manilkara multinervis Dialium guineense, Diospyros mespiliformis, Elaeis guineensis, Cola gigantea, Ceiba pentandra Albizia ferruginea, Combretum collinum, Lonchocarpus sericeus, Millettia thonningii Anogeissus leiocarpus (1) Distance intervals are obtained from TWINSPAN analysis (see Figure 8.8); RS = Riverside Species; MS = Middle of riparian forest Species; ES = Ecotone zone Species; and AES = Adjacent Ecosystem Species. The position of tree individuals from river compared to threshold values of each riparian forest portion (see Table 8.2) allow the detection of distribution patterns, and the investigation of the issue of core and edge species in riparian forests. For each tree species, the t-test compared the positions of all stems from river bed to riparian forest portions threshold distances: 16 m for river front (or riverside); 17 and 29 m for species always in the middle of riparian forest; 30 and 57 m for ecotone zone species; and more than 57 m for adjacent plant communities species. Table 8.3 (annex) summarises the outputs from t-tests. Therefore the following typology and definitions are given for tree species found at riverside in the Benin context: - Riverside tree species (RS) are those that always occur between 0 and 16 m from the river (i.e. river front species); - Tree species in the middle of riparian forest (MS) are those that occur always between 17 and 29 m from the river (i.e. central portion of the riparian forest); 89 Chapter 8: Structural parameters and floristic composition variation across rivers in Benin - Tree species, that always occur between 30 and 57 m, are either riparian forest edge species or ecotone species (ES); and - Tree species typical of adjacent plant communities (AES) are those that always occur outside riparian forests. - We also define typical RF tree species along rivers as those always occurring between 0 and 35.3 m away from the river bed. This typology is only valid for rivers in Benin. The typology related to small waterways (i.e. streams) is not investigated in the present chapter. The abundance distribution models for 18 tree species are shown in Figures 8.9; 8.10; 8.11 and 8.12 (see Annex). We present here results for tree species having at least 20 individuals recorded at riversides across riparian forests. In general the distribution curves follow the theoretical trends of Figure 8.1 for each species type. 90 Chapter 8: Structural parameters and floristic composition variation across rivers in Benin Group I Group II Relevés of the river front (0 to 16 m) Relevés in the middle of riparian forest (17 to 29 m) Group III Group IV Relevés of riparian forest edges and ecotone zone (30 to 57 m) Relevés outside riparian forests (> 57 m) 89 91 87 79 75 93 23 55 99 25 1 3 5 9 13 7 11 15 19 43 33 17 21 41 3935 31 59 49 51 53 65 27 61 85 101 45 83 81 73 57 95 696367 77 47 29 37 97 71 Axis 1 is correlated with distance from the river bed (0 to 100 m) -2.0 +7.0 Figure 8.7: Detrended Correspondence Analysis (i.e. ordination) of the floristic composition of tree species at riverside. It shows a gradual variation of tree species composition at riverside. ( ) subdivisions and distance intervals are obtained from TWINSPAN (see Figure 8.8) and show the partition for riparian flora in three groups (river front, middle and forest edge). 91 Floristic releves from 0 to 100 m away from the river bed (51 releves every 2 m interval across riparian forest and adjacent plant communities) Cut levels 1 2 Releves Group (as in Figure 8.7) Interpretation of each group Distance interval (m) of each group 36 releves between 30 and 100 m from the river bed 21 releves mainly between 58 and 100 m from the river bed 15 releves between 0 and 29 m from the river bed 15 releves mainly between 30 and 57 m from the river bed 5 8 9 7 8 7 9 8 6 7 1 4 4 6 6 8 8 9 6 7 7 9 9 1 5 7 9 9 1 7 7 0 9 5 3 9 3 5 3 5 3 1 0 4 5 6 9 5 5 5 9 3 4 3 3 3 3 4 7 1 1 7 5 7 3 5 9 1 1 5 3 7 3 IV III Adjacent plant communities Ecotone zoen between riparian forest and adjacent plant communities ] 57 - 100] [30 - 57] 8 releves between 17 and 29 m from the river bed 2 1 2 0 2 1 2 2 9 7 5 9 7 9 1 3 II Middle of riparian forest [17 - 29] 7 releves between 0 and 16 m from the river bed 1 0 1 0 1 0 0 5 5 1 7 3 1 3 I River side [0 - 16] NB: Numbers 1, 3, …, 99 represent floristic releves at [0-2[, [2-4[, ….,[98-100[ m distance range Figure 8.8 : TWINSPAN analysis of 51 releves across riparian forests and adjacent plant communities. It shows 4 groups of releves correlated with distance from the river bed: (I) at river side, (II) middle of riparian forest, (III) ecotone zone, and (IV) adjacent plant communities. 92 Chapter 8: Structural parameters and floristic composition variation across rivers in Benin 8.4. DISCUSSION 8.4.1. Partitioning riparian forests in three: river front, middle and forest edge The DCA, TWINSPAN and Clusters analysis outputs, and frequency distribution curves showed interesting features regarding plant species ecology at riverside under the study area conditions, and probably for the whole Dahomey Gap region. The horizontal structure of species exhibit complex patterns across the studied rivers, and the underlying factors are still poorly understood. Rivers and riverbanks constitute very complex environmental gradients, with few (if any) factors having an overriding influence on species richness and composition. Probably, the relative importance of any factor varies greatly over the geographic scale chosen (Nilsson et al. 1989). Also, the predicted covariation in species composition and diversity, and the heterogeneity of environmental factors, which is consistent with the intermediate disturbance hypothesis, are documented from several riparian systems (Minshall et al. 1985, Tabacchi et al. 1990, Kellman & Tackaberry 1993). It is generally assumed that interior and edge areas are affected by patch size and shape (Forman & Godron 1986) and a gradient of environmental variables (e.g. moisture, light) will influence the distribution and frequency of tree and shrub species within and across riparian forests (Porembski 2001). Apart from a few emergent species (e.g. Ceiba pentandra, Cola gigantea, Parinari congensis and Manilkara multinervis) the average tree height is often less than 18 m. The more or less distinct forest layers, each receiving progressively less light, are not always seen across riparian forests and adjacent plant communities. This is in accordance with the climax formations of Benin that are open, dry, or dense semi-deciduous forests with a relatively low height. Generally, riparian forests display a physiognomy that is highly variable, though the understorey is generally dense. Therefore they can be termed as relatively low edaphic and hygrophile forests with irregular canopy (Natta et al. in press, chapter 9), compared to upland tropical rain forests where canopy tree heights fluctuate between 30 to 45 m, and emergent trees may reach heights above 50 m (Popma et al. 1988, Whitmore 1990). Our results reveal that when vegetation formations are undisturbed at the riverside, the floristic composition varies gradually across riparian forest and surrounding plant communities (see Figures 8.7 & 8.8). For Forman & Godron (1986) a gradual change is present where the major controlling environmental factors vary evenly (linearly) with distance. This suggests that, over short distance, water availability, humidity and flood frequency (July to December), as well as soil and micro-climatic variables may vary and differently affect both groups of species and individual species at the riverside. Across a Lake Superior tributary in Minnesota (USA), tree composition was correlated with stream gradient, and at least two effects of floods, oxygen depletion in the root zone and soil modification, could influence riparian vegetation (Ericsson & Schimpf 1986). In riparian ecosystems of South-Western USA, ordinations analyses provided evidence that riparian forests are structured along gradients relating to moisture, salinity, disturbance from fire, and community maturity (Busch & Smith 1995). Certain tree species have a greater fidelity for riparian habitats along rivers. The decreasing affinity of tree species (see Figures 8.9; 8.10 and Table 8.3) to water availability from river to riparian forest edges is as follow: Syzygium guineense, Pterocarpus santalinoides, Parinari congensis, Cola laurifolia, Napoleonaea vogelii, Drypetes floribunda, Cynometra megalophylla, Manilkara multinervis and Dialium guineense. Pterocarpus santalinoides, Parinari congensis, Cola laurifolia and Manilkara multinervis were found to be the most frequent species along the Ouémé, Mono, Couffo, Alibori, Sota and Mékrou rivers in Benin (Sokpon et al. 2001). However, these authors have not made any distinction or ranking among riparian forest species. Typical riparian forest species are those always 93 Chapter 8: Structural parameters and floristic composition variation across rivers in Benin present at river front and in central portions of riparian forests, (i.e. the first eight species in Table 8.2). Therefore they are well adapted to high soil moisture and periodic flooding. Meanwhile, the effects of water-related factors across rivers in Benin are not fully elaborated. Apart from the variation in floristic composition with water-related variables, attention has been focused upon the extent of edaphic and micro-climatic effects in riparian fragments flora because of the large perimeter to area ratio that these patches characteristically possess. In Nigeria, Nye (1954) found that the composition of riverain forest species in the immediate neighbouring of streams varies with the soils of the catena. In Belize, understorey light intensity was highest at riparian forest edge, but rapidly decreased towards the forest interior, where light levels were comparable to other tropical rain forest understoreys. Also the distribution of six species across riparian forest indicates that the lightcorrelated spatial patterning of seedlings persists to adulthood (MacDougall & Kellman 1992). These results suggest that the light regime in riparian forests has sufficient variation to support several regeneration and adult strategies, despite their small patch sizes. 8.4.2. Transition zone between riparian forests and neighbouring plant communities In the study area, the boundary or ecotone between riparian forest and surrounding savanna or forest is on average between 30 and 57 m from the river bed. Under low human disturbance, this overlap zone is more characterised by a gradual change in species composition (Figure 8.7) than a change of structural parameters (i.e. stem density, average height, and basal area; see Figures 8.2, 8.3, and 8.4 respectively), and species richness (see Figure 8.6). The most frequent tree species in this interval, with regards to their abundance, are Dialium guineense, Diospyros mespiliformis, Elaeis guineensis, Cola gigantea and Ceiba pentandra. Interestingly, statistical analysis shows that these five species are not typical to riparian forest found along rivers (see Table 8.3 in Annex). Our results match with empirical knowledge from various floras (e.g. Keay & Hepper 1954-1972, Brunel et al. 1984, Berhaut 1967). Dialium guineense and Diospyros mespiliformis reach their optimal growth inside dense deciduous upland forests. Elaeis guineensis, a heliophytic and pioneering species, is naturally abundant in all West and Central African forests (Sowunmi 1999, Maley & Chepstow-Lusty 2001). Also, Ceiba pentandra is a late secondary pioneer species widely distributed in the tropics (Whitmore 1990, Meave & Kellman 1994), while Cola gigantea, typical of the Guineo-Congolian basin, is among the largest tree species of dense semideciduous forests. Another issue revealed by species lateral distribution patterns (Figures 8.9 to 8.12 in Annex) is that the ecotone zone is composed mainly of species from both sides (i.e. riparian forest and adjacent savanna or upland forest). Numerous authors (e.g. Williamson 1975, Simberloff & Gotelli 1984, van der Maarel 1990, Kent & Coker 1992) have shown how one vegetation type grades into another through a transitional or ecotone zone, which concentrates a certain percentage of the two sides’ species. A sharper ecotone often occurs where the amount of an environmental factor changes abruptly. In the West-African savanna region, abrupt borders seem to be controlled by an interplay of large herbivores, fires (mostly lit by humans), deforestation, shifting cultivation and edaphic factors (Porembski 2001). Between 8º30 and 9º N in Central Benin and North East Côte d’Ivoire, the gallery forest/savanna contact is made through a belt either of light demanding species (e.g. Anogeissus leiocarpus, Mitragyna inermis, Alchornea cordifolia) or species from dense semideciduous and open forest (Pouteria alnifolia, Ceiba pentandra, Albizia spp. and Antiaris toxicaria) (Poilecot et al. 1991, Natta & Porembski, see chapter 9). At the Mountain Pine Ridge savanna site in Belize (Central America), the transition zone between riparian forest and savanna is narrow and often dominated by a tall light-loving grass (Tripsacum latifolium) 94 Chapter 8: Structural parameters and floristic composition variation across rivers in Benin that grows in dense swards up to the edge of the forest (MacDougall & Kellman 1992). Fringing (i.e. edge specialist) species, such as species of savanna affinity, may act as a resilient buffer preventing fire intrusion into the riparian forest interior thus conserving species at the riparian habitat core (Meave & Kellman 1994). Forman & Godron (1986) hypothesise that the interdigitation zone is an area both of high total species diversity, since two vegetation types are in close contact, and of low diversity of interior species. Our own data on tree species at riversides could not yet confirm or reject this hypothesis. Beyond the ecotone zone, in adjacent plant communities, the five most frequent tree species are either of large distribution in the Guineo-Congolian basin, Sudano-Guinean, or Sudano-Zambesian regions, but reach their optimal growth in upland forest or savanna woodland. Albizia ferruginea is a common semi-deciduous forest tree. Lonchocarpus sericeus and Millettia thonningii are deciduous tree species usually seen in fringing and semideciduous forest near water. Anogeissus leiocarpus extends from the driest savanna to the borders of the forest zone, usually in moist situations but also in relatively dry situations, while Combretum collinum is a common savanna tree (Keay et al. 1964, Keay & Hepper 1954-1972). 8.4.3. Modelling tree species distribution across rivers Tree stems are not evenly distributed across riparian forests in Benin. The decrease of the number of individuals at the riverside to 50 m away (Figure 8.2) suggests a decrease in tree species density within riparian forests. This is linked to the biology and adaptation of typical riparian forest tree species to recurrent floods and shallow water table. Such natural selection and adaptation are revealed by the fact that all river front species (i.e. Syzygium guineense, Pterocarpus santalinoides, Cola laurifolia and Cynometra megalophylla) are multi-stemmed species. Provided sufficient data, it is eventually possible to get best fitting models for the lateral distribution patterns of plant species at the riverside. Modelling deals with the construction of a manageable system, which is simpler than the reality that is modelled, but which nevertheless shares interesting features and behaviour with the real systems (Tongeren & Prentice 1986). In general polynomial and linear functions are the best-fit models, and low values of R2 should be linked to complex relationships of tree species to environmental variables under natural conditions. These results accredit the thesis of the presence of a riparian vegetation width effect for tree species, which is a pattern where species distributions are related to the width of vegetation formations at riverside. Nevertheless, the effects of riparian forest width, and disturbance on the width, species composition and community structure along waterway types (i.e. streams and rivers) need further investigation in Benin. 8.4.4. Plot size and shape for plant diversity assessment in riparian forests The variation of tree species composition at riverside according to environmental factors has several implications upon sampling designs, designated at assessing riparian plant diversity. The primary criteria of site selection should be appearance (i.e. physiognomy), species composition and dominant species encountered at riverside. According to the research objectives, random or systematic sampling designs are suitable in continuous and dense vegetation belts by waterways. On the contrary, stations for data collection have to be purposely selected in dense riparian forest stands to avoid bare land or degraded plots. Also, sites designated for drawing species-area curve of riparian forests should be carefully selected in stands uniform with regards to physiognomy and species composition, and avoid ecotone 95 Chapter 8: Structural parameters and floristic composition variation across rivers in Benin zones where edge species are frequent. Including the transitional (i.e. ecotone) zone at a riverside could cause the curve to rise with increasing plot size. When all terrestrial plants in all stages are included, plot sizes of 500 to 1000 m2 are acceptable for classification of forest relevés (Hommel 1990). Field plots (50 m by 10 m) were established in forests fragment forests in Southern Togo (Kokou et al. 2002). In Benin, rectangular plots of 500 or 1000 m2 were used to collect floristic data in several edaphic forests (Sokpon et al. 2001). In Central Benin, a 500 m2 plot size was taken as the minimal plot size of riparian forest (Natta 2000). From the results presented above, circular and square plots may not always cover the whole riparian forest width (i.e. in average 27 to 43 m wide). Circular plots in particular in riparian forests, are characterised by an unequal probability of sampling species, because our results show that species and stems are not evenly, nor randomly distributed in riparian forests (see also Ericsson & Schimpf 1986, MacDougall & Kellman 1992, Stromberg et al. 1996, Natta 2000, Porembski 2001, Natta & Porembski, in press). Therefore circular plots will tend to oversample stems and species of the central portion of riparian forests compared to those at the river front and riparian forest edge. They also do not guarantee a total coverage of the riparian forest width at every stand. In homogeneous strip-like riparian forest fragments, rectangular plots, with varying length and width, fit with any shape of the waterway (see also Begon et al. 1986, Goudiaby 1998). 8.4.5. Optimal riparian forest width to be protected under Benin conditions Riparian forest width varies from river to river, and along a single river system. This has several vital functional implications. The Benin forest law (no. 93-009 of July the 2nd of 1993), does not allow clearance of wood and shrubs within 25 m from both sides of any water course and stretch of smooth water (article 28). This limit was not set up based on scientific data. From field observations, Natta et al. (2002) suggested a limit of 100 m. Can we answer the important question: How wide should riparian forest corridors be to fulfil their vital ecological functions? A prudent approach for riparian biodiversity conservation is to preserve not only the full range of river front, interior and edges species of riparian forests (i.e. at least 57 m from the river bed), but also parts of the adjacent upland forest or savanna as buffer zone. Moreover dense and continuous canopy at the riverside can extend to more than 100 m (e.g. Samiondji, Idadjo, Pénéssoulou, Yarpao, etc). This could guide legislators to update the optimal threshold distance to be protected at each riverside at 100 m (i.e. four time the actual distance). Overseas studies have shown the importance of the issue of optimal distance to be protected at both waterway sides. In South East Brazil, the optimal width that allows RFs to fulfil its multifunctional roles (e.g. sediment retention and improvement of water quality) was found to be 52 m, which was actually wider and more efficient than to the official legal recommendation of 30 m (Sparovek et al. 2002). In South East Australia, Dignan & Bren (2003) found that the fixed minimum non-disturbance width of 20 m along all permanent streams does not provide maintenance of the light regime that could protect riparian forests values, instead they suggest forested buffers 70 to 100 m wide. 8.5. CONCLUSION The present study investigates the structure, floristic composition variation, and spatial distribution of tree species across riparian forests and their adjacent plant communities at riverside. Horizontal and vertical structures of tree species exhibit complex patterns at riversides. On the one hand, tree stems are characterised by an uneven distribution across riparian forests, on the other hand height and basal area variations at riverside do not show 96 Chapter 8: Structural parameters and floristic composition variation across rivers in Benin any clear patterns. The numerical analysis confirms a gradual variation of the floristic composition at riverside, across riparian forests and neighbouring plant communities. This floristic and abundance gradient may correspond to a change of water-related, as well as soil and microclimate variables. This research not only confirms empirical knowledge about the ecology of certain riparian forest and surrounding vegetations plant species, but also makes a distinction between river front, central portion and edge species of riparian forests. Criteria (e.g. average distance range and models) are provided for the most common tree species along rivers. These results suggest a partitioning of riparian forests in three habitats among tree species along the horizontal gradient. In the present paper we could only draw conclusions about the distribution pattern of 18 tree species out of 106 encountered at the riverside. Investigations on the remaining species should continue. An implication of our results for riparian forest phytodiversity assessment is that plot size, shape and layout in the terrain should take into account the river front, central portion and riparian forests edge species. Due to the non-coverage of the whole riparian forest width and unequal chance of species and stems to be sampled, circular and square plots are not suitable for structural parameters and phytodiversity assessment in riparian forests. Instead rectangular plots with varying length and width, and covering the whole cross section of riparian forest appear to be the most suitable under the study area conditions, and probably for savanna regions. The present study also provides scientific guidelines for an improvement of the forest law regarding the distance to be protected at riverside, i.e. 100 m instead of 25 m. 97 Chapter 8: Structural parameters and floristic composition variation across rivers in Benin Annex Table 8.3: Summary of the t-test outputs. Abundance distribution curves are plotted in Figures 8.9, 8.10 and 8.11. Species 1-Syzygium guineense 2-Pterocarpus santalinoides 3-Parinari congensis 4-Cola laurifolia 5-Napoleonaea vogelii 6-Drypetes floribunda 7-Cynometra megalophylla 8-Manilkara multinervis 9-Dialium guineense 10-Diospyros mespiliformis 11-Elaeis guineensis 12-Cola gigantea 13-Ceiba pentandra 14-Albizia ferruginea 15-Combretum collinum 16-Lonchocarpus sericeus 17-Millettia thonningii 18-Anogeissus leiocarpus Abundance Position (m) from river bed (± SD) 72 213 32 181 25 39 70 25 107 58 21 21 35 22 23 32 40 114 3.66 ± 4.97 7.43 ± 8.6 8.28 ± 5.94 14.15 ± 9.8 14.27 ± 6.51 22.55 ± 12 22.07 ± 20.12 27.39 ± 18.16 30.6 ± 22.59 52.54 ± 35.61 29.93 ± 18.49 39.53 ± 19.18 49.39 ± 24.41 49.41 ± 19.93 53.17 ± 14.31 60.93 ± 24.34 61.36 ± 22.77 71.11 ± 22.96 Compared with Riverside limit (16m) Riverside limit (16m) Riverside limit (16m) Riverside limit (16m) Riverside limit (16m) Middle limits (17-29m) Middle limits (17-29m) Middle limits (17-29m) Riparian forest limit (38m) Riparian forest limit (31m) Riparian forest limit (26m) Riparian forest limit (36m) Riparian forest limit (30m) Riparian forest limit (29m) Riparian forest limit (32m) Riparian forest limit (32m) Riparian forest limit (43m) Riparian forest limit (34m) t-calculated 21.05 14.53 7.34 2.53 1.32 3.35 2.87 2.41 2.81 4.05 0.75 0.77 4.04 4.21 3.87 5.81 3.99 14.87 df 142 424 62 360 48 76 138 48 212 114 40 40 68 42 44 62 78 226 Probability p<0.001*** p<0.001*** p<0.001*** p = 0.0117* p = 0.19 NS p = 0.0012** p = 0.004** p = 0.019* p = 0.005** p < 0.001*** p = 0.45 NS p = 0.44 NS p = 0.001*** p = 0.001*** p < 0.001*** p < 0.001*** p = 0.0014** p < 0.001*** Results Typical riverside species Typical riverside species Typical riverside species Typical riverside species Typical riverside species Riparian forest middle species Riparian forest middle species Riparian forest middle species Riparian forest edge species Ecotone zone outside riparian forest Ecotone zone outside riparian forest Ecotone zone outside riparian forest Ecotone zone outside riparian forest Ecotone zone outside riparian forest Ecotone zone outside riparian forest Adjacent vegetation species Adjacent vegetation species Adjacent vegetation species SD = standard deviation; *** highly significant (p<0.001); ** significant at p=0.01; * significant at p=0.05; NS = non significant at p = 0.05 98 Chapter 8: Structural parameters and floristic composition variation across rivers in Benin Abundance of stems (%) 50 (a) Syzygium guineense : river side species y = 0.0044x 2 - 0.5564x + 15.01 R2 = 0.3629 40 30 20 10 0 0 10 20 30 40 50 60 70 80 90 100 Distance from river bed (m) Abunbance of stems (%) 28 (b) Pterocarpus santalinoides: river side species y = 0.0036x 2 - 0.4723x + 13.586 R2 = 0.6475 24 20 16 12 8 4 0 Abundance of stems (%) 0 10 20 30 40 50 60 70 Distance from river bed (m) 32 28 24 20 16 12 8 4 0 10 20 20 Abundance of stems (%) 90 100 80 90 100 80 90 100 80 90 100 (c) Parinari congensis: river side species y = 0.0031x 2 - 0.4131x + 12.223 R2 = 0.4304 0 30 40 50 60 70 Distance from river bed (m) (d) Cola laurifolia: river side species y = 0.0019x 2 - 0.2723x + 9.3742 R2 = 0.7428 16 12 8 4 0 0 10 28 Abundance of stems (%) 80 20 30 40 50 60 70 Distance from river bed (m) (e) Napoleonaea vogelii : river side species y = 0.0016x 2 - 0.2469x + 8.951 R2 = 0.2787 24 20 16 12 8 4 0 0 10 20 30 40 50 60 70 Distance from river bed (m) Figure 8.9: Typical riverside tree species (RS) mainly found less than 16 m from the river bed 99 Chapter 8: Structural parameters and floristic composition variation across rivers in Benin Abundance of stems (%) 28 (a) Cynometra megalophylla : middle of riparian forest species y = 0.0011x 2 - 0.1738x + 7.0832 R2 = 0.2437 24 20 16 12 8 4 0 0 10 Abundance of stems (%) 20 20 30 40 50 60 70 Distance from river bed (m) 80 90 100 (b) Drypetes floribunda: middle of riparian forest species y = 0.0002x 2 - 0.079x + 5.4468 R2 = 0.2142 16 12 8 4 0 0 10 20 30 40 50 60 70 80 90 100 Distance from river bed (m) Abundance of stems (%) 12 (c) Manilkara multinervis : middle of riparian forest species y = -0.0524x + 4.6312 R2 = 0.2177 10 8 6 4 2 0 0 10 20 30 40 50 60 Distance from river bed (m) 70 80 90 100 Figure 8.10: Typical species of the middle of riparian forests (MS) mainly found between 17 and 29 m away from the river bed 100 Chapter 8: Structural parameters and floristic composition variation across rivers in Benin Abundance of stems (%) 10 (a) Dialium guineense : riparian forest edges species y = 0.0001x 2 - 0.0565x + 4.4502 R2 = 0.3536 8 6 4 2 0 0 10 20 Abundance of stems (%) 10 90 100 6 4 2 0 10 20 20 Abundance of stems (%) 80 (b) Diospyros mespiliformis : ecotone zone species y = 0.0013x 2 - 0.1257x + 3.9463 R2 = 0.2164 8 0 30 40 50 60 70 Distance from river bed (m) 80 90 100 90 100 90 100 90 100 (c) Elaeis guineensis : ecotone zone species y = -0.0004x 2 - 0.0042x + 3.602 R2 = 0.0943 16 12 8 4 0 0 10 20 30 40 50 60 Distance from river (m) 70 80 (d) Cola gigantea : ecotone zone species y = -0.0013x 2 + 0.1049x + 1.0386 R2 = 0.1403 16 Abundance of stems (%) 30 40 50 60 70 Distance from river bed (m) 12 8 4 0 0 10 20 30 40 50 60 70 80 Abundance of stems (%) Distance from river bed (m) 12 10 8 (e) Ceiba pentandra : ecotone zone species y = -0.0009x 2 + 0.093x + 0.4664 R2 = 0.0702 6 4 2 0 0 10 20 30 40 50 60 70 Distance from river bed (m) 80 Figure 8.11: Typical species at riparian forest edges (ES) or in the ecotone zone with neighbouring plant communities (i.e. between 30 and 56 m away from river bed) 101 Chapter 8: Structural parameters and floristic composition variation across rivers in Benin Abundance of stems (%) 16 (a) Albizia ferruginea : adjacent plant community species y = -0.0016x 2 + 0.1618x - 0.7031 R2 = 0.1177 12 8 4 0 0 Abundance of stems (%) 16 Abundance of stems (%) 30 40 50 60 70 Distance from river bed (m) 80 90 100 8 4 0 0 16 10 20 30 40 50 60 70 Distance from river bed (m) 80 90 100 (c) Lonchocarpus sericeus : adjacent plant community species y = -0.0006x 2 + 0.0871x - 0.3127 R2 = 0.0709 12 8 4 0 0 16 Abundance of stems (%) 20 (b) Combretum collinum : adjacent plant community species y = -0.0012x 2 + 0.1275x - 0.396 R2 = 0.0726 12 10 20 30 40 50 60 Distance from river bed (m) 70 80 90 100 (d) Millettia thonningii : adjacent plant community species y = -0.0008x 2 + 0.1069x - 0.6855 R2 = 0.1019 12 8 4 0 0 16 Abundance of stems (%) 10 10 20 30 40 50 60 70 Distance from river bed (m) 80 90 100 (e) Anogeissus leiocarpus : adjacent plant community species y = 0.0003x 2 + 0.0187x + 0.0413 R2 = 0.4249 12 8 4 0 0 10 20 30 40 50 60 70 Distance from river bed (m) 80 90 100 Figure 8.12: Typical species of plant communities adjacent to riparian forests (AES), and beyond the ecotone zone (i.e. found constantly beyond 57 m from the river bed). 102 Chapter 9 OUÉMÉ AND COMOÉ: FOREST-SAVANNA BORDER RELATIONSHIPS IN TWO RIPARIAN ECOSYSTEMS IN WEST AFRICA (*) Notulae Florae Beninensis 8 In press, Botanische Jahrbücher. Natta A.K. and Porembski S. Chapter 9: Ouémé and Comoé: forest-savanna border relationships in two riparian ecosystems in West Africa Chapter 9 OUÉMÉ AND COMOÉ: FOREST-SAVANNA BORDER RELATIONSHIPS IN TWO RIPARIAN ECOSYSTEMS IN WEST AFRICA (*) Notulae Florae Beninensis 8 Natta A.K.(1) and Porembski S.(2) (1) (2) Department of Environment Management, Faculty of Agronomic Sciences. University of Abomey-Calavi, FSA/DAGE/UAC 01 BP 526 Cotonou, Benin; aknatta@yahoo.com; Institut für Biodiversitätsforschung, Allgemeine und Spezielle Botanik, Universität Rostock, Wismarsche Str. 8, D-18051 Rostock, Germany; stefan.porembski@biologie.uni-rostock.de (*) In press, Botanische Jahrbücher. ABSTRACT The floristic composition, species richness and structure at river edges are compared between two gallery forests/savanna ecosystems along the Ouémé (Central Benin) and Comoé (NorthEast Côte d’Ivoire) rivers. Although the overall physiognomy of the two gallery forest sites seems similar and they share the most prominent families, there are marked differences in terms of canopy density and height, herb layer density, number of individuals, tree richness and diversity (H’), and species composition. Gallery forest width, top canopy density and height were more developed along the Comoé than along the Ouémé, but the herb layer is more luxuriant at the latter site. Concerning absolute density and basal area Cynometra megalophylla accounted for 9 to 16 %, and 26 to 33 %, respectively in the two sites. At the two sites this species was time and again the most frequent and dominant one at both riverside and in the middle of the gallery forest. A detailed comparison of the three gallery forest plot types (riverside, middle and savanna edge) revealed considerable differences and contrasting results concerning the dominant trees, species composition, the number of species and individuals, average height and diameter, and basal area. Species characteristic for the river front and the central portion of gallery forests were Cynometra megalophylla, Dialium guineense, Cassipourea congoensis, Syzygium guineense and Parinari congensis at both sites, with Pterocarpus santalinoides being particularly abundant along the Ouémé river. The savanna edge has the most distinct floristic composition with Pouteria alnifolia at both sites; Antidesma venosum, Fagara zanthoxyloides along at the Comoé site; and Ceiba pentandra, Albizia spp. and Antiaris toxicaria at the Ouémé site. The difference between the plot types concerning species diversity (H’) was not statistically significant. The present study shows the variability and complexity of ecological processes between and within gallery forests sites. Key words: gallery forest, forest edge, savanna, core species, Benin, Côte d’Ivoire, West Africa. 9.1. INTRODUCTION Gallery forest (also riparian forest) is one of the major vegetation formations, which underlines the outline of waterways in savanna or in forest-savanna mosaic ecosystems (MONNIER 1990, NATTA 2000). They sustain a type of vegetation that is distinct from the 105 Chapter 9: Ouémé and Comoé: forest-savanna border relationships in two riparian ecosystems in West Africa surrounding areas in species composition and vegetation structure (POREMBSKI 2001). Although linear in shape and small in area, as compared to large extents of dense forests, the role played by tropical gallery forests in fluvial processes, and animal and plant species conservation is well documented (FORMAN & GODRON 1986, DECAMPS et al. 1988, BAKER 1990, LEVAUX 1990, MEAVE et al. 1991, ROGGERI 1995, MEDLEY 1992, THOMAS 1996, PIEGAY 1997, LEINARD et al. 1999, VAN ETTEN 1999, NATTA et al. 2002). The study of West-African gallery forests has hitherto been limited (POREMBSKI 2001). So far, most of the published work concerns floristic and structural characteristics of genuine or single gallery forests along specific waterways and the gallery forest/savanna edge relationships, both in species composition and structure, were little investigated. The Ouémé and Comoé, which are among the largest rivers in Benin and Côte d’Ivoire respectively, with their forested banks harbour a rich birdlife and serve as major habitat for primates and other animals of many kinds. Moreover, they are important extrazonal corridors, acting as refugee ecosystem for forest tree species in a fire-prone environment, harbouring species requiring moist habitats, and allowing the migration of species far beyond their zonal distributional area. At the two sites, more than 600 km apart, a field survey shows the presence and dominance of some typical gallery forests tree species, such as Cynometra megalophylla, Dialium guineense, Cassipourea congoensis and Drypetes floribunda. The present paper aims at comparing the floristic composition and structure of two genuine gallery forest/savanna ecosystems, at the same latitude, along the most important rivers in two different West-African countries, Benin and Côte d’Ivoire. The spatial distribution of certain species inside and along the gallery forest edge is investigated as well. 9.2. MATERIAL AND METHODS 9.2.1. Study sites The two gallery forest ecosystems (hereafter GF), which are approximately at the same latitude, are located within protected areas, the Comoé National Park and the Kouffé Mountains (Figure 9.1). They are representative of GFs occurring between 8°30 and 9° N along the Ouémé (Central Benin) and the Comoé (North East Côte d’Ivoire) rivers, which are among the largest waterways in the two countries. At this latitude, the belts of GFs are almost continuous with only a few local gaps. Apart from bush fire traces, the two sites showed no signs of human impact. Their climates are similar in many points, particularly average rainfall (1100 - 1200 mm) and a pronounced seasonality with a dry season from November to March/April and a rainy season from May to October. From January to March the dry-hot Harmattan results in daytime temperatures in excess of 40° C whilst air humidity drops below 20%. During the night temperature can fall below 20° C, and the sudden decrease is occasionally accompanied by dew. The hydrological regime of the Comoé and Ouémé rivers is of tropical transition type, characterised by a unique rise in the water level in AugustSeptember-October. At the end of the dry season, in March/April, the two rivers usually stop to flow but large water holes are still present. The Ouémé river GFs at the latitude of Ouèssè are located at the Eastern limit of the Kouffé Mountains forest reserve that lies along the boundary between the Sudanian regional centre of endemism and the Guinea-Sudanian regional transition zone. It is moister than the protected areas of Northern Benin, so its plant and animal species are more diverse. Despite many decades of neglect, the reserve still contains dozens of small patches of GF and dense moist forest. These remnants appear undisturbed internally, and are therefore of very high biodiversity value (PGRN/IUCN 1994). GF ecosystems and the surrounding vegetation are among the least degraded in the Sudano-Guinean zone. This zone is characterised by the 106 Chapter 9: Ouémé and Comoé: forest-savanna border relationships in two riparian ecosystems in West Africa decrease of rain and merging of the two peaks of rainfall typical of the Guinean region of Southern Benin. The mean annual temperature is 28° C and the elevation ranges from 190 to 390 m a.s.l. Dense semi-deciduous and dry forests represent climax vegetation formations at this latitude. The maximum GF width may reach 200 to 300 m, but the dense and continuous canopy is generally less than 100 m wide. The investigation of GF along the Comoé river focused on the South-Western part of the Comoé National Park (CNP) where semi-deciduous forest represents a large proportion the natural vegetation. Meanwhile, due to annual burning, the vegetation consists mainly of different savanna types (>85%), forests islands (4.4%) and gallery forests (2.3%) (FGU KRONBERG 1979). The climate of the Central and Southern areas of the CNP can be assigned to the ‘secteur sub-soudanais’, which is of tropical sub-humid type (ADJANOHOUN & AKÉ ASSI 1967, GUILLAUMET & ADJANOHOUN 1971, ELDIN 1971). Mean annual temperature is in the range of 26-27° C. March is the hottest month with a mean daily temperature of 37° C. The lowest temperature (15° C) is recorded in January. The river Comoé runs through the park for about 200 km and drains with its tributaries some 78,000 km2 (POILECOT et al. 1991). Permanent and semi-permanent water occurs in many places. The width of GFs may reach up to 600 m, but at the study site, they attain a maximum width of 130 m (POREMBSKI 2001). NIGER MALI 14° Niamey BURKINA FASO Bamako Ouagadougou 12° BENIN OuèssèBétérou NIGERIA 10° Comoé TOGO GUINEA COTE D’IVOIRE 8° GHANA Yamoussoukro Lomé PortoNovo 6° Accra LIBERIA GUINEA GULF States limits 4° Capital 8° Major rivers Scale : 1/ 4 147 000 Study areas 0 6° 125 250 Km 2° 4° 2° 0° 2° 4° 6° Figure 9.1: The Comoé National Park and Kouffé Mountains gallery forests sites in Côte d’Ivoire and Benin in West Africa. 107 Chapter 9: Ouémé and Comoé: forest-savanna border relationships in two riparian ecosystems in West Africa 9.2.2. Data collection and analysis The investigation and comparison of the two sites are based on 9000 m2 of GF (i.e. 4500 m2 per site). Floristic data were collected through phytosociological survey in plots (300 m2 to 500 m2), laid out along a gradient from the stream-side to the border with the savanna. Three plot types were considered: riverside, middle of GF and savanna edge. Per site and at each plot type 1500 m2 were surveyed. All trees and lianas of a diameter at breast height (dbh, at 1.3 m above the ground level) of at least 5 cm were recorded and identified. Nomenclature of the species recorded follows KEAY & HEPPER (1954-1972), BRUNEL et al. (1984), BERHAUT (1967), and LEBRUN & STORK (1991-1997). Tree (i.e. dbh ≥ 5 cm) species diversity for each plot type was calculated using the Shannon-index H’. A t-test based on the Shannon index (H’) and its variance was used to test for differences in alpha-diversity (including evenness) between the plot types, and the two GF sites. This method was successfully used on riparian forests of Central Benin (NATTA 2000) and in the CNP (POREMBSKI 2001). For further details see MAGURRAN (1988). The floristic similarity of the two GF sites was assessed through the Jaccard and SØrensen coefficients. 9.3. RESULTS 9.3.1. General characterisation of the Ouémé and Comoé GF sites GFs investigated along the Comoé and Ouémé rivers have generally a well-marked boundary with the surrounding savanna, which is regularly burnt in the dry season. Elsewhere the contact is more abrupt, with a grass layer bordering the GF. The two GF sites show no clear evidence of stratification, with the different strata transgressing into each other. Here and there, there is a continuum of the top layer canopy with adjacent dense semi-deciduous or dry forests, but ground cover of the dominant trees is always lower at GF fringe than inside. At the Comoé river site, the herbaceous stratum at GF/savanna border is dominated by Vetiveria fulvibarbis, Loudetia simplex, and among the important trees there are Anogeissus leiocarpus, Mitragyna inermis, Holarrhena floribunda, Tamarindus indica, Acacia sieberiana. The interior of this GF is made up of species such as Cymonetra megalophylla, Cola cordifolia, Manilkara multinervis, and Diospyros mespiliformis at a height of 25-30 m. Occasionally, emergent trees such as Ceiba pentandra, reach a height of up to 40 m. The lower storeys consist of smaller trees (e.g. Dialium guineense, Drypetes floribunda, Cassipourea congoensis, Diospyros abyssinica, Celtis brownii) and shrubs (e.g. Oxyanthus racemosus, Rinorea kibbiensis, Tapura fischeri, etc.). Lianas are abundant, with Cnestis ferruginea, Dioscorea spp., Cissus spp., Strophanthus spp., and Strychnos spp. Both towards the Comoé river as well as towards the savanna border, light demanding trees (e.g. Pterocarpus santalinoides, Cassia sieberiana, Christiana africana) can be found. A continuous herb layer is only developed in some gaps. On the contrary, at the Ouémé site Andropogon spp. and Pennisetum spp. are among the dominant herbaceous species at GF/savanna border whilst the most important tree species are Millettia thonningii, Anogeissus leiocarpus, Albizia spp., Combretum spp., Pouteria alnifolia, Acacia sieberiana and Lonchocarpus sericeus. The top canopy tree species inside the GF are Cynometra megalophylla, Parinari congensis, Cola laurifolia, and Pterocarpus santalinoides with a maximum height of 16-20 m. Here, Dialium guineense is among the top layer trees. The emergent tree Cola gigantea, reaches a maximum height of 26 m, which is much lower than those along the Comoé river. The lower storeys consist of smaller trees such as Drypetes floribunda, Lepisanthes senegalensis, Cassipourea congoensis, Salacia pallescens, Vitex chrysocarpa and Morelia senegalensis. Lianas are also abundant: Strychnos 108 Chapter 9: Ouémé and Comoé: forest-savanna border relationships in two riparian ecosystems in West Africa spp., Loeseneriella spp., Alafia spp., Taccazea apiculata, Connarus africana, Paullinia pinnata, Lonchocarpus cyanescens, Cremaspora triflora and Canthium horizontale. The herb layer consists of Achyranthes aspera, Phaulopsis spp., Polygonum salicifolium, Heliotropium indicum, Ruspolia hypocrateriformis, Ruellia praetermissa, and Cyperus spp., and, in contrary to the CNP, is continuous even under dense canopy. 9.3.2. Floristic diversity at the Ouémé and Comoé GF sites The two sites share the most prominent families Leguminosae, Rubiaceae and Sapotaceae, but they have only 16 tree species, out of 88, in common. The Jaccard and SØrensen similarity coefficients are therefore very low, 13.3 and 23.5% respectively. The GFs along the Comoé have higher tree species richness and abundance than those along the Ouémé (see Table 9.1). The t-test, based on the Shannon index that combines species richness and abundance, reveals that the two sites differ highly significantly (t = 8.48, df = 1305, p < 0.001) in terms of tree species diversity. At the same latitude, the GFs along the Comoé are thus more diverse than those along the Ouémé. In absolute density (individuals/ha), Cynometra megalophylla ranks first in the two sites, followed by Cola laurifolia along the Ouémé and Dialium guineense along the Comoé. Table 9.1: Major characteristics of gallery forests along the Ouémé and Comoé rivers. Major characteristics GF along the Ouémé (Central Benin) 50 Tree species richness 624 Number of individuals Most prominent families Leguminosae (8); Moraceae (4); Rubiaceae (4); Annonaceae (3); (species number) GF along the Comoé (CNP) (North East Côte d’Ivoire) 54 906 Leguminosae (8); Euphorbiaceae (7); Rubiaceae (3) and Sapotaceae (3) Sapindaceae (3) and Sapotaceae (3) Most frequent trees Cynometra megalophylla (102); Cola laurifolia (65); Lepisanthes senegalensis (58); Pterocarpus santalinoides (56); Drypetes floribunda (53) and Xylopia parviflora (43) Cynometra megalophylla (84); Dialium guineense (74); Cassipourea congoensis (63); Ouratea affinis (55); Tapura fischeri (54) and Drypetes floribunda (47) Shannon index (H’) Pielou index (E) 3.97 0.70 4.87 (most diverse site) 0.85 The most frequent tree species is Cynometra megalophylla, which accounts for 9 to 16 % of the total abundance, and 26 to 33 % of the total basal area at the Comoé and Ouémé sites respectively (Table 9.2). Table 9.2: Measured variables for Cynometra megalophylla at the Ouémé and Comoé sites. Variables Abundance (%) Basal Area (%) Mean diameter (cm) Mean height (m) Ouémé (Central Benin) Comoé (North East Côte d’Ivoire) 16.3 33.06 19.53 9.48 9.3 26.1 18.6 16.2 109 Chapter 9: Ouémé and Comoé: forest-savanna border relationships in two riparian ecosystems in West Africa 9.3.3. Floristic and structural variability at the riverside, inside GF and at GF edge A detailed comparison of the three GF plot types (riverside, middle of GF and savanna edge) revealed considerable differences concerning the number of individuals and dominant tree and shrub species. At the Comoé GF site, the number of individuals increased from ‘riverside’ (263) and ‘middle’ (317) to ‘savanna’ (326), whilst it decreased (345, 176 and 103 respectively) at the Ouémé site (see Table 9.3). ‘Savanna edge’, ‘middle of GF’ and ‘riverside’ were clearly differentiated when considering the most important species at the two sites and certain species showed particular preferences for specific plot types. Table 9.3: Dominant tree species (with regards to number of individuals, dbh ≥ 5 cm) in different GF plot types at the Ouémé and Comoé sites. (a) The Ouémé GF site at Ouèssè latitude GF sites Plot type Trees abundance Number of tree species Mean height (m) Average diameter (cm) 2 Total basal area (m ) Shannon index (H’) Dominant tree species Ouémé (Central Benin) Riverside 345 32 7.44 13.5 169.6 3.9 Middle of GF 176 26 7.47 13.32 83.2 3.8 Savanna edge 103 22 8.1 16.37 74.1 3.5 Cynometra megalophylla (57) Pterocarpus santalinoides (49) Cola laurifolia (48) Syzygium guineense (36) Morelia senegalensis (22) Drypetes floribunda (20) Vitex chrysocarpa (18) Xylopia parviflora (15) Dialium guineense (11) Cynometra megalophylla (29) Lepisanthes senegalensis (28) Xylopia parviflora (27) Cola laurifolia (18) Drypetes floribunda (10) Dialium guineense (9) Manilkara multinervis (7) Morelia senegalensis (6) Vitex chrysocarpa (6) Drypetes floribunda (23) Lepisanthes senegalensis (19) Cynometra megalophylla (17) Manilkara multinervis (7) Ceiba pentandra (5) Albizia ferruginea (3) Antiaris toxicaria (3) Dialium guineense (3) Pouteria alnifolia (3) (b) The Comoé GF site in the South-West of the Comoé National Park GF sites Plot type Trees abundance Mean height (m) Average diameter (cm) 2 Total basal area (m ) Number of tree species Shannon index (H’) Dominant tree species Comoé (North East Côte d’Ivoire) Riverside 263 8.82 14.2 101.4 27 2.69 Middle of GF 317 9.21 14.7 112.3 23 2.41 Savanna edge 326 8.44 13.3 98.5 34 2.5 Cynometra megalophylla (25) Syzygium guineense (15) Parinari congensis (15) Cassipourea congoensis (14) Manilkara multinervis (14) Drypetes floribunda (12) Mallotus oppositifolius (11) Dialium guineense (9) Lecaniodiscus cupanioides (9) Diospyros abyssinica (8) Cynometra megalophylla (48) Dialium guineense (42) Cassipourea congoensis (37) Drypetes floribunda (35) Tapura fisheri (35) Ouratea affinis (32) Mallotus oppositifolius (26) Holarrhena floribunda (17) Diospyros abyssinica (17) Lecaniodiscus cupanioides (17) Pouteria alnifolia (25) Dialium guineense (23) Antidesma venosum (18) Tapura fischeri (17) Fagara zanthoxyloides (15) Lannea kerstingii (15) Alchornea cordifolia (14) Anogeissus leiocarpus (13) Christiana africana (13) Diospyros mespiliformis (13) At the Comoé site, species characteristic for the central portion of the GF were Cynometra megalophylla, Dialium guineense and Cassipourea congoensis. Cynometra 110 Chapter 9: Ouémé and Comoé: forest-savanna border relationships in two riparian ecosystems in West Africa megalophylla was also abundant along the Comoé river front where Syzygium guineense and Parinari congensis were likewise frequent. At the Ouémé site the difference between ‘river edge’ and ‘middle of GF’ is best shown by a decrease of abundance of certain species such as Cynometra megalophylla, Lepisanthes senegalensis, Cola laurifolia, Drypetes floribunda, Morelia senegalensis, Vitex chrysocarpa, Dialium guineense, Xylopia parviflora and Manilkara multinervis. Meanwhile, Pterocarpus santalinoides and Syzygium guineense are particularly abundant along the Ouémé river. ‘Savanna edge’ has the most distinct floristic composition with Pouteria alnifolia at both sites; Antidesma venosum, Fagara zanthoxyloides along the Comoé river and Ceiba pentandra, Albizia spp. and Antiaris toxicaria at the Ouémé site. Contrasting results were obtained concerning the tree species richness and diversity (H’) between the three plot types. For example, the number of species was highest at ‘savanna edge’ and ‘riverside’ plots respectively for both Comoé and Ouémé. On the contrary, species diversity was at a maximum in the ‘riverside’ plots, but minimum diversity was reached in the ‘middle’ plots at the Comoé site and at ‘savanna edge’ at the Ouémé site. However, the difference between the plot types concerning species diversity (H’) was not statistically significant (at p = 0.05) in both GF sites. Concerning the structure variability at the three plot types, contrasting results were obtained as well. Along the Comoé, both ‘middle’ and ‘riverside’ attained higher values for tree height, diameter and basal area (m2/ha) than ‘savanna edge’, where tree height was clearly lower. Likewise, maximum tree height (30 m) was found for an individual of Ceiba pentandra occurring in ‘Central’ plots. On the contrary, at the Ouémé site, ‘savanna edge’ plots had the highest average diameter and height. Total basal area decreased from the ‘riverside’ to ‘savanna edge’ plots. The tallest tree, Cola gigantea (26 m), was found at the GF/savanna boundary. 9.4. DISCUSSION With regard to the most prominent families of the Comoé and Ouémé GFs, there are some similarities with several tropical GFs and dense forests. Leguminosae, Euphorbiaceae, Rubiaceae and Annonaceae are the most species-rich families always found in Benin riparian forests (NATTA et al., in prep.). In the Dja fauna reserve (Cameroon) Euphorbiaceae, Rubiaceae, Annonaceae, Meliaceae, Caesalpiniaceae, Sapindaceae and Sapotaceae have the highest tree species richness (LEJOLY et al. 1996, SONKÉ & LEJOLY 1998). In a dense 2-ha forest at Kade (Ghana) Leguminosae was the best represented family (SWAINE et al. 1987). MEAVE & KELLMAN (1994) found similar results in a 1.6 ha of riparian forest in the Mountain Pine Ridge (Belize). They further found similarities between this riparian forest flora and the flora of neotropical continuous rain forests. The three tropical rain forest blocks are abundant in Leguminosae, Annonaceae, Euphorbiaceae, Lauraceae, Myristicaceae, Rubiaceae and Sapotaceae (WHITMORE 1990). The floristic dissimilarity between the two West African sites is illustrated by the comparison of the Shannon indices, which combine species richness and abundance. The higher diversity of the Comoé site over the Ouémé site can be attributed to the regional flora and vegetation diversity. Benin is located in ‘the Dahomey gap’, a discontinuity of the West African rain forest belt, which is the product of erratic rainfall, topographic and oceanographic interactions (JENIK 1994). This gap includes essentially the drier types of the Guineo-Congolian forest belt in South-Eastern Ghana, Southern Togo and parts of Southern Benin (ERN 1988). So in Benin, savanna and agro-ecosystems are widespread, and gallery forests are always among the most dense and diverse vegetation formations (NATTA et al. 2002). Côte d’Ivoire is a major rain forest country in West Africa with a much more 111 Chapter 9: Ouémé and Comoé: forest-savanna border relationships in two riparian ecosystems in West Africa diversified vegetation and flora than Benin. Due to their elevated moisture regime (compared to adjacent regions), GFs in the South of the CNP enable the establishment and persistence of plant species from wetter zones such as evergreen rain forests from lower latitudes in Côte d’Ivoire (POREMBSKI 2001). At both sites, species characteristic for river front and central portion of GF were Cynometra megalophylla, Dialium guineense, Cassipourea congoensis, Syzygium guineense and Parinari congensis, while Pterocarpus santalinoides is particularly abundant along the Ouémé. For the Benin riparian forests, Pterocarpus santalinoides is the most common species at all latitudes (NATTA et al., in prep.). The Ouémé and Comoé GFs act as outpost for certain species (e.g. Dialium guineense, Drypetes floribunda, Diospyros abyssinica, Holarrhena floribunda, Fagara zanthoxyloides, Ceiba pentandra, Celtis spp., Mallotus oppositifolius, Spathodea campanulata, Elaeis guineensis, Albizia spp., etc.) which occur in humid forest on plateau between 8° and 9º N or which can be found under more humid climatic conditions in the rain forest region of Southern Côte d’Ivoire. The presence of these species substantiates some hypotheses on the ecological importance of these hygrophilous and edaphically controlled forest formations. They play an important role as migration channels, which provide opportunities for genetic exchange between geographically isolated populations. In many seasonally dry regions, but also in formerly forested rain forest areas, gallery forests possess the character of a refugium for plants and animals (POREMBSKI 2001). FORMAN & GODRON (1986) have documented the importance of gallery forests as routes for movement of plants and terrestrial animals across the landscape. For MEDLEY (1992), gallery forests are the most suitable vegetation for plant species adapted to a moist climatic regime. Tropical riparian forest fragments represent plausible refugia for forest species in fire-prone landscapes, and are known for their potential, albeit limited, to act as safe sites for tropical rain forest species (KELLMAN & TACKABERRY 1993, KELLMAN et al. 1994, MEAVE & KELLMAN 1994, KELLMAN & MEAVE 1997). Cynometra megalophylla, a typical riverine forest species, is the most frequent and dominant species at the riverside and in the middle of GF in the two sites. It may occur occasionally in other forest types (KEAY & HEPPER 1964, HALL & SWAINE 1981). In Benin, apart from riparian forests, this species is dominant only in a periodically flooded forest on vertisol (dark-clayey soil), the Lama forest (AGBANI 2002). Concerning absolute density and basal area Cynometra megalophylla accounted for 9 to 16 %, and 26 to 33 %, respectively in the two sites. In fact, the dominance of a group or a single species is not a new phenomenon within tropical forest in general, and GFs in particular. According to WOLTER (1993) cited by SONKÉ & LEJOLY (1998), dominant species can comprise up to 58 % of the total individuals and 75 % of the basal area. In Benin, 9 tree species (Pterocarpus santalinoides, Cola laurifolia, Syzygium guineense, Dialium guineense, Berlinia grandiflora, Cynometra megalophylla, Elaeis guineensis, Diospyros mespiliformis and Uapaca togoensis) out of 224 contribute to 46.3 % of the total abundance and occupy 55.3 % of the total basal area. Pterocarpus santalinoides alone contributes to 14.8 % and 12.8 % of total abundance and basal area respectively (Natta et al., in prep). In a riparian forest of Belize, MEAVE & KELLMAN (1994) found that the seven most abundant species (2.4% of the species richness) account for 1/3 of the total number of stems. At least three reasons might explain the dominance of Cynometra megalophylla at the two GF sites. Firstly, mono-dominance may develop if disturbance does not occur over long periods and if the regional pool of shade tolerant species is limited (HART 1990). Secondly, according to POREMBSKI (2001), this species may possess particular competitive traits in the specific abiotic conditions of a GF. Thirdly field data show that Cynometra megalophylla occurs along larger waterways South of 9º10’ N (latitude of Bétérou in Central Benin, see NATTA et al., in press Chapter 6) and as far North as Ferkéssédougou (9º30’ N) in Northern 112 Chapter 9: Ouémé and Comoé: forest-savanna border relationships in two riparian ecosystems in West Africa Côte d’Ivoire (AKÉ ASSI, pers. comm.). This species seems to reach its optimal ecological range in the Guinean zone and in sub-humid areas along rivers. It is hypothesized that frequent or lasting environmental stresses, such as floods, result to a decrease in species diversity (at least for trees), which in turn will lead to the increased abundance and dominance of those species that can successfully maintain themselves in isolated or harsh sites. HANSON et al. (1990) described a similar phenomenon in GF patches along rivers in the upper mid-West of the Iowa river (USA). The dominant species along waterways are seen as keystone species and intrinsic components in the functioning of complex and self-organizing riparian ecosystems. They are generally essential for ecosystem resilience (FOLKE et al. 1996). The phenomenon of dominance by a few or single species should also be investigated from the viewpoint of micro-climatic and soils conditions. The plots situated at the forest/savanna boundary were distinguished from the middle of GF and riverside by their physiognomy and structure (i.e. height, diameter, basal area of trees), floristic composition and species richness; but with contrasting trends between the two sites. Along the Ouémé river there is a constant decrease of the number of individuals, tree species, basal area and Shannon indices (H’) from the “riverside’ to ‘middle’ of GF ending at the ‘savanna edge’. At the Comoé site, the number of individuals and species are highest at savanna edges whilst average tree height, diameter and basal area are highest in the ‘middle of GF and at ‘river edge’. Surprisingly, the difference between the three plot types concerning tree species diversity (H’) was not statistically significant at both sites. At both sites, the GF/savanna contact is made through a belt of either light demanding species such as Anogeissus leiocarpus, Mitragyna inermis, Alchornea cordifolia (POILECOT et al. 1991, POREMBSKI 2001) or species from dense semi-deciduous and open forest (Pouteria alnifolia, Ceiba pentandra, Albizia spp. and Antiaris toxicaria). Light regime and flood frequency somehow play an important role for GF edge communities. From the trends detected here, the assumption that a gradient of environmental variables (e.g. moisture, light) influences the distribution and frequency of tree and shrubs species within GF hold true. More detailed data about microclimate and edaphic conditions, as well as seedling establishment, growth, light demand, and other ecological factors for typical GF and savanna boundary species, would enable to draw more definitive conclusions. The present study illustrates the complex variability of ecological processes within gallery forest and between gallery forest and the adjacent savanna. ACKNOWLEDGEMENTS We gratefully acknowledge the assistance and financial support from several institutions and persons. Professor L. Aké Assi (Centre National Floristique de l’Université d’Abidjan) kindly provided plant determinations and information about the ecology of typical gallery forest species in Côte d’Ivoire. The BIOTA-West staff supported additional field check at the Comoé National Park. We would like to thank Professor K.E. Linsenmair (Würzburg, Germany), Dr. F. Fischer (Würzburg, Germany) and Dr. K. Hahn-Hadjali (Frankfurt., Germany) for field assistance in the Comoé National Park. We are also indebted to Professor L.J.G. van der Maesen (Wageningen University, the Netherlands), co-ordinator of the Benin Flora Project, which provides financial assistance for data collection in Benin. We also wish to thank Professor B. Sinsin (Faculty of Agronomical Sciences / University of AbomeyCalavi, Benin) for his guidance during field work and data analysis in Benin. 113 Chapter 10 ASSESSING THE DENSITY OF KHAYA SPECIES THROUGH SIMPLE RANDOM SAMPLING, STRATIFIED SAMPLING AND SYSTEMATIC SAMPLING Natta A.K., de Gier A. and van der Maesen L.J.G. Chapter 10: Assessing the density of Khaya species through SRS, StRS, and SS in Central Benin Chapter 10 ASSESSING THE DENSITY OF KHAYA SPECIES THROUGH SIMPLE RANDOM SAMPLING, STRATIFIED SAMPLING AND SYSTEMATIC SAMPLING Natta A.K.(1), de Gier A.(2) and van der Maesen L.J.G.(3) (1) MSc. DAGE/FSA/UAC 01 BP 526, Tel/fax: 00 229 303084 Cotonou, Benin; aknatta@yahoo.com (2) Professor of sampling design and forest management at ITC - Department of Natural Sciences, Enschede; The Netherlands. degier@itc.nl (3) Professor of Plant Taxonomy, Biosystematics Group, Wageningen University, The Netherlands; jos.vandermaesen@wur.nl ABSTRACT Three sampling designs (i.e. Simple Random, Stratified and Systematic) were used to estimate the density of Khaya senegalensis and K. grandifoliola trees in the Pénéssoulou forest reserve (Central Benin). For both Khaya species, stratified random sampling provided the lowest variance, coefficient of variation, standard error, and sampling error, thus was taken as the most precise and reliable design over simple random and systematic samplings. Densities of trees per ha (dbh ≥ 10 cm) were 5.4 ± 1.9, and 2.3 ± 1.1 for Khaya senegalensis and K. grandifoliola respectively. Stratification with proportional allocation of sampling units has proven valuable in providing precise and reliable density estimates in the study area. On the contrary, systematic sampling gave the least precise and reliable estimates. Other things being equal we recommend stratified random sampling for the assessment of population estimates of Khaya trees in the Pénéssoulou reserve forest. Results have confirmed empirical knowledge about the ecology of Khaya species and shown that the selection of the most precise sampling design, with regards to estimating a given parameter, can be useful for the sustainable management of forest trees in the study area. Meanwhile the assessment of the properties of more complicated sampling designs, such as adaptive sampling and distance sampling, could bring out issues not covered or revealed by the three conventional methods that were applied here. Key words: sampling designs, endangered species, Khaya senegalensis, Khaya grandifoliola, density, Pénéssoulou forest, Benin. 10.1. INTRODUCTION Khaya senegalensis (Desv.) A. Juss. and K. grandifoliola C. DC. (Meliaceae) are among the most best-known endangered species in the West African savanna and pre-forest regions. They are very important for the local economy in the whole West African region. They are well-known for their good timber, fuel wood and charcoal quality (Eyog Matig et al. 2002), medicinal values (Kerharo & Adam 1974, Berhaut 1979, Adjanohoun et al. 1989, Keita et al. 1999, Sokpon & Ouinsavi 2002), as well as for the high nutritive value of their leaves as livestock fodder (Sinsin 1991, Touré 1991, Meurer 1994). Overexploitation of natural stands, weak recruitment of seedlings, and habitats degradation have resulted in a high level of threat and genetic erosion. These two species of Khaya are classified among those threatened most on a world scale (Oldfield et al. 1998). They are mostly found in riparian forests in the savanna regions, but also in semi-deciduous forest and forest outliers, and Northern limits of the Guinea and Congo rain forest blocks (Keay & Hepper 1958). In the Ibadan region and in 117 Chapter 10: Assessing the density of Khaya species through SRS, StRS, and SS in Central Benin Southern Kaduna State (Nigeria), the two Khaya species were once dominant in riparian forests (Onochie 1979). K. senegalensis (also called dry-zone mahogany, or Acajou wood), with a shining foliage, has a wide range up to the Sahelo-Sudanian region in West Africa. In Benin, it spreads from 8°N to the far Northern part, but has disappeared from many natural stands in the guineo-congolese zone (see Sinsin et al. 2002a). It usually grows on clayey-silty or sandy-silty soils. K. senegalensis constitutes the first timber species for Sudano-Sahelian countries (Eyog Matig et al. 2002), and has the highest conservation priority ranking for five West African countries: Benin, Togo, Ghana, Niger and Chad (Adjanohoun et al. 1989, Eyog Matig et al. 2001). It also contributes to the cure of 55 diseases in Benin (Sokpon & Ouinsavi 2002). Meanwhile, missing regeneration indicate that K. senegalensis stands are jeopardised in several sites, particularly in North Benin (Sinsin et al. 2002a). Also the practice of roadside plantation of K. senegalensis, promoted during colonial times for shade and landscape beautification, has all but ceased in West Africa. The remaining trees are heavily lopped for fodder and damaged for the medicinal bark. On the other hand, K. grandifoliola is a large forest tree, and an important timber tree occurring more frequently in dry semi-deciduous forest and forest outliers (Guillaumet & Adjanohoun 1971). It does not extend beyond the Sudano-Guinean zone (Aubreville 1950). This species seems to reach its optimal growth in riparian and deciduous forests in the Bassila and Djougou regions (Central Benin). Current management activities of the endangered species of Benin in general, and Khaya species in particular, is hampered by lack of knowledge about the spatial distribution, and structure related parameters. The Project Flora of Benin has set as major objectives to contribute in filling the gap of knowledge on vegetation formations, as well as the distribution, structure and threats of most Benin’s plants. Although endangered, little is known about the distribution and occurrence of Khaya species in Benin. Apart from ethnobotanical studies, there are only a few publications on Khaya species in Benin. Sinsin et al. (2002a) have documented the spatial distribution of Khaya senegalensis in Benin, and linked the distribution map to levels of human disturbance. So far no distribution map have been published for Khaya grandifoliola. Not only the spatial distribution of the two species still needs to be fully investigated, but also there is no comprehensive and reliable study on density estimation of Khaya species in various ecological regions in Benin. To estimate a given parameter, it is now acknowledged by scientists and forest resource managers that it is desirable to select an efficient sampling design that provides precise estimates. This can eventually be used in conservation or sustainable harvest planning. Likewise the selection of candidate sampling methods must fulfil certain requirements related to the target species and study area characteristics. To study Khaya species we have selected one of the least degraded protected forests, which provides good ecological conditions for the two species. The driving mechanisms of the spatial organisation of the two Khaya species are mostly unknown in the country, and as yet unpredictable in the selected study area. If we assume that the distribution of these species is random, simple random sampling can be applied to be representative of the whole population. This design is the base of most field survey designs, and is also commonly used as a basis of comparison. Also it is sometimes convenient to apply sampling designs that are easy to draw and execute, and that ensures a uniform (i.e. even) coverage of sampling points throughout the whole study area. Therefore systematic sampling was applied. As the study area is relatively flat, we assume that the systematic layout of sampling points is not related to a periodic variation of ecological factors or vegetation types. On the other hand, the vegetation map of the study area allowed us to divide the population into homogeneous and non-overlapping vegetation types (riparian forests, open and dry forests, and savanna), each one having its own 118 Chapter 10: Assessing the density of Khaya species through SRS, StRS, and SS in Central Benin characteristics. The assumption here is that the distribution of the target species may be related to one or more strata, hence stratified random sampling could be more efficient. The stratification of the study area is also suggested by empirical knowledge about optimal vegetation type for each Khaya species (Keay & Hepper 1958, Aubreville 1950, personal observations). This study aims at estimating the density and assessing the precision of population estimates of two Khaya species through three conventional sampling designs, simple random, systematic and stratified random samplings. A research question is put: what are the densities of Khaya species, and which sampling design provides the most precise population estimates? Results from this study are intended to improve knowledge of the ecology of Khaya species, and eventually to sustainably manage the remaining natural stands in the area. 10.2. METHODS 10.2.1. Study area The Pénéssoulou forest (Figure 10.1) was selected for several reasons. It is located in the Sudano-Guinean zone of the Bassila district (Centre West Benin), which is the westernmost part of the semi-deciduous forest, fire-subtype in the sense of Hall & Swaine (1981). According to Adjanohoun et al. (1989), the climate of the region is transitional between the subequatorial climate of the Guinean region (South Benin), and the Sudanian climate (North Benin). It is characterised by progressive fusion of the two peaks of rainfall typical of the subequatorial climate. There are two seasons: a rainy season from mid-April to end-October, and a dry season from end-October to mid-April. Annual rainfall ranges between 1100 and 1300 mm/year, and average temperature from 26 to 32º C. The relative humidity varies from 15 to 99 % in the dry and rainy seasons respectively. The mean annual insolation is 2420 hours, and the average annual Potential Evapotranspiration is 1536 mm. The relief is flat with gentle slopes. This hygrophile enclave stretches out to Central Togo and Central and South East Ghana (Adjanohoun et al. 1989). Forests in the Bassila region are floristically diverse with a emergent stratum, 20 to 25 m high, where species such as Khaya grandifoliola, K. senegalensis, Zanha golungensis, Cola gigantea, Albizia spp., Holoptelia grandis, Antiaris toxicaria, Milicia excelsa, Vitex doniana, Diospyros mespiliformis, Clesitopholis patens, Dalbergiella welwitschii and Anogeissus leiocarpus are frequent (PRRF 1998). Until recently, the whole district was among the least disturbed forest zones of Benin, but now deforestation due to timber gathering is alarming (PRRF 1998, personal observations). Selective tree cutting has resulted in the disappearance of most large specimens of timber trees (e.g. Khaya spp., Antiaris toxicaria, Milicia excelsa, Afzelia africana, Prosopis africana). People now have turned to less valuable savanna trees, such as Isoberlinia doka and I. tomentosa. Within the Bassila region, the Pénéssoulou reserve forest is well known for its dense riparian forest network, which occupies 13% of the total area of the forest, and is the most dense vegetation formation (PRRF 1998). The two Khaya species seem to reach their optimal growth in the Sudano-Guinean zone where this protected area is situated. Also, since 1998 the Pénéssoulou reserve forest experiences a management plan in partnership with local populations. We noticed that current management activities appear to have proven efficient in controlling poaching, trapping, fire, and encroaching cultivation. Other things being equal, this protected forest, which provides good conditions for the survival of Khaya stands, was selected for the present study. 119 Chapter 10: Assessing the density of Khaya species through SRS, StRS, and SS in Central Benin 1° 2° 3° NIGER 12° AFRICA BURKINA FASO 12° Malanville N 11° 11° Natitingou 10° 10° Pénéssoulou forest NIGERIA Parakou 9° 9° Pénéssoulou reserve forest TOGO Major cities 8° 8° States limits Major waterways Bohicon 7° 7° 0 100 km Porto - Novo Cotonou 1° 2° 3° Figure 10.1: The Pénéssoulou forest reserve in Benin. The Pénéssoulou forest covers some 5470 ha, from which a rectangular area (Figure 10.2), 1 km large and 3.6 km long (i.e. 360 ha or 6.6% of the protected forest), was surveyed in that is known to be the densely wooded and least disturbed part. 10.2.2. Sampling designs and field procedures Three conventional sampling designs (Simple Random, Stratified and Systematic) were used to estimate tree densities. The population consisted of 9,000 plots from which 180 were actually selected and surveyed. Plot centres were located in the terrain through a GPS, a clinometer, and tapes. Slope-corrected rectangular plots of 400 m2 were used. In each plot, the diameter of all Khaya trees (i.e. dbh ≥ 10 cm) was measured. Field procedures were specific to each sampling design. Simple random sampling (SRS) is the base of most sampling designs and is widely applied in forest resources inventory. Most other sampling procedures that are designated to achieve greater economy and precision are modification of SRS. The central idea behind SRS is the equal probability selection, i.e. every possible combination of n units is equally likely to be the sample selected. The estimates of population mean and variance are designunbiased, as they do not depend on any assumptions about the population itself (Cochran 120 Chapter 10: Assessing the density of Khaya species through SRS, StRS, and SS in Central Benin 1977, Freese 1990, Thompson 1992, Ott & Longnecker 2001). 180 out of 9000 plots were randomly selected without replacement from their x and y coordinates on the study area map, and then located accurately in the field. The recorded characteristic in each sampling unit is the number of stems of Khaya trees. Units in a systematic sampling (SS) are selected according to a pre-specified pattern throughout the population. Here the study site is divided in 90 blocks, each containing 100 plots. Two starting points were randomly selected in the first block, resulting in two sets (primary units) of systematically selected plots. Therefore 180 plots (secondary units) were surveyed under the systematic design. The chosen design permits the calculation of means and standard deviations. All properties of estimates are obtained based on the design by which the sample of primary units is selected (Thompson 1992). In stratified random sampling (StRS), the population is divided in strata on the basis of similarity of some characteristics, here vegetation types. The underlying assumption was that the variable of interest (tree density) was related to the different strata. The vegetation map of the Pénéssoulou forest (Figure 10.2), which is derived from satellite images from 1998, was used to determine the size of each vegetation type. The three vegetation types were proportionally sampled with regards to their size (Table 10.1). Within each stratum, a random sample without replacement has been independently selected, and estimates computed per stratum. The strata estimates are combined to give a population estimates. StRS usually offers two major advantages over SRS: it provides separate estimates of the mean and variance of each stratum, and for a given sampling intensity, it often gives more precise estimates of the population parameters than would a SRS of the same size. Towards Djougou N Nioro Pénélan Pénéssoulou Deciduous and open forests Riparian forests Tree and shrub savanna Plantations Towards Bassila Boundary of the protected forest Villages Paved road Secondary road 0 0.5 1 Km Study site in the forest Figure 10.2: Location of the study site in the Pénéssoulou forest (From PRRF/Bassila, 1998) 121 Chapter 10: Assessing the density of Khaya species through SRS, StRS, and SS in Central Benin Table 10.1: Vegetation types and plot characteristics in the study area Vegetation types Plots in Percentage Simple Random Stratified Random (or strata) the site (%) Sampling (SRS) Sampling (StRS) Riparian forests 1620 18 32 Dry & open forests 1800 20 36 Savanna 5580 62 112 Total 9000 100 180 180 Systematic Sampling (SS) 180 10.2.3. Assessment of the precision of population estimates Many authors have used or suggested various estimators (e.g. variance, variance ratio, standard error, sampling error, coefficient of variation, etc.) to compare sampling designs and assess their relative precision, reliability, accuracy or efficiency. An unbiased estimate of variability can be used to assess the reliability of a survey result. Along with getting unbiased estimates come goals of precise or low variance and procedure that are convenient or costeffective to carry out. Standard error, or standard error of the means is considered as the standard deviation among sample means. It gives an indication of the reliability of the mean: the smaller its value, the more reliable the mean. As pointed out by Cochran (1977) and Ott & Longnecker (2001), standard error of the population mean and total can be used to estimate the precision actually attained in a survey that has been completed, compare the precision obtained by SRS with that given by other sampling methods, or indicate how accurate any estimate should be. Likewise confidence intervals (t*standard error) are used to assess the accuracy of the computed estimate (Thompson 1992). Likewise, the ratio of the variance, also called design effect by Kish (1965), obtained from any sampling design over the one obtained from SRS for the same number of units, is used in appraising the efficiency of sampling designs. The coefficient of variation is the standard deviation expressed as a percentage of the mean. It is often used as an index of process or population variability (Ott & Longnecker 2001). According to Snedecor & Cochran (1989), one may judge the success of a statistics, evaluate experiments or compare populations by looking at the coefficient of variation. For Cochran (1977) better precision or lower variance is sometimes reached while minimising the cost, expressed as time or money to locate and enumerate sampling points. As the objective of study was not to assess the efficiency (i.e. lowest variance for a given cost) of the three sampling designs, time involved in data collection was not measured. However, we have used the same number of sampling units (i.e. 180) for the three designs, and have assumed that locating and enumerating these sampling units for the 3 designs cost the same time. Following the authors cited above, we used five estimators (i.e. variance, variance ratio, standard and sampling errors, and coefficient of variation) to compare the precision and reliability of the three designs with regards to density estimation. They were selected because of their well-known utility. Ample explanations and more complicated issues related to the five estimators are provided by classical and specialised statistical publications. Generally the lower (or smaller) the variance, variance ratio, standard and sampling errors, or coefficient of variation, obtained from a given sampling design, the more reliable and precise is the estimated parameter. The combination of two or more lowest values of the selected estimators is much sough-after, and indicates the potential of a sampling design as being the most reliable and/or precise. The formulae that were used to estimate population mean, density, total, variance, variance ratio, standard error of the mean, sampling error, coefficient of variation (see Table 122 Chapter 10: Assessing the density of Khaya species through SRS, StRS, and SS in Central Benin 10.2) are given by Cochran (1977), Snedecor & Cochran 1989, Freese 1990, Thompson (1992), and Ott & Longnecker (2001). 10.2.4. Data processing Data were processed through Excel, using the formulae given in Table 10.2. Table 10.2: Unbiased estimators of population parameters from Simple Random Sampling (SRS), Stratified Random Sampling (StRS) and Systematic Sampling (SS). Estimators Simple Random Sampling (SRS) X 1 n = Mean/plot n ¦ xi i=1 Stratified Random Sampling (StRS) 1 L X st = ¦ N hxh N h =1 Systematic Sampling (SS) y= 1 Mn Variance Variance ratio (design effect) Standard error Sampling error (*) S2 = i − X )2 i =1 n −1 S 2 S 2 =1 S2 n SX = Density/ha Total (i.e. 360 ha for the study area) 2 h L S 1 S (yst ) = 2 ¦Nh (Nh − nh ) nh N h=1 2 (t * S ) / x * 100 % ¦( y − My) i N −n var y = 2 * i=1 n −1 M Nn S 2 ( y st ) S 2 2 2 L n· § * ¨1 − ¸ S = 1 * ª N h S h ¦« © N ¹ X st N 2 h=1 ¬« nh X Coefficient of variation i i =1 n n ¦ (x n ¦y (t * S )/ x X st st var y S 2 § nh ·º ¨¨1− ¸¸» © N h ¹¼» * 100 s * 100 % CV = x s * 100 % CV = x x ± t * S x ( ha ) x ± t * S x (360 ha ) x st ± t * S x st ( ha ) x st ± t * S x st (360 ha ) SX = S2 n n· § * ¨1 − ¸ N ¹ © (t * S ) / x * 100 % X CV = s * 100 % x x ± t * S x ( ha ) x ± t * S x (360 ha ) (*) All confidence intervals are calculated at probability level of 0.05. 10.3. RESULTS Outputs from the three sampling designs are summarised in Tables 10.3 and 10.4 for Khaya senegalensis and K. grandifoliola, respectively. For Khaya senegalensis, stratification resulted in the lowest variance (0.001), ratio variance (0.004), coefficient of variation (18.55 %), standard error (0.040), and sampling error (36.36 %). Therefore it gave the most precise estimate of the population density (i.e. 5.4 ±1.9 trees/ha), and total (i.e. 1945 ± 707 trees/360 ha). Mean and variance per stratum show the absence of Khaya senegalensis in riparian forests in the study area. With Khaya grandifoliola, stratification gave also the lowest variance (0.0006), ratio variance (0.001), coefficient of variation (25.62 %), and standard error (0.024). The sampling error from StRS was a bit higher (50.2 %) than the one obtained from SRS (47.6 %), and this is related to a smaller value of the mean (i.e. 0.095) obtained with StRS compared to 0.172 from SRS. A close look at the formula of sampling error (see Table 10.2) shows that for a given t-value and standard error: the smaller the mean, the higher the sampling error. This is actually the case as: sampling error (StRS) = 1.059*sampling error (SRS). The lowest 123 2 Chapter 10: Assessing the density of Khaya species through SRS, StRS, and SS in Central Benin stratified mean results from the fact that K. grandifoliola is absent from the savannas, which occupy 62 % of the study area. Therefore the mean/plot in savanna is null and this has lowered the overall stratified mean/plot. Because it provided the lowest values for variance, ratio variance, coefficient of variation and standard error (4 estimators over 5), StRS was taken as the most precise sampling for the estimation of K. grandifoliola density. Therefore the most precise estimate of population density and total for K. grandifoliola were 2.3 ± 1.1 trees/ha, and 855 ± 429 trees/360ha respectively. SS appeared to be, in both cases, the least precise sampling design, as in all cases it gave the highest variance, variance ratio, standard error, sampling error and coefficient of variation. Table 10.3: Results from the three sampling designs for Khaya senegalensis Estimated parameters Simple Random Sampling (SRS) Mean/plot (0.004 ha) 0.211 Variance 0.379 Ratio of variance (design effect) Density/ha (*) Total abundance (360 ha) Coefficient of variation in % Standard error of the mean Sampling error in % (*) 1 5.2 ± 2.2 1900 ± 802 291.9 0.045 42.21 Stratified Random Sampling (StRS) 0.216 Stratified mean 0.000 (riparian forests) 0.250 (dry, open forests) 0.267 (savanna) 0.001 0.000 0.364 0.360 Systematic Sampling (SS) 0.270 Stratified variance (riparian forests) 6.396 (dry, open forests) (savanna) 0.004 5.4 ± 1.9 1945 ± 707 18.55 0.040 36.36 16.844 6.8 2450 929.09 1.770 1274.72 * 95% confidence interval Table 10.4: Results from the three sampling designs for Khaya grandifoliola Estimated parameters Simple Random Sampling (SRS) Mean/plot (0.004 ha) 0.172 Variance 0.3221 Ratio of variance (design effect) Density/ha (*) Total abundance (360 ha) Coefficient of variation in % Standard error of the mean Sampling error in % (*) 1 4.3 ± 2.05 1550 ± 738 329.55 0.041 47.66 Stratified Random Systematic Sampling (StRS) Sampling (SS) 0.095 Stratified mean 0.160 0.250 (riparian forests) 0.250 (dry, open forests) 0.000 (savanna) 0.0006 0.3226 0.2500 0.0000 Stratified variance (riparian forests) 2.0897 (dry, open forests) (savanna) 0.001 2.3 ± 1.1 855 ± 429 25.62 0.024 50.21 (*) 95% confidence interval 124 6.487 3.9 1400 929.3 1.011 1275 Chapter 10: Assessing the density of Khaya species through SRS, StRS, and SS in Central Benin 10.4. DISCUSSION Considerable study is often required to be able to select the proper design for a given problem in a given situation. So far such a study, comparing several sampling designs, is the first one done in the Pénéssoulou reserve forest. Own data collected in several riparian forest sites throughout the country showed that densities of Khaya species varied from 1.6 to 18 trees/ha, and 5 to 13 trees/ha for Khaya senegalensis and K. grandifoliola, respectively. Khaya grandifoliola is meanly encountered in the Bassila region, where the Pénéssoulou forest is located. We could not compare our results with other similar studies in various vegetation types in Benin, as published data are non-existent. Field observations confirm empirical knowledge about the ecology of the two Khaya species in the study area. Khaya senegalensis is not observed in riparian forests, the wettest vegetation type in the study area. This suggests that K. senegalensis favours less humid conditions such as savannas and open forests. On the contrary Khaya grandifoliola is not observed in savanna because it favours wetter vegetation types. Therefore, values for mean densities of Khaya species can be misleading if they are not linked to the ecological preferences of each species. Although we have not collected data on the regeneration of these species, observations throughout the whole country suggest that the study area with its current management activities, is comparatively well-off, and has a great potential for the long term conservation of Khaya species. For both species, StRS gave the lowest values for all the five estimators for K. senegalensis, and four out of five (i.e. variance, variance ratio, coefficient of variation and standard error) for K. grandifoliola. Considering variance ratio, stratification has reduced the variance obtained from SRS to about 235 and 545 times for K. senegalensis and K. grandifoliola, respectively. Generally the reduction of sampling variation is obtained by subdividing the population into relatively homogeneous strata. But the efficacy of stratification depends on foreknowledge of the behaviour of experimental material (Snedecor & Cochran 1989.), i.e. an expected relationship between stem density and vegetation type for the species concerned. We have used a recent vegetation map (from 1998), and stratification of the study area in three vegetation types appeared to be effective, in comparison to SRS and far better than SS, for the same number of sampling units. If we assume that 180 sampling units of the three designs cost the same in time to locate and enumerate, then the designs can be compared for efficiency. Under this assumption, we might say that StRS is also more efficient than SRS and SS. Additional field investigations will provide new insights in this issue. Although relatively easy to draw, to execute, and to control, the coverage of the whole population in a systematic way is not a guarantee for a better precision of Khaya species density. The performance of SS in relation to that of StRS or SRS is greatly dependent on the properties of the population (e.g. observations/items in random order, with linear trend, stratification effects, periodic variation, auto correlated, or unpredictable) (Cochran 1977). Following this author we might say that it is difficult to give general advice about the situations in which SS is to be recommended in the study area, because the specific properties of the Khaya species are not yet well known. From the data available, the vegetation map and our own knowledge of the study area, we could not yet detect any particular property of Khaya populations. Further research should document this issue. Once a particular distribution pattern or population property is suspected or has been detected, appropriate surveys designated to getting more precise and cost-efficient estimates are available. It is now acknowledged that if a species is known or can be expected to cluster 125 Chapter 10: Assessing the density of Khaya species through SRS, StRS, and SS in Central Benin (i.e. gregarious distribution of individuals), more efficient adaptive sampling designs are available (Thompson 1992, Thompson & Seber 1996). Also, distance sampling provides biologists with a powerful yet practical method for estimating density of various populations, including plant populations (Buckland et al. 1993). The approach followed here for Khaya species can be tested with other valuable timber trees, such as Milicia excelsa, Antiaris toxicaria, Afzelia africana, Vitex doniana, Diospyros mespiliformis, Albizia spp., Anogeissus leiocarpus, Pterocarpus erinaceus, Holoptelea grandis, etc. Therefore we could get consistent results for most of the timber tree in the study area. Starting from 1998, the Pénéssoulou protected forest management plan has set as objective to harvest 30% of the increment of wood on the basis of a 10 year rotation. Knowledge on the reliability of species density and precision of sampling designs can eventually be useful for the management of tree populations in the study area, if consistent results are obtained for other timber species. Although the Pénéssoulou protected forest management plan has specified the attributions of each stakeholder, there is still room to improve current and future management activities (e.g. number of trees to be harvested). 10.5. CONCLUSION The present study facilitates the choice between several sampling designs for the estimation of the density of Khaya species, which are endangered in Benin. In the study area, StRS with proportional allocation of sampling units was the most precise sampling design in comparison to SRS and SS to estimate Khaya stem density. The mean densities/ha were 5.4 ± 1.9 and 2.3 ± 1.1 for Khaya senegalensis and K. grandifoliola trees, respectively (p = 0.05). Our results also have confirmed empirical knowledge about the ecology of the two Khaya species. If consistent results are also obtained for most timber trees in the study area, this may be helpful to manage (e.g. for sustainable harvest) the remaining natural stands of these valuable trees. Ongoing research should clarify several issues, such as the detection of population characteristics and the assessment of non-conventional sampling designs. This could bring out issues not covered by the three conventional sampling designs compared here. 126 Chapter 11 GENERAL DISCUSSION A.K. Natta Chapter 11: General discussion Chapter 11 GENERAL DISCUSSION 11.1. CHARACTERISTICS OF RIPARIAN FORESTS Riverine landscapes are heterogeneous, dynamic, and biologically and spatially complex (Ward et al. 2002). Within riverine landscapes, riparian forests (RFs) are transitional between terrestrial and aquatic ecosystems and are distinguished by gradients in biophysical conditions, ecological process, and biota. They are portions of terrestrial ecosystems that influence exchanges of energy and matter between aquatic and uplands ecosystems (NRC 2002), and contribute to the diversity and function of both terrestrial and aquatic ecosystems (Acker et al. 2003). At all latitudes, riparian or streamside forests are recognized as distinct components of the landscape because of their unique ecotonal nature, linking aquatic and terrestrial ecosystems (Naiman & Décamps 1990, Gregory et al. 1991, Malanson 1993, Hancock et al. 1996, Cordes et al. 1997). They occupy areas of intense land-water interaction and are known to support concentrated and diverse assemblages of wildlife and plant species (Nilsson 1992, Naiman et al. 1993). This enhanced biodiversity is believed to be related to landscape mosaic, spatial heterogeneity, complex environmental gradients and unique natural disturbance regimes (Robinson et al. 2002), which instigate a wide variety and abundance of resources and substrates (Manuwal 1983, Knopf 1985, Szaro & Jakle 1985, Acker et al. 2003). As pointed out by Brinson (1990), RFs own their dynamics, structure, and composition to river processes of inundation, transport of sediments, hence the erosive forces of water. The characteristic plant species, plant communities and associated aquatic or semiaquatic animal species in RFs are intrinsically linked to the role of water as both agent of natural disturbance and as critical requirement of biota survival. Alternating environmental stress (Hancock et al. 1996), such as periodic flooding, is a form of disturbance to which many of the taxa occurring in riparian communities appear well adapted (Jongkind 1996), by sprouting from remaining root or trunk (e.g. Pterocarpus santalinoides, Cola laurifolia, Syzygium guineense, Cynometra megalophylla, personal observations). According to Acker et al. (2003), flooding apparently promotes complexity at the life form, species composition, and stand structural levels, thus the significance of channel constraint for severity of floods may be different in mountainous versus relatively flat terrain. Frequent floods are known to maintain the native riparian vegetation in a mosaic of different successional stages (Salo et al. 1986, Kalliola & Puhakka 1988, Naiman et al. 1993). Sediments deposited by floods can create initial seedbeds for germination, but also injure seedlings (Levine & Stromberg 2001). Research on the effects of flooding have shown that it increases wood and leaf decomposition rates, and sites that are flooded generally have a lower accumulation of organic matter (Bell & Sipp 1975, Bell et al. 1978, Peterson & Rolfe 1982, Shure et al. 1986, Ellis et al. 1999). As a result RFs tend to be very diverse in species composition and physical structure (Gregory et al. 1991). Also, RFs at either waterway side are often related to gradients of environmental factors perpendicular to the stream, all linked with water availability. Changes in environmental factors across RFs are shown by: (a) decreasing humidity, soil moisture, nutrient level, productivity (Hancock et al. 1996, Ettema et al. 1999); (b) decreasing flood intensity, frequency and duration (Stromberg et al. 1993a, Hupp & Osterkamp 1996, Piegay & Bravard 1997, Swanson et al. 1998); and (c) increasing light availability (Dignan & Bren 2003). 129 Chapter 11: General discussion In the West-African savanna region, riparian forests are the natural wooded vegetation types, of varying successional states, bordering (i.e. lining) fresh water bodies such as rivers, streams, brooks and lakes. Riparian forests occupy 2.37 % of Benin’s the total area of the country, and dense deciduous and semi-deciduous forests 1.05 %. Therefore dense forests in Benin occupy some 3.42 % of total area of Benin (CENATEL 2001). Compared to the figure given by FAO in 1980 (i.e. 0.55 % for deciduous, semi-deciduous forests, and important riparian forests), the updated value seems more reliable because of computerised data collection, rectification and analysis of remotely sensed data. Yet, it shows how big the difference can be when various sources at different periods interpret and map tropical vegetations. 11.2. ECOLOGICAL IMPORTANCE RFs are under the influence of fluvial and geomorphic processes (Baker 1990, Fetherston et al. 1995), and because of short- and long-term in waterways basin hydrological cycles, they are in constant state of adjustment, with different abiotic and biotic components adjusting with different lag-times, producing a complex-response situation (Cordes et al. 1997). Many authors have documented on dynamic processes within riparian ecosystems, and interactions of RFs with waterborne particles and chemicals, microbial assemblages, flood, flood plain, water table, climate and micro-climate, soils, local geology, channel morphology, topography and stream banks (Guillaumet 1967, Hall & Swaine 1976, 1981, Devineau 1975, 1984, van Rompaey 1993, Leinard et al. 1999, NRC 2002). The contribution of RFs to the functioning of rivers and biodiversity protection is recognised worldwide (Adjanohoun 1965, Baker 1990, Tabacchi et al. 1990, Polansky 1994, Lykke 1996). More than other vegetation types, they are well-known for their roles in controlling water runoff and quality, mineral nutrient flows, bank erosion and stabilisation, sedimentation, flood, air and soil pollution, water shading and cooling, sediment filtering, aquifer recharge, carbon storage, erosion control, etc. (Devineau 1984, Forman & Godron 1986, Gregory et al. 1991, Haycock et al. 1993, Harper et al. 1994, Large & Petts 1994, Forniss 1996, Mander et al. 1997, Aide & Rivera 1998). The processes involved in water quality improvement are: sediment retention, preservation of the floodplain channels from excessive erosion, filtering and decomposition of nutrients and water pollutants (e.g. NO3-, see Willems et al. 1997, Groffman et al. 2001), improved soil water infiltration rates, etc. (Gilliam 1994, Lowrance et al. 1997, Trimble 1999). Riparian zones with various types of vegetation are reported to be effective in reducing, removing, or assimilating nutrients, coming from upslope agricultural field via shallow lateral flow (Peterjohn & Correl 1984, Pinay & Décamps 1988, Weller et al. 1994, Hefting & Klein 1998, Ettema et al. 1999). Here, the processes involved include reduced velocity of runoff flow, increased sedimentation of coarse particles, retention of suspended particles by leaf and the soil, plant uptake, microbial immobilisation, nitrification and denitrification (Lowrance et al. 1984, Ambus & Lowrance 1991, Groffman et al. 1991, Ambus & Christensen 1993). Contrasting results were found concerning denitrification in temperate countries: rates of denitrification in RFs were higher than in grassland zones (Hefting & Klein 1998, Haycock & Pinay 1993), and lower than in grassland (Groffman et al. 1991, Schnabel et al. 1996). Maintaining biodiversity is one of the most important functions of RFs, which are, directly or indirectly, home to an abundance of animal life (e.g. fisheries, birds, invertebrates and micro-organisms, amphibian and reptiles, mammal species, etc.). For Johnson et al. (1977), riparian zones perform an essential and unique function in areas where water availability is limited. According to Thomas (1996), typical riparian forest tree species are dependent on river flows, a shallow aquifer, and the community and population structure of 130 Chapter 11: General discussion riparian forests are related to spatial and temporal patterns of flooding. In most cases, they do not occupy large areas (Varty 1990, Monnier 1990, Medley 1992), nevertheless they are important in the conservation of a large range of plants, animals, water sources, soils and watersheds. In the Amazon Basin, numerous streams and rivers provide huge potential for increasing the conservation value of deforested and fragmented landscape through the protection of linear remnants along waterways (de Lima & Gascon 1999). Not only do they constitute a natural habitat or the last refuge for many species, but also they lodge many endemic species and extinction-menaced species (Roggeri 1995), and usually act as routes for movement of terrestrial plants and animals across the landscape (Forman & Godron 1986). Mature riparian forests are known to store large quantities of biomass (above ground, roots, litterfall, woody debris, soil carbon, submerged aquatic vegetation) than younger or degraded stages (Burke et al. 1999, Giese et al. 2003). RFs regulate inputs to streams, including solar radiation, dissolved nutrients, foliage and other food resources (Gregory et al. 1991), and large woods, which add to structural and biological diversity of streams by creating pools and other hydraulic features needed by many fishes, algae, and invertebrates (Sedell et al. 1988, Fetherston et al. 1995, Knutson & Nae 1997, Bragg & Kershner 1999). All these fundamental and critical ecological functions that RFs perform better than uplands and aquatic ecosystems, fall into three major categories: (a) hydrology and sediments dynamics, (b) biogeochemistry and nutrient cycling, (c) species habitat diversity and food web maintenance. 11.3. PLANT COMMUNITIES DIVERSITY IN RIPARIAN FORESTS In addition to being characterized by a variety of plant species, and unique assemblages of plants, compared to uplands and wetlands, RFs in Benin harbour a newly described rare and endemic ornamental Acanthaceae (Thunbergia atacoriensis), which is found along certain streams at hills feet. Compared to other tropical forests, such occurence is not negligible and cannot be underestimated within the Dahomey Gap. The structural complexity and species richness of tropical forests make difficult their study by the traditional phytosociological procedures, but numerical methods have proved successful, especially in demonstrating correlations between vegetation and environmental factors in fairly small areas (Hall & Swaine 1976). Also studies utilising direct and indirect gradient analysis have shown that each plant species is distributed in a unique pattern in the landscape and that this pattern results from the interaction of each species with its physical and biological environment. As result, the composition of plant communities changes more or less continuously along environmental gradients as changing conditions cause species to enter or leave (Whittaker 1975). The direct ordination technique based on environmental variables, does not recommend itself in the present study, which main interests are botany, phytosociology and ecology. Meanwhile, ongoing research should clarify this issue. In the meantime, indirect gradient analysis (here DCA) was applied and environmental gradients were inferred from species composition data. Also it is worthwhile noting that, sometimes classification of vegetation into communities or associations may obscure the true situation in the terrain that is characterised by gradual changes of the floristic composition of plant assemblages over environmental gradients. Traditional methods of plant community ecology usually rely on the dominance of overstorey trees to define species associations and establish plant community boundaries, however, as shown by Sagers & Lyon (1997) species associations in the overstorey are not necessarily good predictors of understorey associations. To overcome this, we have taken all species of the least degraded relevés (presence/absence data file), and no species was down131 Chapter 11: General discussion weighted in the classification analysis. Ordination and classification analyses provided evidence that RFs are structured along gradients relating to moisture, flood conditions, regional variations in climate and relief. These factors have created and maintained several riparian forests plant communities, some of which have not yet been described. Also, microtopographic variations across waterways, created by diverse fluvial process, support floristic richness that would not otherwise occur. The repetitive plant assemblages in riparian forests are made of species that can tolerate greater water table and flow variation, physical disturbance, and hydric stress either due to drought or soil saturation. Elsewhere, e.g. in riparian ecosystems of the South Western USA, salinity, disturbance from fire and community maturity were the most important gradients (Busch & Smith 1995). Also, the distribution of vegetation, in general and forests in particular, within riparian zones has been linked to numerous abiotic factors such as hydroperiod, floodplain landform and sediment types as well as to competition and life history factors (Medley 1992, Dunn & Stearns 1987, Nilsson et al. 1989, Hughes 1990). The type and severity of disturbance also have a major impact on the diversity of RFs plant communities. In East Gippsland in Victoria (Australia), Melick (1990) found that undisturbed riparian rain forests were the most species poor plant community, and disturbed ones the most species rich. In a riparian forest of the Buffalo National River (Arkansas, USA), species associations were influenced by environmental gradients dominated by pH and elevation, but secondary gradients differed among forest layers (Sagers & Lyon 1997). Moreover dispersal (Skoglund 1990, Malanson 1993, Johansson et al. 1996, Malanson & Armstrong 1996), floating (Danvind & Nilsson 1997, Jansson et al. 2000a), as well as colonisation (Auble et al. 1997, Stromberg 1997) strategies of seeds in riparian ecosystems, and the ability of riverbanks to trap waterborne diaspores (Andersson et al. 2000), are known to influence patterns of species richness and abundance of RF and predict vegetation changes. However, diversity patterns of riparian plant community were found to be complicated, with contrasting relationships. 11.4. RIPARIAN FORESTS AND WILDLIFE CONSERVATION Waterways and their dynamic forested banks are known to serve as a focal point for animals of many kinds. They host a diverse and abundant fauna of aquatic, terrestrial and amphibious species (Robinson et al. 2002). Also, studies in very different areas have documented the importance of riparian areas for birds, that these have typically different assemblages and higher species richness than neighbouring areas (Skagen et al. 1998, Rottenborn 1999, Saab 1999, Woinarski et al. 2000). The importance of RFs as critical, seasonally or temporary habitat for a variety of animals is reported by numerous authors (e.g. Knopf et al. 1988, Hunter 1990) and confirmed by our own field observations in Benin. The reproduction, growth and nutrition depend on the presence of more or less dense and safe stands of riparian forests: e.g. nesting in trees along streams and rivers, food consisting of fish or other aquatic animals from inundation plains near water courses. In Benin, animals such as Cephalophus rufilatus, Tragelaphus scriptus, Kobus defassa, Loxodonta africana africana, Python sebae, Crocodylus niloticus, Osteolamus tetrapis, Heliosciurus gambianus, etc. very much like gallery forests and forests edges (Sinsin et al. 1997). Likewise Cercopithecus mona, Cercopithecus nictitans, Cercopithecus diana live in forest areas and large and dense galleries. Cercopithecus erythrogaster erythrogaster, the endemic primate found so far in Benin (Grubb et al. 1999, Sinsin et al. 2002b) shares a part of only its time in riparian forests and forest patches in the Ouémé valley region. Raynaud & Georgy (1969) reported that Syncerus cafer nanus that was usually seen in woodlands and humid sites in the South and Central Benin follows gallery forests up to the 132 Chapter 11: General discussion latitude of Djougou to Bembèrèkè. Tragelaphus scriptus and Cephalophus rufilatus main habitats are forests, its edges and gallery forests. Limnotragus spekei lives in humid or swampy areas, gallery forests and upland forests. Cephalophus sylvicultor lives in dense forest or large and dense galleries. Phacochoerus aethiopicus lives in groups in savannas, gallery and edges of forest. The Potamochoerus porcus was found in dense forest and Guinean savanna and gallery forests as far as in the W National Park along the Mékrou stream. Generally in regions where dense forest does not exist, riparian forests play the role of residual forest and provide, when they are undisturbed, the best conditions for wildlife that normally lives in forest zone. In Central Africa, species such as squirrels (Helioscirus gambianus var. punctatus, Protoxerus stangeri, Paraxerus poensis), anomalure (Anomalurops beecrofti, Anomalure derbianus), and civet (Nandinia binotata) are mainly seen either in dense forest or RFs (Depiere & Vivien 1992). Likewise certain bird species (e.g. Tigriornis leucolopha, Butorides striatus atricapillus, Hagedashia hagedash brevirostris, Stephanoaetus coronatus, Podica senegalensis, Tympanistria fraseri, Centropus monachus occidentalis, Musophaga violacea) depend on RFs, dense forests or stream and river shores with periodic inundation for feeding and nesting. Many studies on wildlife ecology highlight the potential of linear remnant forests along waterways to serve as habitat for small forest vertebrates and suggest they could function as corridors for some species to increase landscape connectivity (Forman 1997, Machtans et al. 1996). Examples are given by de Lima & Gascon (1999) in Central Amazonia for small mammal and litter-frog communities; Burbrink et al. (1998) in Illinois (USA) for frogs and lizards; Darveau et al. (2001) in the riparian zone of the Laurentian mountains (Québec) for small mammals. In the USA, RFs are much more important to bird communities than surrounding habitats because the vegetation is structurally complex and floristically diverse within the vertical and horizontal dimensions (Anderson & Ohmart 1977, Best et al. 1979, Hehnke & Stone 1979, Szaro 1980, Bull & Skovlin 1982, Szaro & Jakle 1985). It has also been shown that a mosaic of riparian woodlands, containing mixture of native tree and shrub species of different classes, is necessary to maintain avian species richness (Farley et al. 1994). 11.5. WATERWAYS ORDER AND SIZE RELATIONSHIP WITH RIPARIAN FOREST BIODIVERSITY Outputs from the DCA and TWINSPAN suggest that waterways can be grouped into two types: large (= rivers) and small (= streams). Field observations also support this simplification of the reality. This analysis was based neither on measured environmental variables nor on previous floristic data and hydrographic network maps. Indeed, physical characteristics of streams vary from low to high order (Brierley & Hickin 1985, Forman & Godron 1986, Billy & Ward 1989), and RF characteristics are known to change with increasing stream order (Swanson & Lienkamper 1982, Swanson et al. 1982, Minshall et al. 1985, Nilsson et al. 1989). As pointed out by Gregory et al. (1991) and Naiman et al. (1992) stream size or order, is an important factor that determines the nature of interactions between geomorphology, fluvial and terrestrial disturbance, and riparian community characteristics. Diversity of RFs and their distinctiveness from upslope forests tend to increase with increasing stream size (Naiman et al. 1993). Therefore we expect concomitant changes in the flora and plant community structure, and associated animals. In Benin, RFs along streams are more species rich than those along rivers. In South West Ghana, Jongkind (1996) found that the floristic difference between large rivers and small streams is striking, and this was probably caused by the much smaller seasonal changes in the water level and the absence of an open space above the water of the small streams (see 133 Chapter 11: General discussion also Devineau 1984 in Côte d’Ivoire). Similar stream order patterns were identified in numerous overseas studies. In South West of Western Australia, RFs along tributaries (low order streams) had higher species richness (at p < 0.05) than those along the largest river (higher order streams) (Hancock et al. 1996). This was related to the more intimate association between streamside and upland communities along first order streams due to less developed environmental gradient away from the usually ephemeral streams. In several RFs in the USA, Naiman et al. (1987) and Lock & Naiman (1998) found that bird communities on large rivers differ from those on small rivers and these differences occur predictably with stream order. Also deciduous vegetation and patch types appeared to be linked to river size through stream valley shape and disturbance history. As waterways size and order have been shown to have a great impact on vegetation types as well as wild- and bird life, ongoing research in Benin should integrate several environmental (abiotic factors) variables (e.g. river order and size, sediment deposition, debris flow, position in the watershed, patch size, RFs width and elevation within the watershed, etc.) in multivariate analysis. Likewise, the concepts of ephemeral, perennial, and quasi-perennial waterways should be re-investigated from field observations (e.g. hill headwaters, stream related to hills, waterways on plateau with or no extended flood plains). 11.6. RIPARIAN FORESTS RELATIONSHIPS WITH UPLAND PLANT COMMUNITIES RFs occur in various climatic, geomorphic, edaphic, hydroperiod and geographic settings, and have similar as well as contrasting features with surrounding upland plant communities, with regards to species composition, structure, and ecosystem diversity. In most savanna regions, narrow bands of vegetation in the vicinity of the river contrast strongly with open forests and savanna woodlands which otherwise dominate the landscape (Woinarski et al. 2000). Numerous studies confirm not only the great richness of RFs over upland forests (Brinson 1990, Nilsson 1992, Nilsson et al. 1994, Spackman & Hughes 1995), but also that the structure of RFs is generally more complex than upland forests (Hancock et al. 1996, Melick & Ashton 1991). It is now accepted that RFs harbour more species of birds (Stauffer & Best 1980, Gates & Giffen 1991, McGarigal & McComb 1992, Larue et al. 1995, Whitaker & Montevecchi 1997), and mammals (Geier & Best 1980, Doyle 1990) than uplands habitats. Roché (1993) found bird species richness in riverine flood plains to be twice that in adjacent uplands, while Wakentin & Reed (1999) documented that the majority of bird species in arid regions of Great Basin (USA) are associated with riparian habitats. Also in the extensive natural landscape of Australia’s tropical savanna, species richness and abundance of birds was significantly greater in riparian forests than in matched non-riparian areas, and riparian vegetations allow many species to extend their distributions into lower rainfall areas (Woinarski et al. 2000). The physical and vegetation characteristics of the streamside area differ from those upslope because of frequent inundation, soil saturation, and physical disturbance of streamside vegetation due to flood flows, mass soil movement, etc. (Brinson 1990, Gregory et al. 1991, Fetherston et al. 1995). On the contrary, along the Helena River and its tributaries, and in the Grampians (Western Australia), RFs were generally less species rich than adjacent upland communities (Enright et al. 1994, Hancock et al. 1996). According to these authors their results were corroborated by other studies in other parts of Australia and South Africa (see Cowling & Holmes 1992), but not in Europe. RFs in the Oregon Coast Range are said to have lower tree density than upland stands (Gregory et al. 1990, Pabst & Spies 1999, Nierenberg & Hibbs 2000). Reasons of such richness pattern are not yet fully investigated. Light penetration, which depends upon topographic position, slope and vegetation characteristics, plays a major role in the ecology of the riparian forest, and protection of the 134 Chapter 11: General discussion light environment is generally seen as an important function of buffers that aim to protect riparian habitat values (Barton et al. 1985, MacDougall & Kellman 1992, Dignan & Bren 2003). The ecotone between the riparian and upland vegetation may be very narrow or gradual depending on the fact that gradients are sharp or not. In the latter case it could be the site of highest species richness because of the presence of species from both communities (Tilman 1988, Hancock et al. 1996). In Benin, field observations show that RFs tree flora is not only more diverse, but also has marked differences in structure, abundance and composition with surrounding upland plant communities (e.g. in Yarpao, Pénéssoulou, Idadjo, etc.). However, this issue was not analysed with environmental data and numerical analysis of the whole floristic data. Therefore we could not yet draw any conclusion about the entire RF flora (i.e. trees, shrubs and herbaceous). 11.7. DEGRADATION OF RIPARIAN FORESTS It is generally agreed that species extinction is largely related to the reduction and fragmentation of their habitats, therefore the protection of species can best be done through protecting habitats (Swaminathan 1990). So, a wide range of plants and animals will benefit from the protection of forest ecosystems along rivers and streams. Despite their importance to biodiversity in many countries, native RFs have been severely degraded. This is attributed to a variety of factors including dam construction, river channelisation, cattle ranching, agricultural development, and recreational development (Décamps et al. 1988, Knopf et al. 1988, Roods & Mahoney 1990, Busch & Smith 1993, Busch 1995, Braatne et al. 1996, Lonard et al. 2000). Dams and water diversions are known to modify surface flow rates, flood periodicity, and sediment and nutrient transport, often to the detriment of riparian plants (Jansson et al. 2000b, Levine & Stromberg 2001). When channelling occurs within riparian systems, removal of sediment and nutrients from surface runoff is less effective (Norris 1993). Modification of the frequency, duration and intensity of floods, draining and lowering of floodplain water tables, contribute to change in RF communities (Stromberg et al. 1996, Crawford et al. 1996, Molles et al. 1998). As sites become drier, width of riparian corridor declines and species composition shifts from obligate and facultative wetland plants to those with xerophytic adaptations (Baker 1989, Szaro 1990), and changes in physiognomy and biomass may also occur, sometimes independently with compositional shifts (Stromberg et al. 1993b, Stromberg 2001). Many authors (e.g. Medina 1990, Stromberg & Patten 1990, Smith et al. 1991, Rood et al. 1995) observed that water reduction impact on riparian tree populations include lowered tree densities, loss of class diversity due to death of susceptible age classes, reduced reproductive output, increased seed size. Changes in water regime inhibit recruitment of native plants, and, thus create ‘functional gaps’ in riparian communities (Levine & Stromberg 2001). Urbanisation on lands adjacent to intact riparian woodland has substantial impacts on riparian bird species richness, density, and community composition (Katibath 1984, Smith & Schaefer 1992, Cubbedge & Nilon 1993, Ohmart 1994, Rottenborn 1999). Likewise invasion of the naturalized shrubs potentially alters competitive hierarchies and disturbance regimes in riparian systems (Busch & Smith 1995, Ellis 2001). Riparian plants are not dependent upon fire for renewal, but fire can influence the composition and structure of riparian ecosystems (Kellman & Meave 1997), in particular the understorey (personal observations), in combination or not with other factors such as flooding (Kellman & Tackaberry 1993, Bendix 1994, DeBano & Neary 1996, Ellis 2001, Everett et al. 2003). In Central New Mexico (USA), the suppression of flooding along RFs has also increased forest floor litter and woody debris, which may have contributed to the increased 135 Chapter 11: General discussion frequency and severity of fires (Ellis et al. 1998, 1999). In tropical savanna regions, when the gallery forests degrade, forest species become sparse, leaving the vegetation open for the savanna species to invade (Lykke 1996). The environmental impacts of such degradations are generally tremendous. For example, the changes in plant physiology, population dynamics, and community structure affect functional roles such as provision of habitat for ripariandependent animals (Stromberg 2001). Many studies have shown that in Benin riparian forests have waned during the last decades. Interpretations of aerial photographs and satellite images show that degradation rates varied from 0.14 to 8.6 during the period 1949 -1998 (Table 11.1). So far, no study has shown an increase in riparian forests area during that period. Table 11.1: Degradation rates of riparian forests in Benin Study area Latitude (°N) Degradation rate (%) Klouékanmè 6°55’ – 7°10 1.58 Dekpo-Lonkly 2 7° – 7°10 Kétou 7°20’ – 7°35 8.59 Dogo 7°30’ – 7°40 3.32 Ouémé-Boukou 7°45’ – 7°55 0.3 Bantè 8°05’ – 8°36 0.14 Wari-Maro 0.51 9°– 9°10 Djougou 3.4 9°45 Natitingou region 10°05’– 10°30 0.8 Péhunco 10°05’– 10°35 1.14 Alibori forest 10°40’– 10°50 3.5 Malanville 3.4 11°45 Period (number of years) 1982-1994 (12) 1982-1994 (12) 1949-1994 (45) 1949-1995 (46) 1949-1998 (49) 1979-1987 (8) 1975-1997 (22) 1949-1975 (25) 1975-1994 (19) 1975-1997 (22) 1975-1998 (23) 1949-1975 (25) Authors Adambiokou-Akakpo (2001) Gnele (1999) Houndagba et al. (in press) Houndagba et al. (in press) Houndagba et al. (in press) Akoègninou et al. 2001 Sounon (2001) FAO (1980) Tenté (2000) Wotto (2001) Arouna (2002) FAO (1980) Periodic removal of trees (either by selective tree cutting or farming) along waterways has modified the structure, composition, hence the functions of riparian forests. Beside farming and illegal tree cutting, livestock animals have a disproportionate effect on riparian forests ecosystems, because they tend to concentrate in areas, which are close to rangelands and water reservoirs. Long-term cumulative effects of domestic livestock grazing involve changes in the structure, composition, and productivity of plants and animals at community, ecosystem and landscape level (e.g. along the Ouémé river at Bétérou). Less often seen disturbances in RFs in Benin are industrial, urbanization and recreational activities. Nevertheless their impact is very destructive. 11.8. FLORISTIC RELATIONSHIPS OF RFS IN BENIN WITH OTHER TROPICAL DENSE FORESTS There are marked floristic affinities as well as differences between RFs in the Dahomey Gap and those in the Guinean and Congo forest blocks. Certain RF plants in the Dahomey Gap are also found in forests along waterways in the Guinean and Congo rain forest blocks (Table 11.2). Among the most important we have Cola laurifolia, Pterocarpus santalinoides, Ficus asperifolia, Entada manii, Leptoderris brachyptera, Psychotria calva, Manilkara multinervis, Synsepalum brevipes, Parinari congensis and Uapaca heudelotii. These species can be termed typical stream and riverside forest plants of West and Central Africa. They also can be seen as specialists of a fixed habitat of limited area, because they are tied to special topographic or edaphic site factors (see Hubbell & Foster 1986). However, riparian habitat specialists are often common where RFs are extensive. 136 Chapter 11: General discussion Among fresh water (swamp) forest species found in RFs in the Dahomey Gap, Syzygium guineense var. guineense, Cleistopholis patens, Cynometra megalophylla, and Mitragyna inermis are common. Floristic data indicate that large numbers of upland forest species in drier parts of the rain forest zones (e.g. Dialium guineense, Diospyros mespiliformis, Elaeis guineensis, Xylopia parviflora, Dennettia tripetala, Hexalobus crispiflorus, Isolona thonneri, Drypetes floribunda, etc.), coexist locally within RF patches in the Dahomey Gap. Also, large specimens of Cola gigantea, Ceiba pentandra, Milicia excelsa, and Antiaris toxicaria are frequent at RF edges. Most of these species are typical of the three drought resistant or dry semi-deciduous forest types of the Guineo-Congolian forest belt (Ern 1988, Hall & Swaine 1981). Generally at RF edges we have a host of open forest, forest outliers or woodland species, such as Albizia spp., Combretum spp., Terminalia spp., which augment RF phytodiversity. Table 11.2: Typical riparian forest species of West and Central African regions Species Cola laurifolia Cynometra megalophylla Entada mannii Ficus asperifolia Gardenia imperialis Irvingia smithii Leptoderris brachyptera Manilkara multinervis Mimosa pigra Morelia senegalensis Napoleonaea vogelii Ouratea glaberrima Parinari congensis Pterocarpus santalinoides Psychotria calva Synsepalum brevipes Syzygium spp. Uapaca heudelotii Xylopia aethiopica Life Geographic Habitat forms* affinity** mPh GC River banks tree in forest and savanna zone mPh GC River banks and fresh water swamp forest species mph GC A climbing forest shrub in streamside forest nph GC Shrub near water mph GC River banks and fresh water swamp forest species mPh PRA Riparian forest tree LmPh GC Climbing shrub in riparian forest mPh TA Riparian forest tree nph Pan Shrub often on sandy bars along rivers mph SZ Shrub of gallery and swamp forests mph G Riparian forest tree or near to sea shore nph GC Erected shrub or small tree in forest by streams MPH GC Evergreen tree on river banks and fringing forest mPh PRA Forest tree on river banks nph GC Erected shrub beside streams in forest mph GC Riparian forest tree mPh River banks and fresh water swamp forest species mPh GC River banks and fresh water swamp forest species mPh GC Riverine forest tree, also in dry forest Source (1),(2),(7),(8) (1),(2),(3),(6),(8) (2) (1),(2),(4),(9) (1),(2),(4) (4),(9) (2) (1),(4) (9) (1),(4),(9) (1),(2) (1) (1),(2),(4),(6),(8),(9) (1),(2),(3),(6),(7),(8),(9) (2),(4) (4) (9) (1),(4),(6),(9) (5),(6) * Life forms follow Raunkaier (1934), Schnell (1971), and Keay & Hepper (1954-1972). ** Geographic affinity of each species (i.e. phytogeographic types) follows White (1986), Keay & Hepper (1954-1972), Keay et al. (1964). (1) Keay et al. (1964); (2) Keay & Hepper (1954-1972); (3) Aubreville (1950); (4) Pauwels (1993); (5) de Koning (1983); (6) Hawthorne (1996); (7) Onochie (1979); (8) Adjanohoun (1968); (9) Schmitz (1988). On the other hand, there are marked floristic differences between RFs in the rain forest zone and those in the Dahomey Gap. Numerous plants bound to waterways in the rain forest zones are absent in the Dahomey Gap. Among the most common riverine plants, we have Brachystegia spp., Cathormion spp., Coelocaryon spp., Cynometra spp., Gilbertiodendron spp., Hymenocardia heudelotii, Millettia spp., Psychotria spp., Sacoglottis gabonensis, etc. (see Keay & Hepper 1954-1972). So far, apart from a few evergreen forest trees adapted to swampy situations (e.g. Cynometra megalophylla, Pentadesma butyracea, Xylopia aethiopica, etc.), no true rain forest species of the Guinean-Congo regions are present in riparian forests in the Dahomey Gap. 137 Chapter 11: General discussion 11.9. STRUCTURAL RELATIONSHIPS OF RFS IN BENIN WITH OTHER TROPICAL DENSE FORESTS In contrast of savanna regions, the physiognomy of RFs in rain forest regions is similar to that of adjacent rain forests. The distinction between the two forest types can therefore only be made through their specific floristic composition and topographic position. In the Zaire basin (Central Africa), RFs that fringe waterways are narrow where the banks are high but expanding into great swamp forests in the sump-lands (Kingdon 1990). The difference between RF in the Dahomey Gap and dense forest in the rain forest zone is well illustrated in terms of tree density, basal area and plant richness. Although our own data indicate that high plant species numbers can be sustained in isolated RFs habitats, tree richness and diversity in RFs are lower in the Dahomey Gap than in the continuous West and Central Africa rain forests (see Table 11.3). Table 11.3: Characteristics of some riparian and rain forest sites in Africa and tropical America Locality Shannon Tree density Basal area TR/ha SR/ha Rainfall Source 2 (mm) index (H’) (stem/ha)* ** *** (m /ha) RIPARIAN FORESTS IN THE DAHOMEY GAP (BENIN & TOGO) South Benin (Samiondji 7°20N) 2.4 726 Central Benin (Ouèssè 8°30N) 3.9 748-785 Benin: various RFs 3.7-4.9 312-665 South Togo: various RFs 1.89 - 41.7±17.4 45.6±15.8 21.8-41 - 27 34 - 249 129-195 - 1200 1150 1100 1000 (1) (1) (2) (3) RIPARIAN FORESTS IN SOUTH AND CENTRAL AMERICA Belize: Río Bravo Brazil: Mobi-Guaçu 476 Belize: Mountain Pine Ridge 766±241 21.9±8.8 56 52 - 1517 1280 - (4) (5) (6) 360-523 523 520 375 461 - 30.5 - 34-70 42 109 138 - 2400 2080 2400 3250 3900 3900 (7),(8) (8) (9) (9) (10) (11) 605-649 376-555 539-565 - 35-43 25.3-36.3 25.5 28.4-32.2 - 69-131 77 86-122 120 < 100 - 3000 2100 3000 2000 1640 - (12),(13) (14) (15) (16) (17) (11) WEST AND CENTRAL AFRICAN RAIN FORESTS Nigeria: Okumu forest reserve Nigeria: Shasha forest reserve Nigeria: Omo forest reserve Cameroon: Bakundu reserve Cameroon: Dja 5.4 Cameroon: Korup 0.65 ha plot (African pleistocene refugia) Gabon: various rain forests Côte d’Ivoire: Yapo Côte d’Ivoire: Taï National Park Ghana: dense forests Ghana: Kade research station African non-refugial sites - * (dbh ≥ 10 cm); ** TR = Tree Species Richness per ha; *** SR = Species Richness per ha. Sources: (1) Natta et al. (in prep.); (2) Sokpon et al. (2001); (3) Kokou et al. (2002); (4) Kellman et al. 1994; (5) Gibbs & Leitão (1978); (6) Meave & Kellman (1994); (7) Jones (1955); (8) Richards (1939); (9) Richards (1963); (10) Sonké & Lejoly (1998); (11) Whitmore (1990); (12) Reitsma (1988); (13) Hladik (1982), (14) Corthay (1996); (15) Dengueadhe (1999); (16) Hall & Swaine (1981); (17) Swaine et al. (1987). Also, most RFs of the Dahomey Gap are characterised by many small-stemmed communities of low height (12-18 m) compared to tropical rain forest canopy height. In rain forest regions, the uniformity in the physiognomy at riversides makes it difficult to differentiate between RFs and adjacent dense evergreen forests. Swamps can be distinguished easily, but lower slopes and riverbank forests, which are fed by groundwater as well, are 138 Chapter 11: General discussion more difficult to distinguish from upland forest (van Rompaey 1993). Therefore from the viewpoint of forest physiognomy, we prefer to use the term ‘riverine forest species’ instead of ‘riparian forest’ for rain forest regions. 11.10. CONSERVATION OF DENSE FOREST SPECIES IN RIPARIAN FORESTS IN THE DAHOMEY GAP: implications for periods of climate change As many as 350 out of 410 (i.e. 85.4 %) species recorded in RFs in South and Central Benin are not specialised to riparian systems, but are rather a subset of the general flora without endemic species. One would expect many rain forest taxa in RFs of the Dahomey Gap, because it is the wettest vegetation formation in this relatively dry savanna corridor. However, present days distribution pattern of plants show that the greater water availability owing to higher water tables in RFs seems not to favour many rain forest taxa. In Northern Australia, riparian vegetation has many elements of monsoonal rain forests, and has probably been an important refugium area for rain forest species in period of increased aridity during the Pleistocene (Russell-Smith 1991), as it was postulated for RFs in neotropical systems (see Meave & Kellman 1994). In Belize and Venezuela, RF fragments often contain a nonspecialised tree flora with species characteristic of continuous forests. Provided that the local diversity of riparian systems is complemented by high regional diversity (this is lacking in the Dahomey Gap), RFs may have provided plausible refugia for tropical forest taxa in drier periods (Meave et al. 1991). Also, studies on the geographic patterns of genetic diversity of certain lowland species in Central America showed that these species might have persisted in riparian zones during the late Pleistocene (Aide & Rivera 1998). Likewise in the Neotropical savanna region, it has been suggested that tropical rain forest taxa may have persisted in small mesic, fire-protected pockets within Pleistocene non-forest communities (Leyden 1985, Colinvaux 1987). For Whitmore (1990) parts of the world’s rain forests that are most rich in species are those that the evidence shows have been the most stable, where species have evolved and continued to accumulate with the passage of time without episodes of extinction caused by unfavourable climatic periods. Therefore the floristic dissimilarity between RFs in the Dahomey Gap and those in the African rain forest zones as well as the presence of only a few rain forest taxa in the former, suggest three hypotheses: 1 - The most recent regression of dense forests in West Africa, and subsequent reopening of the Dahomey Gap, which occurred from the end of the middle Holocene to the beginning of the late Holocene (see Sowunmi 1986, Maley 1991, 1997, 2001, 2002, Tossou 2002) has resulted in a localised extinction of African rain forest taxa in RFs of the Dahomey Gap. Support to this hypothesis is given by e.g. Léonard (1965) and White (1979), who found that the Dahomey Gap constitutes a barrier to the distribution of very few typical forest plant species in West Africa (see also Kingdon 1990). 2 - Even during favourable periods, the ‘Gap’ may have been too much of an obstacle for some species (Kingdon 1990), probably for rain forest taxa. We now known that the shaping of the Dahomey Gap vegetation is linked to historical retreats and invasions of forests, which occurred several times during the Pleistocene and Holocene (Guillaumet 1967, Maley 1989, Dupond & Weinelt 1996, Dupond et al. 2000, Tossou 2002). Therefore the absence of many rain forests taxa in RFs might be the result of cumulative effects of these historical events at the actual location of the Dahomey Gap. 3 - Despite the higher water balance of RFs over upland forests, the environmental stress (i.e. recurrent flooding but also periodic drought) is unfavourable for the establishment and survival of many rain forest taxa in RFs of the Dahomey Gap. 139 Chapter 11: General discussion These hypotheses need to be documented from palynologic records and comprehensive floristic data from little disturbed sites. As RFs of the Dahomey Gap possess a floristic richness lower than those of continuous tropical rain forests (see Table 11.3), and sustain few typical rain forest species, they can not effectively act as refugia for many rain forest taxa in the Dahomey Gap. Nevertheless, a significant number of species from deciduous and semi-deciduous forest types can survive in RFs; therefore they could act as refuge ecosystems for less luxuriant forest types taxa in the Dahomey Gap. In addition to their importance in conserving biological diversity, RFs also provided important ecosystems functions in geological times (Naiman et al. 1993) and may have helped to maintain and create biodiversity in evolutionary time (Aide & Rivera 1998). RFs are typically narrow, but can have high species diversity, and palaeoecological data suggest that this diversity can be maintained for thousands of years, even during dry periods (Kellman et al. 1996). This viewpoint is shared by Kingdon (1990) when he points out that during the most arid periods, the riverine forest trailed over a much larger area than the vestigial jungles of Biafra and Central African uplands. At such times, riverine galleries became by far the most extensive microrefugia of all forest habitats in Africa. For Meave et al. (1991), patches of RFs differ from neighbouring non forests plant communities in that they are protected by a river, have high levels of humidity, and some occur within steep ravines (see also RFs at hills feet in North Benin). These factors contribute to making RFs an important habitat during periods of climate change. RFs are known to have played a crucial role in maintaining biodiversity for more than 10,000 years in the past, so we expect that these forests may be able to do the same in the future (Kellman et al. 1996). The mentioned properties show the potential of riparian forests to act as refuge ecosystem during dry periods in savanna as they are in rain forest regions. 140 Chapter 12 GENERAL CONCLUSION TOWARDS A BETTER PRESERVATION AND REHABILITATION OF RIPARIAN FORESTS IN BENIN A.K. Natta Chapter 12: General conclusion Chapter 12 GENERAL CONCLUSION TOWARDS A BETTER PRESERVATION AND REHABILITATION OF RIPARIAN FORESTS IN BENIN 12.1. RIPARIAN FORESTS AS BIODIVERSITY VEGETATION HOTSPOTS IN BENIN Hotspots of biodiversity are areas particularly rich in species, rare, endemic, threatened species or some combination of these attributes (Reid 1998). In Benin, more than other vegetation types, RFs can be considered as hotspots of biodiversity for several reasons: - they are among the most vulnerable forest formations at all latitudes, yet of high ecological importance. - their high species richness on relatively small areas (patches and narrow corridors). We collected 1/3 of the estimated number of species in the Benin flora in just 20 ha of RFs. At this stage of the investigation on RFs diversity, we might say that 2.37 % of Benin’s total area harbours at least 33 % of the estimated number of plant species in the country. - their numerous species either rare, threatened, or with superior adaptability to a specific habitat. - they are habitat for an endemic plant species in a fire-prone environment. So far just a few endemic plant species have been found in Benin. - they are a vital ecosystem for numerous wild and bird life species. - they protect many water resources all over the country. Riparian ecosystems are essential multifunctional elements of the worldwide ecological network (Mander et al. 1997), and the maintenance of riparian forests is recognised as being vital to the integrity of waterways (Hancock et al. 1996). Given that RFs can support high levels of biodiversity, and have played an important role during periods of climate change over geological time (Kingdon 1990, Meave et al. 1991, Naiman et al. 1993, Kellman & Tackaberry 1993, Meave & Kellman 1994, Kellman et al. 1994, Kellman et al. 1996, Aide & Rivera 1998), their conservation in Benin is of utmost importance with the threat of a rapid increase in human occupation of land and possible change in climate. 12.2. LEGAL PROTECTION OF RIPARIAN FORESTS IN BENIN The uniqueness, vulnerability and diversity of riparian forests are the most important criteria to protect them. In Côte d’Ivoire, 29 out of 46 biological reserves are protected to preserve genuine riparian forests, inter alia, because of their fragile soils in relationship with slope and hydromorphy (Kadio 1999). The uniqueness of individual populations suggests that conservation efforts directed only at species preservation can result in the loss of genetic diversity (Aide & Rivera 1998). Therefore the preservation of endemic animal species (e.g. Cercopithecus erythrogaster erythrogaster, a primate sub-species still found in Benin (Oates 1996, Grubb et al. 1999, Sinsin et al. 2002b), and the ornamental Thunbergia atacoriensis (Acanthaceae), should be coupled with the protection of their habitats. The distribution pattern of the endemic primate sub-species of Benin, suggest the protection of all RFs of the Ouémé valley, which are among its most important habitat. Likewise, RFs at Yarpao (Natitingou district) and many others in the Atacora hills should have higher legal protection priority for their great plant diversity, and because they are remnant habitats for Thunbergia atacoriensis. Likewise the boundaries of certain protected areas (East of Kouffé Mountains; West of Kétou, Dogo and Ouémé-Boukou; and North East 143 Chapter 12: General conclusion of the Pendjari Biosphere Reserve) that coincide with rivers should be extended to include RFs at both riversides. Legislation and policies for the protection or rehabilitation of RFs in Benin, like in many developing countries, must take into account the economic and social costs as well as the environmental benefits at various scales. Before improving the actual forest law, there is an urgent requirement to increase enforcement of existing legislation, which is often unknown or ignored by many stakeholders. As suggested by de Lima & Gascon (1999), proactive educational initiatives are direly needed and should be coupled with legislation enforcement to reduce deforestation of riverbanks and maintain corridors between ecosystems in populated regions. 12.3. STRATEGIES FOR SUSTAINABLE MANAGEMENT OF RIPARIAN FORESTS Cost-effective conservation of biodiversity requires that maximum biodiversity should be protected in a minimum area (Williams et al. 1991). The current degraded status of many riparian forests in Benin represents the cumulative and persistent impacts of unplanned and uncontrolled land uses and poor management of RFs themselves but also of adjacent uplands. Strategies that reflect a spectrum of goals linked to improving the ecological functions and sustainability of riparian ecosystems (i.e. natural integrity and diversity of the system) are needed. For example, strategies that focus on returning the hydrologic regime to a more natural state have the greatest potential for restoring riparian vegetation, and forests in particular that have evolved with and adapted to the patterns of changing flows in waterway environments. As pointed out by Cordes et al. (1997), to maintain existing and restore riparian woodland and to ensure their long-term regeneration, it has become important to understand in some detail both the contemporary and historical relationships between riparian vegetation and hydrological and geomorphological factors. Also, management practices to maintain RF biodiversity need to be tailored to the conditions of each particular area (Hancock et al. 1996), in particular one must take into consideration the types of disturbances that typically affect these forests (Agee 1988). Therefore, activities aiming at protecting and rehabilitating RFs in their watershed should be implemented at the national, province, district and local levels, while taking into account ecological benefits (i.e. physical variables) as well as social costs represented by the loss of arable land (Cooper et al. 1987, Flanagan et al. 1989, Xiang 1993). Because riparian forests perform a more than proportionate number of critical biological and physical functions on a unit area basis, their restoration when degraded, and protection of undisturbed stands can have a major influence on the local and regional environment. Meanwhile, deciding on priorities between rehabilitation and preservation may be difficult. Following Geraghty & Vollebergh (1993), we might say that it is more cost effective initially to maintain systems that are in reasonable condition and build from them, rather than at first attempting difficult rehabilitation tasks of very degraded river sections. The preservation of intact riparian forest stands is vital because they represent valuable reference sites for understanding the goals and efficacy of restoration efforts, and are important sources of genetic material for the reintroduction of native species into degraded areas. The restoration of RFs should refer firstly to the process of removing the degrading factor or reducing its negative effects, and secondly repairing the condition and functioning of degraded riparian vegetation. Ecological restoration of degraded RFs, in which biodiversity is greatly decreased, has the stated goal to enhance the value of the ecosystem, moderate degrading influences and regain pre-disturbance characteristics (NRC 2000). This approach may include planting native trees, facilitating development of certain species, preserving and rehabilitating threatened and endemic species. The problem of isolated 144 Chapter 12: General conclusion populations could be partially solved by reforesting corridors between riparian forests patches and belts along watercourses. Qualitative and quantitative knowledge of hydrology and ecology (including the range of natural variability, disturbance regimes, soils and landforms, and vegetation) are therefore required for successful rehabilitation of RFs. 12.4. CHALLENGES TO RIPARIAN FORESTS ASSESSMENT IN BENIN As significant as climate changes are likely to be, land- and water-use changes have had and will continue to have the greatest effects on riparian vegetation. Therefore, more qualitative and quantitative assessments of the effects of land and water management on RFs have to be undertaken. The main variables to be considered should be related to physical, biological and anthropogenic factors. Research on RFs in Benin should document or clarify: - ecological conditions (e.g. seed bank, germination, gap dynamics, edge effects, patterns in the succession, and long-term survival of plant communities), geomorphic, as well as hydrologic issues related to riparian systems; - relationships between land forms, and RFs composition and structure dynamics; - the effects of hydrologic (i.e. ground water table, surface water) changes on the composition, physiognomy and productivity of RFs; - relationships between RFs diversity and structure, and waterways order and size; - the relationship between flooding, accumulation of organic debris, and severity of fire; - the effects of dissolved nutrients, foliage, or woody debris from RFs, as food resources for fishes, algae, and invertebrates; - the influence of interactions of abiotic factors on relative growth and flood survivorship rates of native RF species; - the rate and pattern of tree mortality in RFs, and the role of subsequent delivery of wood to waterways; - the influence of fire on regeneration, recruitment patterns of plant species and RFs diversity; - the potential role of such forest fragments as animal habitat or refuge, and movement corridors throughout the landscape; - the usefulness of direct gradient analysis based on environmental variables (e.g. soils profile, structure, texture, fertility and nutrient content, water and flood related variables, types and levels of disturbance) to explain RFs diversity and dynamics; - the genetic diversity within and among little known, valuable or rare plant populations; From the list of RFs species, a database of suitable riparian species needs to be established to provide information on the type of plants and their ecological requirements at various latitudes and river systems. Research should set up tolerances and thresholds at least for the dominant RF trees regarding ground and surface waters, as a basis for predicting and preventing adverse ecological effects from stream draining or severe drought. Likewise dispersal and colonisation strategies of seeds, as well as ability of riverbank to trap waterborne diaspores on riparian vegetation diversity should be investigated. Remote sensing and geographical information systems are known to be powerful tools for assessing spatial and temporal changes of natural resources in general, and riparian habitats in particular. Effective management strategies must be devised to preserve and restore riparian corridors (Brinson et al. 1981, Albernathy & Turner 1987, Gosselink & Lee 1989, Taylor et al. 1990), while relying on accurate understanding of the structure and dynamics of RF communities (Sagers & Lyon 1997). However, simply protecting RFs in a buffer zone may not be adequate to ensure their existence in the long term (Hibbs & Bower 2001). Instead the management of riparian forests must be a component (i.e. subset) of good watershed or landscape management. So awareness has to be raised to various stakeholders on riparian forests as unique physical and natural systems in their own right, and warranting special 145 Chapter 12: General conclusion management and protection. Integrated management of riparian forests that optimise their values as habitat for native plants and animals requires planning and acting, with all stakeholders, at both site-specific and watershed levels. Through improved and long-term scientific, educational and recreational programs, the ecological importance and intrinsic human values associated with riparian forests may be better balanced against the competing wants of unplanned urbanisation, short-term local inhabitants’ demands, and unknown motivations of other stakeholders. 146 SUMMARY A.K. Natta Summary SUMMARY The present research deals with the flora, phytosociology and ecology of riparian forests in Benin. In chapter 1 (General introduction), the research background, objectives and approach for riparian forests biodiversity assessment, and the organisation of the thesis are presented. Chapter 2 introduces the study area which covered about 70 % of Benin, from 7° 10’ to 12° 20’ N. Chapter 3 presents an overview of riparian forests biodiversity, their importance and the threats they face making them endangered ecosystems. A definition of riparian forests (or gallery forests) is given in the Benin context. The floristic characteristics of riparian forests in each phytogeographic district are presented. Issues related to legal protection and rehabilitation of the function and resources of riparian forests are documented: specifications and weaknesses of the forest law regarding riparian forests are presented; challenges for various stakeholders are discussed, and some improvements of the current forest law are proposed. Chapter 4 assesses plant species diversity, as well as species abundance models that best fit representative collections of plant species of riparian forests throughout the country. This study shows the richness and diversity of riparian forests in Benin, in comparison to other vegetation types in this country. They harbour about 1/3 of the estimated total number of plant species of the whole country in sample plots totalling 19 ha. This flora shares many features with riparian forests and dense forests worldwide: e.g. most abundant families, species richness/ha, trees species richness/ha, Shannon index, equitability index of Pielou, and species abundance models. Endemism is very low compared to that in rain forests, what is not surprising in the Dahomey Gap. The main conclusion is that relatively large numbers of species are still maintained in small forest fragments along waterways. These remnants with their specific plant species composition can be used for the restoration of degraded forest stands. Chapter 5 assesses the structure and ecological spectra of 19 ha of riparian forests through selected parameters (e.g. life form, geographic affinity, diameter class distribution, basal area, stem density, species dominance) that give a general picture of different vegetation types present. Figures obtained for these parameters show that riparian forests in Benin are on the one hand similar to many riparian forests in West Africa as well as in South and Central America, and on the other hand to many tropical upland forests. A brief description of the process of riparian forests degradation is also presented. Chapter 6 deals with the phytosociological assessment of representative relevés of riparian forests of Benin. Floristic ordination (DCA analysis) and classification (TWINSPAN) were derived from a comprehensive floristic inventory of a data set of 818 plant species and 180 relevés. This yielded 12 plant communities or associations, most of which had not yet been formally described: 1 - Community of Isolona thonneri and Callichilia barteri (10 relevés) along streams. This community occurs at the lowest parts of the gallery forest with frequent inundation in the centre of Pénéssoulou protected forest. 2 - Community of Motandra guineensis and Pararistolochia goldieana (24 relevés) along streams at the East and West parts of Pénéssoulou reserve forest. This community is mainly present on drained sites (i.e. seldom inundated). 3 - Community of Chrysobalanus icaco subsp. atacoriensis and Pentadesma butyracea (22 relevés) along streams at hill feet in the Atacora mountain chain. 4 - Community of Alchornea cordifolia and Ficus trichopoda (9 relevés) along streams on regularly inundated plateaus all over the country. 149 Summary 5 - Community of Berlinia grandiflora and Khaya senegalensis (8 relevés) along streams on drained plateaus (i.e. seldom inundated), mainly in the Sudanian region of the country. 6 - Community of Raphia sudanica and Oxytenanthera abyssinica (8 relevés) along streams on drained plateaus, mainly in the Sudanian region. 7 - Community of Cynometra megalophylla and Parinari congensis (31 relevés) along the Ouémé river in the Guinean region of Southern Benin. 8 - Community of Capparis thonningii and Crateva adansonii (30 relevés) along the Ouémé river in the Sudano-Guinean region of Central Benin. 9 - Community of Lepisanthes senegalensis and Drypetes floribunda (17 relevés) along the Ouémé river in the Sudano-Guinean region of Central Benin. 10 - Community of Uapaca heudelotii and Irvingia smithii (8 relevés) along the Sota river in the Sudanian region of North East Benin. 11 - Community of Garcinia livingstonei and Combretum acutum (12 relevés) along the Pendjari river in the Sudanian region of North West Benin. 12 - Community of Mimosa pigra and Ficus asperifolia (20 relevés) widely distributed on sandy banks along rivers. Ordination proved invaluable in the exploration of environmental characteristics of the phytosociological groups. The environmental factors (waterways, relief, topography, latitude and longitude) helped in the grouping of floristic relevés in the above mentioned 12 plant communities. The distinguished plant communities were compared with syntaxonomic data in literature. Riparian forests in Benin belong to the Mitragynetea Schmitz 1963, which is the phytosociological class of hygrophile fresh water forests of tropical Africa. Based on similarities of ecological conditions and floristic composition, the 12 plant communities can be classified into 3 orders that are Alchornetalia cordifoliae Lebrun 1947, LanneoPseudospondietalia Lebrun & Gilbert 1954 and Pterygotetalia Lebrun & Gilbert 1954. Chapter 7 presents the spatial distribution and ecological factors determining the occurrence of Pentadesma butyracea (Clusiaceae), a rain forest and multipurpose species found in Benin only along certain streams. Among the 224 tree species found along waterways, Pentadesma is one of the least known, yet of great ecological and economic importance. Field survey reveals the presence of this rain forest species in four non-contiguous remnant riparian areas, some located far from its optimal ecological range. If urgent actions are not taken to protect the remaining fragmented and dispersed riparian habitats, current human-induced disturbance could result in the disappearance of this species in Benin. Chapter 8 deals with the variation of the floristic composition, structural parameters (e.g. abundance, average height, basal area, tree richness) and spatial distribution of tree species at river edges across riparian forests. Horizontal and vertical structures of tree species exhibit complex patterns at riverside. On the one hand, tree stems are characterised by an uneven distribution across riparian forests, on the other hand height and basal area variations at riverside do not show any easily interpretable patterns. The numerical analysis confirms a gradual variation in the floristic composition across riparian forests and neighbouring plant communities. These results suggest a partitioning of riparian forests in three habitats (i.e. river front, middle and riparian forest edge). An implication for diversity assessment is that plot size, shape and layout in the terrain should take into account the river front, the middle and the edge of riparian forest. Due to the non-coverage of the whole riparian forest width and unequal chance of species and stems to be sampled, circular and square plots are not suitable for structural parameters and phytodiversity assessment in riparian forests. Instead rectangular plots with varying length and width, and covering the whole cross section of riparian forest are the most suitable sampling units under the study area conditions, and probably for savanna regions too. The present study also provides scientific guidelines for an 150 Summary improvement of the forest law regarding the distance to be protected at riverside, and suggests 100 m instead of 25 m. In chapter 9 the floristic composition, species richness and structure of two riparian ecosystems in West Africa (the Comoé in Ivory Coast and the Ouémé in Benin), are compared. The overall physiognomy of the two gallery forest sites seems similar and they share the most prominent families. However, there are marked differences in terms of canopy density and height, herb layer density, number of individuals, tree richness and diversity (H’), and species composition. The phenomenon of single species dominance at both sites is documented from Cynometra megalophylla, an evergreen tree species, which was time and again the most frequent and dominant tree at both riversides and in the middle of the gallery forests. Only detailed comparison shows the difference and complexity of ecological processes between and within gallery forests sites. The research carried out in chapter 10 facilitates the choice between several sampling designs for the estimation of a population parameter for endangered species. This study was carried out in the Pénéssoulou forest, in Central Benin. Stratified random sampling provided the lowest variance, coefficient of variation, standard error and sampling error. This method was taken as the most precise and reliable design over simple random and systematic samplings for the density estimation of Khaya senegalensis and K. grandifoliola trees. Results have confirmed empirical knowledge about the ecology of Khaya species and shown that the selection of the most precise sampling design, with regards to estimating a given parameter, can eventually be useful for the sustainable management of forest trees in the study area. A reliable density estimate for Khaya species within the given vegetation types facilitates the selection of areas to be protected and sustainably exploited. Chapter 11 presents a general discussion on issues discussed in this thesis. Sustainable rehabilitation and restoration of riparian forests biodiversity in Benin are discussed in the general conclusion (chapter 12). This study has provided detailed site-specific data on plant species that can serve for further scientific research, as well as for conservation management and planning. It fills a gap of knowledge on the flora of Benin, and can contribute to better land-use planning and conservation of riparian forests. 151 RESUME A.K. Natta Résumés RESUME TITRE DE LA THESE ETUDE ECOLOGIQUE DES FORETS GALERIES DU BENIN: PHYTODIVERSITE, PHYTOSOCIOLOGIE ET DISTRIBUTION DES ESPECES LIGNEUSES. Le chapitre 1 (Introduction générale) présente les fondements, objectifs et l’approche suivie pour l’étude de la biodiversité des forêts galeries (i.e. riparian forests en Anglais) du Bénin. Le chapitre 2 (Milieu d’étude) introduit la zone d’étude qui couvre environ 70 % de la surface totale du Bénin de 7° 10’ à 12° 20’ N. Tous les districts phytogéographiques de Bénin reconnus par Adjanohoun et al. (1989), et Houinato et al. (2000) y ont été sillonnés et des sites représentatifs installés. Le chapitre 3 (Forêts galeries du Bénin: un écosystème unique mais en voie de disparition) fait le point des connaissances actuelles sur la diversité de la flore, l’importance écologique, économique et socioculturelle des forêts galeries du Bénin. Une définition des forêts galeries dans le contexte du Bénin est proposée. Les causes de la dégradation actuelle des forêts galeries sont présentées. Un accent particulier est mis sur la protection légale de ces bandes boisées longeant les cours d’eau et les défis actuels pour la protection et l’aménagement durable de leur biodiversité. Le chapitre 4 (Diversité floristique des forêts galeries du Bénin) évalue la flore et la phytodiversité de l’ensemble des 19 ha de forêts galeries étudiées de 1999 à 2002 au Bénin. Au total quelque 1003 espèces ont été recensées, ce qui représente environ le 1/3 du nombre total d’espèces estimé de la flore du Bénin. Ces espèces se regroupent en 120 familles et 513 genres. 224 espèces ligneuses ont été recensées. Les (sous) familles les plus représentées sont les Papilionioideae, Poaceae, Rubiaceae et Euphobiaceae. Les forêts galeries sont caractérisées par une flore typique adaptée aux inondations récurrentes et par une multitude d’espèces typiques de forêt et savane de plateau. Selon les sites, l’indice de Shannon varie de 2,4 à 5,8 bits et l’Equitabilité de Pielou de 0,51 à 0,86. La richesse spécifique varie de 120 à 358 plantes par ha et celle des espèces ligneuses (dbh ≥ 10 cm) de 27 à 99 par ha. Cette étude montre que dans la majorité des paysages savanicoles, les forêts galeries du Bénin constituent des îlots de grande biodiversité qu’il faudrait restaurer et conserver durablement. Le chapitre 5 (Structure et spectres écologiques des forêts galeries du Bénin) traite des principales caractéristiques structurales de 19 ha de forêts galeries du Bénin, à travers : les types biologiques, phytogéographiques, la distribution par classe de diamètre, la surface terrière et la densité des ligneux. La physionomie des forêts galeries est très variable en terme de stratification verticale et hauteur de la canopée, cependant le sous-bois est généralement dense. Les forêts galeries du Bénin sont similaires à de nombreuses forêts galeries et forêts semi-décidues d’Afrique de l’Ouest, concernant: les types phytogéographiques (forte présence de espèces Guinéo-Congolaises), les types biologiques (forte abondance des thérophytes et des micro et méso phanérophytes, faible abondance des méga-phanérophytes et lianes ligneuses), la distribution par classe de diamètre (J renversé), la surface terrière (23 à 59 m3/ha) et la densité des ligneux (253 à 785 tiges/ha). La dominance d’une ou de plusieurs espèces en terme d’abondance et de surface terrière est une caractéristique 155 Résumés essentielle des forêts galeries du Bénin. Malheureusement les sites les plus denses et ayant une structure assez développée sont assez rares de nos jours. Le chapitre 6 (Phytosociologie des forêts galeries du Bénin) traite en large de la classification et de l’ordination de relevés phytosociologiques les moins dégradés et les plus représentatifs de l’ensemble du milieu d’étude. L’analyse multivariée de 180 relevés et 818 espèces a permis d’individualiser 12 associations qui s’ordonnent suivant cinq facteurs ou gradients: l’importance du cours d’eau (rivière ou fleuve), le relief, la topographie, la latitude et la longitude. Il s’agit des associations végétales à : 1 - Isolona thonneri et Callichilia barteri (10 relevés), le long de rivières dans le centre de la forêt classée de Pénéssoulou, sur sol très inondable. 2 - Motandra guineensis et Pararistolochia goldieana (24 relevés), le long des rivières à l’Est et à l’Ouest de la forêt classée de Pénéssoulou, sur sol rarement inondé (i.e. bon ressuyage du sol). 3 - Chrysobalanus atacoriensis et Pentadesma butyracea (22 relevés), le long des rivières de bas de collines. 4 - Alchornea cordifolia et Ficus trichopoda (9 relevés), le long des rivières de plateaux sur substrats engorgés d’eau pendant la saison de pluie (mauvais ressuyage du sol). 5 - Berlinia grandiflora et Khaya senegalensis (8 relevés), le long des rivières de plateaux à bon ressuyage du sol en milieu Soudanien. 6 - Raphia sudanica et Oxytenanthera abyssinica (8 relevés), le long des rivières de plateaux à bon ressuyage du sol en milieu Soudanien. 7 - Cynometra megalophylla et Parinari congensis (31 relevés), le long du fleuve Ouémé en région Guinéenne. 8 - Capparis thonningii et Crateva adansonii (30 relevés), le long du fleuve Ouémé en région Soudano-Guinéenne. 9 - Lepisanthes senegalensis et Drypetes floribunda (17 relevés), le long du fleuve Ouémé en région Soudano-Guinéenne. 10 - Uapaca heudelotii et Irvingia smithii (8 relevés), le long de la Sota dans le Nord Est du Bénin. 11 - Garcinia livingstonei et Combretum acutum (12 relevés), le long de la Pendjari dans le Nord Ouest du Bénin. 12 - Mimosa pigra et Ficus asperifolia (20 relevés), sur les bancs de sable proche du lit des cours d’eaux. Une classification syntaxonomique est aussi présentée. Les forêts galeries du Bénin sont rangées dans la classe des Mitragynetea Schmitz 1963 qui regroupe toutes les forêts édaphiques hygrophiles d’eau douce d’Afrique tropicale. Les associations individualisées se regroupent en trois ordres (Alchornetalia cordifoliae Lebrun 1947, LanneoPseudospondietalia Lebrun & Gilbert 1954, et Pterygotetalia Lebrun & Gilbert 1954). Trois associations sur les douze ont pu être classifiées dans des Alliances préexistantes. L’approche phytosociologique de Braun-Blanquet couplée aux analyses multivariées se sont avérées concluantes quant à l’individualisation des groupements végétaux de forêts galeries et la détection de gradients / facteurs majeurs de cette typologie. Le chapitre 7 (Distribution spatiale et facteurs écologiques de la présence de Pentadesma butyracea (Clusiaceae) au Bénin), traite de la répartition spatiale d’une espèce de forêt sempervirente qui se retrouvent curieusement au Bénin uniquement dans les forêts galeries en milieux Soudano-Guinéen (région de Bassila et près du village Agbassa, commune de Tchaourou) et Soudanien (région de Perma à Tandafa et près du village de Gbèssè, commune de Ségbana). Pentadesma est une espèce relativement peu connue du monde scientifique 156 Résumés béninois, pourtant elle est à but multiple, et pourrait jouer un rôle de premier choix dans la diversification des filières agro-forestières. La présence de cette espèce de forêt dense sempervirente dans les forêts galeries du Bénin dans la zone Soudano-Guinéenne, et aussi loin qu’à 11° N, nous a permis de valider quelques hypothèses concernant l’importance écologique de ces formations édaphiques d’eau douce. Le chapitre 8 (Variation de la structure et de la composition floristique dans la direction perpendiculaire aux fleuves) fait usage de l’analyse multivariée (DCA et TWINSPAN) et des courbes de distribution des ligneux pour tester l’existence d’espèces typiques du front d’eau, du milieu et de lisière des galeries. L’abondance des ligneux décroît constamment du lit du cours d’eau vers la lisière de la galerie. Cependant la hauteur des arbres et la surface terrière varient très peu lorsqu’on s’éloigne du cours d’eau. Quant à la composition floristique, elle varie graduellement dans la galerie perpendiculairement au cours d’eau vers les groupements végétaux adjacents aux galeries. Ces résultats nous ont permis de définir des espèces du front des fleuves, du milieu des galeries, de lisière de galeries, et des espèces typiques des groupements végétaux contigus aux galeries. Les espèces typiques de galeries constamment au contact de l’eau des fleuves sont: Syzygium guineense, Pterocarpus santalinoides, Parinari congensis, Cola laurifolia et Napoleonaea vogelii; Les espèces constamment au centre des galeries sont: Cynometra megalophylla, Drypetes floribunda et Manilkara multinervis; Les espèces de bordure de galerie ou dans l’écotone avec les groupements végétaux adjacents aux galeries sont: Dialium guineense, Diospyros mespiliformis, Elaeis guineensis, Cola gigantea et Ceiba pentandra. Cette recherche confirme le savoir empirique sur l’écologie des arbres en bordure des cours d’eau, mais montre que, bien que linéaires et fragmentées, les galeries ne peuvent être considérées comme uniformes sur les plans floristique et structural dans la direction perpendiculaire aux cours d’eaux. Une conséquence de ces résultats est que les placeaux rectangulaires avec des longueurs et largueurs variables et qui couvrent toute la section perpendiculaire des galeries sont les mieux adaptés en milieux Soudanien et SoudanoGuinéen. Aussi la distance légale minimale à être protégée de part et d’autre des cours d’eau (i.e. 25 m) apparaît-elle trop faible. Les résultats obtenus suggèrent de porter cette distance à 100 m, cela permettra aux galeries forestières de mieux assurer leurs importantes fonctions. Le chapitre 9 (Comparaison de l’interface forêt galerie/ savanes le long de l’Ouémé (Bénin) et de la Comoé (Côte d’Ivoire)), traite des relations forêt galerie / savane dans deux systèmes riverains en milieu Soudano-Guinéen. Les deux sites comparés sont approximativement à la même latitude, mais distants d’au moins 600 km à vol d’oiseau au Bénin (Idadjo-Bétérou) et dans le Sud Ouest du Parc National de la Comoé en Côte d’Ivoire. On constate une similarité concernant les familles ayant une grande richesse spécifique (i.e. Leguminosae, Rubiaceae et Sapotaceae) mais aussi des différences notables (e.g. densité des espèces de la canopée, hauteur moyenne, densité du sous-bois, richesse spécifique et composition floristique). Cynometra megalophylla est dans les deux sites l’espèce la plus fréquente et dominante. C’est un exemple typique de dominance mono-spécifique dans les forêts galeries en Afrique de l’Ouest. Une analyse plus fine des interactions entre les trois portions des galeries (front du cours d’eau, milieu de la galerie et l’écotone) montre la variabilité et la complexité des phénomènes écologiques entre sites et au sein d’un même site. Le chapitre 10 (Comparaison de trois méthodes d’échantillonnage pour l’estimation de la densité de Khaya senegalensis et K. grandifoliola dans la forêt classée de Pénéssoulou) 157 Résumés évalue les performances relatives des systèmes d’échantillonnage aléatoire, systématique et stratifié pour l’obtention des valeurs crédibles et précises (en terme statistique) des densités des deux espèces de Khaya présentent au Bénin. Il est généralement admis que l’obtention des valeurs précises pour des paramètres d’espèces en voie de disparition est une nécessité pour tout aménagement durable. Ainsi nous avons testé dans la partie la plus dense et la moins perturbée de la forêt classée de Pénéssoulou, les trois méthodes d’échantillonnage conventionnelles. L’échantillonnage stratifié, avec allocation proportionnelle de placeaux, est le plus fiable et précis, puisqu’il donne les valeurs les plus faibles de la variance, du ratio de variance, de l’erreur standard, et du coefficient de variation. Cette approche pourrait être testée pour d’autres espèces de valeur dans la même zone d’étude. Cela permettrait d’avoir des résultats consistants pour plusieurs espèces dans la zone d’étude. Une discussion générale de tous les aspects abordés dans la présente thèse est présentée au chapitre 11. La conclusion générale (Chapitre 12) traite des stratégies les plus adaptées et les nouveaux défis pour une meilleure conservation de la biodiversité des forêts galeries au Bénin. 158 SAMENVATTING A.K. Natta Samenvatting SAMENVATTING Dit onderzoek behandelt de flora, de vegetatiekundige aspecten en de ecologie van oeverbossen in Benin. In Hoofdstuk 1 (Algemene Inleiding) worden de achtergronden van het onderzoek, de doelstellingen en benaderingen om de biodiversiteit van oeverbossen te beoordelen, evenals de indeling van dit proefschrift beschreven. Hoofdstuk 2 beschrijft het gebied in Benin waar het onderzoek plaatsvond. Het studiegebied beslaat ca. 70% van het land en is gelegen tussen 7010 en 120 20’ Noorderbreedte. Hoofdstuk 3 geeft een overzicht van de biodiversiteit in oeverbossen, het belang daarvan en de aantastingen waaraan ze blootstaan. Dit ecosysteem moet in Benin als bedreigd worden beschouwd. Een definitie van oever- (of galerij-)bossen is gegeven in de Beninese context. De floristische eigenschappen van oeverbossen worden vermeld in elk fytogeografisch district. Vraagstukken i.v.m. de wettige bescherming en herstel van de functie en bronnen van het oeverbos komen aan de orde; uitdagingen voor diverse betrokkenen worden besproken, en verbeteringen voor de huidige Boswet worden voorgesteld. Hoofdstuk 4 beschrijft de soortenrijkdom, en de op de representatieve verzamelingen uit oeverbossen uit de meeste delen van het land best passende abundantie-modellen. Dit onderzoek laat zien hoe rijk en divers oeverbossen zijn in Benin, in vergelijking met andersoortige bossen in het land. Oeverbossen herbergen ongeveer 1/3 van het aantal soorten van de gehele flora van Benin, zoals aangetoond in vele proefvlakken die bij elkaar 19 ha besloegen. De oeverbosflora heeft veel gemeen met andere oeverbossen en dichte laaglandregenbossen in de wereld: bijv. de soortenrijkste plantenfamilies, soortenrijkdom per ha, boomsoortenrijkdom per ha, Shannon index, de Equitability Index van Pielou, en abundantie modellen. In de oeverbossen is het endemisme veel lager dan in dichtere regenwouden, dat is echter niet verrassend in de Dahomey Gap. De belangrijkste conclusie is dat relatief veel soorten zich nog steeds handhaven in kleine bosfragmenten langs waterwegen. Deze bosfragmenten en bepaalde soorten kunnen worden gebruikt als uitgangsmateriaal voor het herstel van gedegenereerde vegetaties. Hoofdstuk 5 bekijkt de structuur en ecologische spectra van 19 ha oeverbos door middel van geselecteerde parameters (levensvorm, geografische verbanden, verdeling van diameter klassen, basis oppervlak, dichtheid van stammen, dominantie van soorten) welke een overzicht leveren van vegetatietypes. Cijfers verkregen voor deze parameters bewijzen dat oeverbossen in Benin vergelijkbaar zijn met vele oeverbossen in West Afrika en Zuid- en Centraal Amerika, maar ook met vele tropische bossen op drogere gronden. Het degradatieproces van oeverbossen wordt kort beschreven. Hoofdstuk 6 behandelt de vegetatiekundige opnamen in representatieve proefvlakken in Beninese oeverbossen. Ordinatie (DCA analyse) en classificatie (TWINSPAN) werden toegepast op 180 proefvlakken en 818 plantensoorten. Dit leverde liefst 12 plantengemeenschappen op, waarvan de meeste nog niet eerder formeel zijn beschreven. 1 - Isolona thonneri en Callichilia barteri gemeenschap (10 proefvlakken) langs stroompjes in het midden van het beschermde bos van Pénéssoulou. Deze associatie vindt men in de laagstgelegen delen van het bos welke vaak overstroomd worden. 2 - Motandra guineensis en Pararistolochia goldieana gemeenschap (24 proefvlakken) langs riviertjes in het oosten en het westen van het Pénéssoulou bos. Deze associatie komt voor op goed ontwaterde plekken welke zelden overstromen. 3 - Chrysobalanus icaco subsp. atacoriensis en Pentadesma butyracea gemeenschap (22 proefvlakken) langs stroompjes aan de voet van de Atacora heuvels. 4 - Alchornea cordifolia en Ficus trichopoda gemeenschap (9 proefvlakken) langs riviertjes op regelmatig overstroomde plateaus door het hele land. 161 Samenvatting 5 - Gemeenschap van Berlinia grandiflora en Khaya senegalensis (8 proefvlakken) langs stroompjes op goed ontwaterde plateaus, vooral in de Sudan zone van het land. 6 - Gemeenschap van Raphia sudanica en Oxytenanthera abyssinica (8 proefvlakken), ook langs stroompjes op goed gedraineerde plateaus in de Sudan zone. 7 - Gemeenschap van Cynometra megalophylla en Parinari congensis (31 proefvlakken) langs de Ouémé rivier in de Guinee-zone van Zuid Benin. 8 - Gemeenschap van Capparis thonningii en Crateva adansonii (30 proefvlakken) langs de Ouémé rivier in de Sudano-Guinee zone van midden-Benin. 9 - Gemeenschap van Lepisanthes senegalensis en Drypetes floribunda (17 proefvlakken) langs de Ouémé rivier in de Sudano-Guinee zone van midden-Benin. 10 - Gemeenschap van Uapaca heudelotii en Irvingia smithii (8 proefvlakken) langs de Sota rivier in het noordoosten van het land. 11 - Gemeenschap van Garcinia livingstonei en Combretum acutum (12 proefvlakken) langs de Pendjari rivier in het noordwesten van Benin. 12 - Mimosa pigra en Ficus asperifolia gemeenschap (20 proefvlakken) zeer wijdverspreid op zandbanken langs rivieren. Ordinatie blijkt waardevol om de eigenschappen van het milieu van de fytosociologische groepen te onderzoeken. De gevonden factoren (aard van watergangen, reliëf, topografie, lengte- en breedtegraad) hielpen bij het groeperen van de floristische opnamen in de bovengenoemde 12 gemeenschappen. Een syntaxonomische indeling van de onderscheiden plantengemeenschappen wordt gegeven. Oeverbossen in Benin behoren tot het Mitragynetea Schmitz 1963, de fytosociologische klasse van vochtminnende zoetwaterbossen van tropisch Afrika. Gebaseerd op overeenkomsten van ecologische gesteldheid en floristische samenstelling deelden wij de 12 plantengemeenschappen in 3 ordes in: Alchornetalia cordifoliae Lebrun 1947, Lanneo-Pseudospondietalia Lebrun & Gilbert 1954 en Pterygotetalia Lebrun & Gilbert 1954. Hoofdstuk 7 onderzoekt de ruimtelijke verspreding en de ecologische factoren die het voorkomen van Pentadesma butyracea (Clusiaceae) bepalen. Deze op diverse manieren bruikbare soort uit het regenbos komt in Benin alleen langs sommige riviertjes voor. Onder de 224 boomsoorten gevonden langs grote en kleine stromen, is Pentadesma een van de minst bekende, ofschoon deze van groot ecologisch en economisch belang is. Veldstudies toonden aan dat deze regenwoudsoort in vier disjuncte relictgebiedjes voorkomt, sommige ver verwijderd van de optimale ecologische gebieden. De huidige verstoring door de mens kan leiden tot de verdwijning van de soort in Benin tenzij onmiddellijk actie wordt ondernomen de laatste versnipperde habitats, de oeverbossen, te beschermen. Hoofdstuk 8 behandelt de variatie in de floristische samenstelling, structurele parameters (bijv. abundantie, gemiddelde hoogte, basisoppervlakte, rijkdom aan bomen) en ruimtelijke verdeling van boomsoorten in oeverbossen. Horizontale en verticale structuren laten complexe patronen zien. Enerzijds wordt het aantal stammen gekarakteriseerd door een ongelijke verdeling in het oeverbos dwars op de richting van de stroom, anderzijds zijn de variaties van boomhoogte en basisoppervlakten langs de stromen niet eenvoudig te interpreteren. De numerieke analyse bevestigt een geleidelijke variatie van de floristische samenstelling vanaf de rivier, in het midden van het oeverbos, tot bij de naburige vegetatie. De resultaten wijzen op een onderverdeling van het oeverbos in deze drie genoemde habitats langs de horizontale gradiënt. Bij onderzoek aan diversiteit van oeverbossen, de omvang, vorm en plaatsing van proefvlakken in het terrein moeten steeds de oevers, het middenstuk en de buitenrand in beschouwing dienen genomen te worden. Ronde en vierkantige proefvlakken zijn ongeschikt om structurele parameters en biodiversiteit te meten, omdat hiermee niet de gehele breedte van het oeverbos wordt bemonsterd. Rechthoekige proefvlakken van variabele lengte en breedte lijken beter geschikt te zijn voor het 162 Samenvatting studiegebied, en vermoedelijk ook voor savannegebieden. Deze studie levert ook wetenschappelijke onderbouwing ter verbetering van de Boswet wat betreft de te beschermen breedte langs de rivieren, en stelt 100 m voor in plaats van 25 m. In Hoofdstuk 9 worden de floristische samenstelling, soortenrijkdom en structuur van twee oeverbossystemen in West Afrika (de Comoé in Ivoorkust en de Ouémé in Benin) vergeleken. Hoewel de fysiognomie van de twee bossen overeen lijken te komen en dezelfde plantenfamilies er prominent in voorkomen, zijn er duidelijke verschillen in dichtheid en hoogte van het kronendak, dichtheid van de kruidlaag, aantal exemplaren, rijkdom aan bomen en boomsoorten (H’) en soortensamenstelling. Het verschijnsel dat één soort dominant voorkomt is gedocumenteerd aan de hand van Cynometra megalophylla, een altijdgroene boomsoort, welke steeds de meest algemene de dominante soort is, zowel langs de rivier als in het midden van het oeverbos, in beide landen. Alleen gedetailleerde vergelijkingen maken de verschillen en complexiteit van de ecologische processen tussen en binnen oeverbossen duidelijk. Het onderzoek zoals gerapporteerd in hoofdstuk 10 vergemakkelijkt een keuze tussen verschillende bemonsteringsmethoden om populatieparameters van bedreigde soorten in te schatten. In het bos van Pénéssoulou bleek stratified random sampling de laagste variantie, variatiecoëfficient, standaardafwijking en monsterfouten op te leveren. Deze methode werd dus gekozen als de meest nauwkeurige in vergelijking met simple random en systematic sampling, om de dichtheid te meten van Khaya senegalensis en K. grandifoliola bomen. De resultaten bevestigen de empirische kennis van de ecologie van Khaya soorten en lieten zien dat de keuze van de meest nauwkeurige monstermethode, om bepaalde parameters te schatten, uiteindelijk nuttig kan zijn bosbomen in het studiegebied duurzaam te beheren. Een betrouwbare schatting van Khaya in de betreffende vegetatietypes vergemakkelijkt de keuze van te beschermen of duurzaam te exploiteren arealen. Hoofdstuk 11 bevat een algemene bespreking over de problemen zoals gegeven in het proefschrift als geheel. Duurzaam herstel van oeverbosdiversiteit in Benin wordt gepresenteerd in de algemene conclusie (hoofdstuk 12). Dit onderzoek levert nauwkeurige terrein-specifieke gegevens die kunnen dienen als uitgangspunt voor verder wetenschappelijke studie, en ook voor beheer en planning van behoudsmaatregelen. Het draagt bij aan de leemten in de kennis van de flora van Benin, en wij verwachten dat het bijdraagt aan verbetering van het behoud van oeverbossen en een verhoging van de prioriteit daarvan in landgebruiksplanning. 163 REFERENCES A.K. Natta References REFERENCES Abbiw D.K. 1990. Useful plants of Ghana: West African uses of wild and cultivated plants. Intermediate Technology Publications and the Royal Botanic Gardens, Kew, London. 337p. Acharya B. 1999. Forest biodiversity assessment: A spatial analysis of tree species diversity in Nepal. PhD Thesis. ITC Forest Division, ITC Publication 72, Enschede, The Netherlands. 199 p. Acker A.S., Lienkaemper G., McKee W.A., Swanson F.J. & Miller S.D. 2003. Composition, complexity, and tree mortality in riparian forests in the central Western Cascades of Oregon. Forest Ecology and Management 173: 293-308. Adambiokou-Akakpo J. 2001. Gestion des terroirs et planification de l’utilisation des ressources naturelles dans la commune de Klouékanmè. Mémoire DEA, Option Gestion de l’environnement. FLASH-UAC, Abomey-Calavi, Bénin. 96p. Adjakidjè V. 1984. Contribution à l’étude botanique des savanes guinéennes en République Populaire du Bénin. Thèse de 3ème cycle, Université de Bordeaux III, Bordeaux, France. 245p. Adjanohoun E. 1965. Végétation des savanes et des rochers découverts en Côte d’Ivoire Centrale. Mémoire n° 7. ORSTOM, Paris. 178p. Adjanohoun E. 1966. Existence au Dahomey de savanes côtières analogues à celles de Côte d’Ivoire. Etudes dahoméennes 6-7: 48-71. Adjanohoun, E. & Aké Assi L. 1967. Inventaire floristique des forêts claires soudaniennes et subsoudaniennes en Côte d’Ivoire septentrionale. Annales de l’Université d’Abidjan 3: 89-141. Adjanohoun E. 1968. Le Dahomey. Acta Phytogeographica Suecica 54: 86-91. Adjanohoun E. J., Adjakidjè V., Ahyi M. R. A., Aké Assi L., Akoègninou A., Akpovo F., d’Almeida J., Gbaguidi N., Boukef K., Chadaré F., Cusset G., Dramane K., Eyme J., Gassita J-N., Goudoté E., Guinko S., Houngnon P., Issa Lo, Keita A., Kiniffo H. V., Koné Bamba D., Musampa Nseyya A., Saadou N., Sodogandji Th., de Souza S., Tchabi A., Zinsou Dossa C. and Zohoun Th. 1989. Contribution aux études ethnobotaniques et floristiques en République Populaire du Bénin. Médecine traditionnelle et pharmacopée. ACCT, Paris. 895p. Adomako D. 1977. Fatty acid composition and characteristics of Pentadesma butyracea fat extracted from Ghana seeds. Journal of the Science of Food and Agriculture 28 (4): 384 - 386. Agbani P.O. 2002. Etudes phytosociologiques des groupements forestiers par bandes longitudinales à grandes échelles: Cas du noyau central de la forêt dense semi-décidue de la Lama au Bénin. Mémoire DEA/FLASH/UAC, Abomey-Calavi, Bénin. 73 p + annexes. Agee J.K. 1988. Successional dynamics in forest riparian zones. In: Raedeke K. (ed.), Streamside management: Riparian, Wildlife and Forestry Interactions. Institute of Forest resources, Contribution 59, University of Washington, Seattle, WA. pp 31-44. Aide M.T. & Rivera E. 1998. Geographic patterns of genetic diversity in Poulsenia armata (Moraceae): implications for the theory of Pleistocene refugia and the importance of riparian forest. Journal of Biogeography 25: 695-705. Akoègninou A. 1984. Contribution à l’étude botanique des îlots de forêts denses humides semidécidues en République Populaire du Bénin. Thèse de 3ème cycle, Université de Bordeaux III, Bordeaux, France. 181p. Akoègninou A., Houndagba J.C., Tossou M.G., Essou J.P. & Akpagana K. 2001. La végétation d’une zone de transition entre la forêt dense humide semi-décidue et les savanes: la région de Bantè (Bénin, Afrique de l’Ouest). Journal Botanique de la Société Botanique de France 15: 99108. Albernathy Y. & Turner R.E. 1987. United States Forested wetlands: 1940-1980. BioScience 37: 721727. 167 References Ambus P. & Lowrance R. 1991. Comparison of denitrification in two riparian soils. Soil Science Society of America Journal 55: 994-997. Ambus P. & Christensen S. 1993. Denitrification variability and control in a riparian fen irrigated with agricultural dainage water. Soil Biology and Biochemistry 25: 915-923. Amétépé A.M. 1997. Forêts sacrées et conservation de la biodiversité au Bénin: Cas du département du Mono. Mémoire d’Ingénieur Agronome. FSA/UAC, Abomey-Calavi, Bénin. 172p. Anderson B.W. & Ohmart R.D. 1977. Vegetation structure and birduse in the lower Colorado River valley. Importance, preservation, and management of riparian habitat: a symposium, USDA Forest Service General Technical Report RM – 43, Rocky Mountain Forest and Range Experiment Station, Fort Collins, Colorado. pp 23-34. Andersson E., Nilsson C. & Johansson E. 2000. Plant dispersal in boreal rivers and its relation to the diversity of riparian flora. Journal of Biogeography 27: 1095-1106. Arouna O. 2002. L’exploitation des ressources biologiques et la dynamique de la forêt classée de l’Alibori supérieur (secteur de Bagou). Mémoire de Maîtrise en Géographie. FLASH-UNB, Abomey-Calavi, Bénin. 114p. Auble G.T., Scott M.L., Friedman J.M., Back J. & Lee V.J. 1997. Contraints on establishement of plains cottonwood in an urban riparian preserve. Wetlands 17: 138-148. Aubreville A. 1950. Flore forestière soudano-guinéenne AOF – Cameroun – AEF. Office de la Recherche Scientifique Outre-Mer. Société d’Editions Géographiques Maritimes et Coloniales, Paris. 523p. Aubreville A. 1959. La flore forestière de la Côte d’Ivoire. Tome 2, 2ème edition. Publication CTFT, Paris. 332 p. Baker W.L. 1989. Macro- and micro-scale influences on riparian vegetation in Western Colorado. Annals of the Association of American Geographers 79: 65-78. Baker W.L. 1990. Species richness of Colorado riparian vegetation. Journal of Vegetation Science 1: 119-124. Bamps P. 1971. Répartition de Pentadesma butyracea en Afrique. Bulletin du Jardin Botanique National de Belgique 41: 430. Barton D.R., Taylor W.D. & Biette R.M. 1985. Dimensions of riparian buffer strips required to maintain trout habitat in southern Ontario streams. North American Journal of Fisheries Management 5: 364-378. Baumer M. 1995. Arbres, arbustres et arbrisseaux nourriciers en Afrique Occidentale. ACCT, Paris. 260 p. Begon M., Harper J.L. & Townsend C.R. 1986. Ecology, populations and communities. Blackwell Scientific Publication, Oxford. 876p. Bélem O.M. & Guinko S. 1998. La biodiversité des galeries forestières du Burkina Faso: cas de la réserve de la biosphère de la mare aux hippopotames. In: Bâ A.T., Madsen J.E. & Sambou B. (eds.), Atelier sur flore, végétation et Biodiversité au Sahel. Aarhus University Press, Aarhus Denmark, AAU Reports 39: 75-86. Bell D.T. & Sipp S.K. 1975. The litter stratum in a streamside forest ecosystem. Oikos 26: 391-397. Bell D.T., Johnson F.L. & Gilmore A.R. 1978. Dynamics of litter fall, decomposition, and incorporation in the streamside forest ecosystems. Oikos 30: 76-82. Bendix J. 1994. Among-site variation in riparian vegetation in the Southern California Transverse Ranges. American Midland Naturalist 132: 136-151. Bennett A.F. 1999. Linkages in the landscape: the role of corridors and connectivity in wildlife conservation. IUCN, Gland, Switzerland and Cambridge, UK. 254p. Berhaut J. 1967. Flore du Sénégal, 2ème édition. Editions Clairafrique, Dakar, Sénégal. 485p. 168 References Berhaut J. 1979. Flore illustrée du Sénégal. Tome IV. Edition Clairafrique, Dakar, Sénégal. 329p. Bersier L-F. & Meyer D. 1994. Bird assemblages in mosaic forests: the relative importance of vegetation structure and floristic composition along the successional gradient. Acta Oecologica 15 (5): 561-576. Best L.B., Stauffer D.F. & Geier A.R. 1979. Evaluating the effects of habitat alterations on birds and small mammalls occupying riparian communities. In: Johnson R.R. & McCormick J.R. (eds.), Strategies for protection and management of floodplain wetlands and other riparian ecosystems. USDA Forest Service General Technical Report WO-12. Washington. D.C. pp 117-124. Billy R.E. & Ward J.W. 1989. Changes in characteristics and functions of woody debris with increasing size of streams in western Washington. Trans. Am. Fish. Soc. 118: 368-378. Bocard D. 1992. Partialling out the spatial component of ecological variation. Ecology 73 (3): 10451055. Bongers F., Popma J., Meave del Castillo J. & Carabias J. 1988. Structure and floristic composition of the lowland rain forest of Los Tuxtlas, Mexico. Vegetatio 74:55-80. Bonkoungou E.G. 1984. Inventaire et analyse biogégraphique de la flore des forêts galeries de la Volta Noire en Haute Volta. Not. Doc.Volt. 15: 64-73. Bossou B. 2001. Contribution à l’aménagement de la forêt classée de Dogo-Kétou : Structure et dynamique des différents groupements végétaux et périodicité de coupe. Mémoire de 3ème cycle DESS/FSA/UAC, Abomey-Calavi, Bénin. 141p + annexes. Bowden W.B., McDowell W.H., Asbury C.E. & Finley A.M. 1992. Riparian nitrogen dynamics in two geomorphological distinct tropical rain forest watersheds: nitrous oxide fluxes. Biogeochemistry 18: 77-99. Braatne J.H., Rood S.B. & Heilman P.E. 1996. Life history, ecology, and conservation of riparian cottonwoods in North America. In: Settler R.F., Bradshaw Jr. H.D., Heilman P.E. & Hinckley T.M. (eds.), Biology of Populus and its implications for Management and Conservation. NRC Press, Ottawa, Ontario. pp 57-85. Bragg D.C. & Kershner J.L. 1999. Coarse woody debris in riparian zones. Journal of Forestry 97(4): 30-35. Braun-Blanquet J. 1972. Plant sociology: The study of plant communities (Facsimile of the edition of 1932, translated by Fuller G.D. & Conard H.S.). Hafner Publishing Company, New York. 439 p. Brierley G.J. & Hickin E.J. 1985. The downstream gradation of particle sizes in the Squamish River, Bristish Columbia. Earth Surface Processes and Landforms 10: 597-606. Brinson M.M. 1990. Riparian forests. In: Lugo A.E., Brinson M. & Brown S. (eds.), Forested Wetlands. Ecosystems of the world 15. Elsevier, Amsterdam. pp 87-141. Brinson M.M., Swift B.L., Plantico R.C. & Barclay J.S. 1981. Riparian ecosystems: their ecology and status. U.S. Fish and Wildlife Service, Biological Services Program, FWS/OBS - 81-17, Washington D.C. Brinson M.M. & Verhoeven J. 1999. Riparian forests. In: Hunter M.L.Jr. (ed.), Maintaining Biodiversity in Forest Ecosystems. Cambridge University Press, Cambridge. pp 265-299. Brunel J. F., Hiepko P. & Scholz H. (eds.) 1984. Flore analytique du Togo - Phanerogames. Englera 4. GTZ, Eschborn, Berlin. 751p. Buckland S.T., Laake J.L., Anderson D.R. & Burnham K.P. 1993. Distance sampling. Estimating abundance of biological populations. First edition. Chapman and Hall, London. 441p. Bull E. & Skovlin J.M. 1982. Relationships between avifauna and streamside vegetation. 47th North American Wildlife Conservation 47: 496-505. 169 References Burbrink F.T., Phillips C.A. & Heske E.J. 1998. A riparian zone in Southern Illinois as a potential dispersal corridor for reptiles and amphibians. Biological Conservation 86: 107-115. Burke M.K., Lockaby B.G. & Conner W.H. 1999. Above ground production and nutrient circulation along a flooding gradient in a South Carolina Coastal Plain forest. Canadian Journal of Forest Resources 29: 1402-1418. Busch D.E. 1995. Effects of fire on south-western riparian plant community structure. Southwestern Naturalist 40: 259-267. Busch D.E. & Smith S.D. 1993. Effects of fire on water and salinity relations of riparian forests woody taxa. Oecologia 94: 186-194. Busch D.E. & Smith S.D. 1995. Mechanisms associated with decline of woody species in riparian ecosystems of the southwestern US. Ecological Monographs 65(3): 347-370. Carrol R., Augspurger C., Dobson A., Franklin J., Orians G., Reid W., Tracy R., Wilcove D. & Wilson D. 1996. Strengthening the use of science in achieving the goals of the endangered species act: An assessment by the ecological society of America. Ecological Applications 6 (1): 1-11. CENATEL (Centre National de Télédétection et de Surveillance du couvert forestier) 1995. Carte de végétation du Bénin. DFRN/MDR. Cotonou, Bénin. 15p. CENATEL 2000. The study on cartography, inventory and management of classified forest in Northern area in Benin. Final Report, Main text. Japan International Cooperation Agency (JICA), DFRN-MDR. Cotonou, Benin. 245p. CENATEL 2001. Résultats statistiques de la carte de végétation du Bénin. DFRN/MAEP. Cotonou, Bénin. Chestnut T.J. & McDowell W.H. 2000. C and N dynamics in the riparian and hyporheic zones of a tropical stream, Luquillo Mountains, Puerto Rico. Journal of the North American Benthological Society 19: 199-214. Cochran W.G. 1977. Sampling techniques. 3rd edition. John Wiley, New York. 428p. Colinvaux P. 1987. Amazon diversity in light of the paleoecological record. Quaternary Science Reviews 6: 93-114. Cooper J.R., Gilliam J.W., Daniels R.B., Ponarge W.P. 1987. Riparian areas as filters for agricultural sediments. Soil Science Society of America Journal 51: 416-420. Cordes L.D., Hughes F.M.R. & Getty M. 1997. Factors affecting the regeneration and distribution of riparian woodlands along a northern prairie river: the red Deer River Alberta, Canada. Journal of Biogeography 24: 675-695. Corthay R. 1996. Analyse floristique de la forêt sempervirente de Yapo (Côte d’Ivoire). Mémoire de l’Université de Genève, Suisse. 152p. Cowling R.M. & Holmes P.M. 1992. Flora and vegetation. In: Cowling R.M. (ed.), The Ecology of Fynbos: Nutrients, Fire and Diversity. Oxford University Press, Cape Town. Crawford C.S., Ellis L.M. & Molles Jr.M.C. 1996. The middle Rio Grande bosque: an endangered ecosystem. New Mexico Journal of Science 36: 276-299. Cubbedge A.W. & Nilon C.H. 1993. Adjacent land use effects on the flood plain forest bird community of Minnesota Valley National Wildlife Refuge. Natural Areas Journal 13: 220. Danvind M. & Nilsson C. 1997. Seed floating ability and distribution of alpine plants along a northern Swedish river. Journal of Vegetation Science 8: 271-276. Darveau M., Beauchesne P., Bélanger L., Huot J. & Larue P. 1995. Riparian forests strips as habitat for breedings birds in boreal forest. Journal of Wildlife Management 59: 67-78. Darveau M., Huot J. & Bélanger L. 1998. Riparian forests strips as habitat for snowshoe hare in boreal balsam fir forest. Canadian Journal of Forest Resources 28: 1494-1500. 170 References Darveau M., Labbé P., Beauchesne P., Bélanger L. & Huot J. 2001. The use of riparian forest strips by small mammals in a boreal balsam fir forest. Forest Ecology and Management 143: 95104. DeBano L.F. & Neary D.G. 1996. Effects of fire on riparian systems. In: Folliott P.F., DeBano L.F. & Baker M.B. (eds.), Effects of fire on Madrean Province ecosystems. A symposium proceedings. General Technical Report RMGTR – 289. US Department of Agriculture, Forest Service, Rocky Mountain Forest and Range Experiment Station, Fort Collins, CO. pp 69-76. Décamps H., Fortuné M., Gazelle F. & Pautou G. 1988. Historical influence of man on the riparian dynamics of a fluvial landscape. Landscape Ecology 1: 163-173. de Koning J. 1983. La forêt du Banco. Thèse de Doctorat. Section de Taxinomie et de Géographie botaniques, Université Agronomique de Wageningen, Pays-Bas. 921 p. de Lima M.G. & Gascon C. 1999. The conservation value of linear forest remnants in central Amazonia. Biological Conservation 91: 241-247. Dengueadhe K.Th.S. 1999. Etude de la végétation du Sud du Parc National de Taï: Identification des milieux. Résultats du Projet ‘Flore’, Deuxième partie. Tropenbos, Côte d’Ivoire. 62p. Depiere D. & Vivien J. 1992. Mammifères sauvages du Cameroun. Ministère du Développement Rural. Office National des Forêts, Fontainebleau, France. 249p. de Souza S. 1979. Contribution à l’étude biologique et écologique de quelques Chrysobalanacées des genres Chrysobalanus, Parinari et Maranthes au Bénin. Thèse de Doctorat de l’Université de Bordeaux III, Bordeaux, France. 317p. Devineau J.L. 1975. Etude quantitative des forêts galeries de Lamto (Moyenne Côte d’Ivoire). Thèse 3ème cycle. Université Pierre et Marie Curie, Paris VI, France. 190p. Devineau J.L. 1976. Principales caractéristiques physiognomiques et floristiques des formations forestières de Lamto (Moyenne Côte d’Ivoire). Annales de l’Université d’Abidjan série E 9: 274-303. Devineau J.L. 1984. Travaux des chercheurs de la station de Lamto: Structure et dynamique de quelques forêts tropophiles de l’Ouest Africain (Côte d’Ivoire). Programme MAB savanes no. 5. Université d’Abidjan, Station d’Ecologie tropicale de Lamto, Côte d’Ivoire. 294p. Dignan P. & Bren L. 2003. A study of logging on the understorey light environment in riparian buffer strips in a south-east Australian forest. Forest Ecology and Management 172: 161-172. Djego J.G.M. 2000. Impacts écologiques des plantations forestières sur la diversité biologique des écosystèmes côtiers du Bénin. Cas du périmètre de reboisement de Sèmè. Mémoire DEA/FLASH/UAC. Abomey-Calavi, Bénin. 129p. DNRE 1996. Code of Practice: Code of Forest practices for Timber Production. Department of Natural Resources and Environment, Victoria, Australia. 68p. Doyle A.T. 1990. Use of riparian and upland habitats by small mammals. Journal of mammalogy 71: 14-23. Dunn C.P. & Stearns F. 1987. Relationship of vegetation layers to soils in Wisconsin forested wetlands. American Midland Naturalist 118: 366-374. Dupond L.M. & Weinelt M. 1996. Vegetation history of the savanna corridor between the Guinean and the Congolian rain forest during the last 150,000 years. Vegetation History Archaeobotany 5: 273-292. Dupond L.M., Jahn S., Marret F. & Ning S. 2000. Vegetation change in equatorial West Africa: times-slices for the last 150 ka. Paleogeography, Paleoclimatology, Paleoecology 155: 95122. Eldin M. 1971. Le climat. In: Le milieu naturel de la Côte d’Ivoire. Mémoire ORSTOM 50: 73-108. 171 References Ellis L.M., Crawford C.S. & Molles Jr.M.C. 1998. Comparison of litter dynamics in native and exotic riparian vegetation along the Middle Rio Grande of Central New Mexico, USA. Journal of Arid Environments 38: 283-296. Ellis L.M., Molles Jr.M.C. & Crawford C.S. 1999. Influence of experimental flooding on litter dynamics in a Rio Grande riparian forest of Central New Mexico. Restoration Ecology 7: 113. Ellis L.M. 2001. Short-term response of woody plants to fire in a Rio Grande riparian forest of Central New Mexico, USA. Biological Conservation 97:159-170. Enright N.J., Miller B.P. & Crawford D. 1994. Environmental correlates of vegetation patterns and species richness in the northern Grampians, Victoria. Australian Journal of Ecology 19: 159168. Ericsson K.A. & Schimpf D.J. 1986. Woody riparian vegetation of a Lake superior tributary. Canadian Journal of Botany 64: 769-773. Ern H. 1979. Die Vegetation Togos. Gliederung, Gefährdung, Erhaltung. Willdenowia 9: 295-312. Ern H. 1988. Flora and vegetation of the Dahomey Gap. A contribution to the plant geography of west tropical Africa. Monographs Systematic Botany from the Missouri Botanical Garden 25: 517520. Essou J-P. 1991. Contribution à l’étude des pâturages et systèmes agro-pastoraux du Sud-Bénin. Thèse de Doctorat Université de Bordeaux III, Bordeaux, France. 370p. Ettema C.H., Lowrance R. & Coleman D.C. 1999. Riparian soil response to surface nitrogen input: the indicator potential of free-living soil nematode populations. Soil Biology and Biochemistry 31: 1625-1638. Everett R., Schellhaas R., Ohlson P., Spurbeck D. & Keenum D. 2003. Continuity in fire disturbance between riparian and adjacent sideslope Douglas-fir forests. Forest Ecology and Management 173: 293-308 Eyog-Matig O., Adjanohoun E., de Souza E. & Sinsin B. (eds.) 2001. Réseau “espèces ligneuses Médicinales”. Compte rendu de la première réunion du Réseau, 15-17 Décembre 1999. Station IITA Cotonou, Bénin. 131p. Eyog-Matig O., Gaoué O.G. & Obel-Lawson E. (eds.) 2002. Development of appropriate conservation strategies for African forest trees identified as priority species by SAFORGEN member countries. IPGRI Nairobi, Kenya. 83p. Faber-Langendoen D. & Gentry A.H. 1991. The structure and diversity of rain forests at Bajo Calima, Choco region, western Colombia. Biotropica 23(1): 2-11. FAO 1980. Cartographie du couvert végétal et étude de ses modifications. Projet pilote sur la surveillance de la couverture forestière tropicale. UN 32/6. 1102-75-005. Rome, Italie. 75p + 2 cartes. FAO 2000. Africa Forests Shrinking. Environment News Service. Rome, Italia. Farley G.H., Ellis L.M., Stuart J.N. & Scott Jr. 1994. Avian species richness in different-aged stands of riparian forest along the Middle Rio Grande, New Mexico. Conservation Biology 8(4): 1098-1108. Fetherston K.L., Naiman R.J. & Billy R.E. 1995. Large woody debris, physical process, and riparian forest development in montane river networks of the Pacific Northwest. Geomorphology 13: 133-144. FGU Kronberg 1979. Gegenwärter Status der Comoé – und Tai – Nationalparks sowie des AzagnyReservats und Vorschläge zu deren Erhaltung und Entwicklung zur Förderung des Tourismus. II. Comoé-Nationalpark. 1. Bestandsaufnahme der ökologischen Verhältnisse. GTZ, Eschborn. Germany. 172 References Flanagan D.C., Foster G.R., Neibling W.H. & Burt J.P. 1989. Simplified equations for filter strip design. Transactions of the ASAE 32(1): 2001-2007. Folke C., Holling C.S. & Perrings C. 1996. Biological diversity, ecosystems, and the human scale. Ecological Applications 6(4): 1018-1024. Forman R.T.T. 1997. Land Mosaics. Cambridge University Press. Cambridge, UK. Forman R.T.T. & Godron M. 1986. Landscape ecology. John Wiley & Sons. New York. 619p. Forniss M.J. 1996. Index for riparian bibliography. Interagency Watershed Analysis Centre, USDA – Forest Service. Freese F. 1990. Statistics for land managers. Paeony Press, Edinburg, Scotland. 176p. Ganglo J.C. 2000. Phytosociologie de la végétation naturelle de sous-bois, écologie et productivité des plantations de teck (Tectona grandis L.f.) du sud et du centre Bénin. Thèse Doctorat Université Libre de Bruxelles. Section Interfacultaire d’Agronomie, Belgique. 366p. Gartlan J.S., Newbery D.M., Thomas D.W. & Waterman P.G. 1986. The influence of topography and soil phosphorus on the vegetation of Korup forest Reserve, Cameroon. Vegetatio 65: 131-148. Gates J.E. & Giffen N.R. 1991. Neotropical migrant birds and edge effects at a forest-stream ecotone. Wilson Bulletin 103: 204-217. Geier A.R. & Best L.B. 1980. Habitat selection by small mammals of riparian communities: evaluating effects of habitat alterations. Journal of Wildlife Management 44: 16-24. Geraghty P. & Vollebergh P. 1993. Riparian zone management in Victoria: Policy and practice. In: Bunn S.E., Pusey B.J. & Price P. (eds.), Ecology and Management of Riparian Zones in Australia. LWRRDC Occasional Paper Series No: 05/93. Gentry A.H. 1982. Patterns of neotropical plant species diversity. In: Hecht M.K., Wallace B. & Prance G.T. (eds.), Evolutionary Biology. Plenum Press, New York. pp 123-135. Gentry A.H. 1988. Changes in plant community diversity and floristic composition on environmental and geographical gradients. Annals of the Missouri Botanical Garden 75:1-34. Gentry A.H. 1992. Tropical forest biodiversity: distributional patterns and their conservational significance. Oikos 63: 19-28. Gibbs P.E. & Leitão F.H.F. 1978. Floristic composition of an area of gallery forest near Mogi Guaçu, State of São Paulo, S.E. Brazil. Revista Brasileira de Botanica 1: 151-156. Giese L.A.B., Aust W.N., Kolka R.K. & Trettin C.C. 2003. Biomass and carbon pools of disturbed riparian forests. Forest Ecology and Management 6198: 1-16. Gilliam J.W. 1994. Riparian wetlands and water quality. Journal of Environment Quality 25: 896-900. Gnele J.E. 1999. Dynamique de l’environnement et utilisation des bas-fonds dans le Mono-Nord: Cas du secteur Dekpo-Lonkly. Mémoire de Maîtrise en Géographie FLASH-UNB. AbomeyCalavi, Bénin. 89p. Gosselink J.P. & Lee L.C. 1989. Cumulative impact assessment in bottomland hardwood forests. Wetlands 9: 89-174. Goudiaby A. 1998. Diversité floristique et conservation de la forêt galerie de la cascade de Dindelello, sud-est du Sénégal. In: Bâ A.T., Madsen J.E. & Sambou B. (eds.), Atelier sur flore, végétation et Biodiversité au Sahel. Aarhus University Press, Aarhus, Denmark. AAU Reports 39: 67-74. Green A.A. 1979a. Développement des Parcs Nationaux: La rivière Pendjari (Parc National de la Pendjari). PNUD/FAO/ 77/011 Document de travail No. 6. Cotonou, Bénin. 42p. Green A.A. 1979b. Développement des Parcs Nationaux: La végétation du Parc National de la Pendjari et des régions avoisinantes. PNUD/FAO/ 77/011 Document de travail No. 8. Cotonou, Bénin. 86p. 173 References Gregory S.V., Beschta R, Swanson F.J., Sedell J., Reeves G. & Everest F. 1990. Abundance of conifers in Oregon’s riparian forests. COPE Report, Vol 3, No. 2. College of Forestry, Oregon State University, Corvallis, OR. pp 5-6. Gregory S.V., Swanson F.J., MsKee W.A. & Cummins K.W. 1991. An ecosystem perspective of riparian zones: focus on links between land and water. BioScience 41: 540-551. Groffman P.M., Axelrod E.A., Lemunyon J.L. & Sullivan W.M. 1991. Denitrification in grass and forest vegetated filter strips. Journal of Environmental Quality 20: 671-674. Groffman P.M., McDowell W.H., Myers J.C. & Merriam J.L. 2001. Soil microbial biomass and activity in tropical riparian forests. Soil Biology & Biochemistry 33: 1339-1348. Grubb P., Lernould J.M. & Oates J.F. 1999. Validation of Cercopithecus erythrogaster pococki as the name for Nigerian white-throated guenon. Mammalia 63(3): 389-392. Guelly K.A. 1994. Les savanes des plateaux de la zone subhumide du Togo. Thèse de Doctorat de l’Université Paris 6, Spécialité Botanique tropicale. Paris, France. 163 p. Guillaumet J.L. 1967. Recherche sur la végétation et la flore de la région du Bas-Cavally (Côted’Ivoire). Mémoire ORSTOM 20: 247p. Guillaumet J.L. & Adjanohoun E. 1971. La végétation de la Côte d’Ivoire. In: Le milieu naturel de la Côte d’Ivoire. Mémoire ORSTOM 50:157-263. Guinko S. 1974. Contribution à l’étude écologique des savanes marécageuses du Bas-Dahomey. Thèse 3ème cycle. Université d’Abidjan, Côte d’Ivoire. 142p. Gunasekera S.P., Sivapalan K., Sultanbawa M. & Ollis W.D. 1977. Chemical investigation of ceylonese plants. Extractives of Pentadesma butyracea Guttiferae. J.C.S. Perkin 1: 15051511. Hall J.B. & Swaine M.D. 1976. Classification and ecology of closed-canopy forest in Ghana. Journal of Ecology 64: 913-951. Hall J.B. & Swaine M.D. 1981. Distribution and Ecology of vascular plants in a tropical rain forest, Forest Vegetation in Ghana. Geobotany 1, Dr. W. Junk, The Hague. 383 p. Hancock C.N., Ladd P.G. & Froend R.H. 1996. Biodiversity and management of riparian vegetation in Western Australia. Forest Ecology and Management 85: 239-250. Hanson J.S., Malanson G.P. & Armstrong P.M. 1990. Landscape fragmentation and dispersal in a model of riparian dynamics. Ecological Modelling 49 (3): 277-296. Harper D., Smith C., Barham P. & Howell R. 1994. The ecological basis for the management of the natural river environment. In: Harper D. & Ferguson A.J.D. (eds.), The ecological basis for river management. Wileym Chichester. pp 219-238. Hart T.B. 1990. Monospecific dominance in tropical rain forests. Tropical Ecology and Evolution 5: 6-12. Hartshorn G.S. 1988. Tropical and subtropical vegetation of meso-America. In: Barbour M.A. & Billings W.D. (eds.), North American terrestrial vegetation. Cambridge University Press, Cambridge. pp 366-390. Hawthorne W.D. 1996. Guide de terrain pour les arbres des forêts denses de la Côte d’Ivoire et pays limitrophes. Avec clés végétatifs sur plus que 650 espèces d’arbres, à partir de 5 cm de diamètre. Projet ECOSYN, Université de Wageningen, Pays-Bas. 276 p. Haycock N.E. & Pinay G. 1993. Groundwater nitrate dynamics in grass and poplar vegetated riparian buffer strips during the winter. Journal of Environmental Quality 22: 273-278. Haycock N.E., Pinay G. & Walter C. 1993. Nitrogen retention in river corridors: European perspective. Ambio 22: 340-346. Hefting M.M. & de Klein J.J.M. 1998. Nitrogen removal in buffer strips along a lowland stream in the Netherlands: a pilot study. Environmental Pollution 102(S1): 521-526. 174 References Hegarty E.E. & Caballé G. 1991. Distribution and abundance of vines in forests communities. In: Putz F.E. & Mooney H.A. (eds.), The biology of vines. Cambridge University Press, Cambridge. pp 238-250. Hehnke M. & Stone C. 1979. Value of riparian vegetation to avian populations along the Sacramento River system. In: Johnson R.R. & McCormick J.R. (eds.), Strategies for protection and management of floodplain wetlands and other riparian ecosystems. USDA Forest Service General Technical Report WO-12, Washington D.C. pp 228-235. Hermans M. 2002. Primary Producers at War with Parasites: A survey of medicinal plants used to treat malaria in West Africa with emphasis on Benin. MSc thesis, Laboratory for Plant Taxonomy. Wageningen University, The Netherlands. 60p. Hibbs D.E. & Bower A.L. 2001. Riparian forests in the Oregon Coast Range. Forest Ecology and Management 154: 201-213. Hill M.O. & Gauch Jr.H.G. 1980. Detrended Correspondance Analysis: An improved ordination technique. Vegetatio 42: 47-58. Hill M.O. 1979. TWINSPAN- a FORTAN Program for arranging Multivariate Data in an Ordered Two Way Table by Classification of the individuals and the Attributes. Cornell University, Department of Ecology and Systematics, Ithaca. New York. Hladik C. 1982. Dynamique d’une forêt équatoriale africaine: mesures en temps réel et comparaison du potentiel de croissance des différentes espèces. Acta Oecologica, Oecologia Generalis 3: 373-392. Hommel P.W.F.M. 1990. A phytosociological study of a forest area in the humid tropics (Ujung Kulon, West Java, Indonesia). Vegetatio 89: 39-54. Houinato M., Delvaux C. & Pauwels L. 2000. Les Eragrostis (Poaceae) du Bénin. Belgian Journal of Botany 133(1-2): 21-35. Houinato M.R.B. 2001. Phytosociologie, écologie, production et capacité de charge des formations végétales pâturées dans la région des Monts Kouffè, Bénin. Thèse de doctorat de l’Université Libre de Bruxelles, Belgique. 255p. Houndagba C.J., Tenté A.B.H. & Guédou G.R. (in press). Dynamique des forêts classées dans le cours moyen de l’Ouémé (Kétou, Dogo et Ouémé-Boukou) au Bénin. In : Actes du Premier séminaire régional sur l’Aménagement et la Gestion des Aires Protégées d’Afrique de l’Ouest. Parakou, Bénin, 14-19 Avril 2003. Houngbedji A. 1997. Etude phytotechnique, écologique et des technologies endogènes de transformation du Pentadesma butyracea, espèce des galeries forestières de la région de Bassila. Mémoire du DEAT – LTAM, Sékou, Bénin. 59 p. Hountondji Y.C.H. 1998. Contribution à l’étude des possibilités d’aménagement de la forêt de Niaouli au Sud-Bénin. Mémoire de fin d’étude, CPU/UAC, Abomey-Calavi, Bénin. 114p. Hubbell S.P. & Foster R.B. 1986. Commonness and rarity in a neotropical forests: implications for tropical tree conservation. In: Soulé M.J. (ed.), Conservation Biology: the Science of Scarcity and Diversity. Sinauer, Massachusetts. pp 205-231. Hughes F.M.R. 1990. The influence of flooding regimes on forest distribution and composition in the Tana River floodplain, Kenya. Journal of Applied Ecology 27: 475-491. Hunter Jr.M.L. 1990. Wildlife, Forest and Forestry: Principles of Managing Forests for biological diversity. Prentice-Hall, Englewood Cliffs, NJ. pp 148-149. Hupp C.R. & Osterkamp W.R. 1996. Riparian vegetation and fluvial geomorphic processes. Geomorphology 14: 277-295. Huston M.A. 1994. Biological diversity: The coexistence of species on changing landscapes. Cambridge University Press, Cambridge. 681p. 175 References Hutcheson K. 1970. A test for comparing based on the Shannon formula. Journal of Theoretical Biology 29: 151-4. INSAE (Institut National de la Statistique et de l’Analyse Economique) 1979. Recensement Général de la Population et de l’Habitat (RGPH). Tableaux statistiques, Volume National (Tome 1: mars 1979). Ministère du Plan, Cotonou, Bénin. INSAE 2002. Recensement Général de la Population et de l’Habitat (RGPH). Résultats provisoires. Ministère du Plan, Cotonou, Bénin. 4 p. Jansson R., Nilsson C., Dynesius M, & Andersson E. 2000a. Effects of river regulation on rivermargin vegetation: a comparison of eight boreal rivers. Ecological Applications 10: 203-224. Jansson R., Nilsson C. & Renöfält B. 2000b. Fragmentation of riparian floras in rivers with multiple dams. Ecology 81: 899-903. Johansson M.E., Nilsson C. & Nilsson E. 1996. Do rivers function as corridors for plant dispersal? Journal of Vegetation Science 7: 593-598. Jenik J. 1994. The Dahomey-Gap: an important issue in African phytogeography. Mémoire de la Société Biogéographique (3ème série) IV: 125-133. Johnson R.R., Haight L.T. & Simpson J.M. 1977. Endangered species versus endangered habitats: a concept. Importance, preservation, and management of riparian habitat: a symposium. USDA Forest Service General Technical Report RM – 43, Fort Collins, Colorado. pp 68-79. Jones A.P.D. 1945. Notes on terms for use in vegetation description in southern Nigeria. Farm and Forest 6: 130-136. Jones E.W. 1955. Ecological studies on the rain forest of southern Nigeria. IV. The plateau forest of the Okumu Forest Reserve. Part I. The environment, the vegetation types of the forest, and the distribution of species. Journal of Ecology 43: 564-594. Jongkind C.C.H. 1996. Environmental Impact Assessment of Black Volta River in south west Ghana. Report for the ODA - Ghana. (Unpublished). Kadio A. 1999. Constitution d’un réseau de “série de protection” dans les forêts classées gérées par la Société de Développement des Forêts (SODEFOR): une contribution à la sauvegarde de la diversité biologique. In: ECOSYN (ed.), Guide, Atlas et Ecologie des arbres et lianes de l’Afrique de l’Ouest, de la Guinée au Ghana, pour soutenir la gestion écologique des forêts denses tropicales humides. Séminaire d’Abidjan 13-14 décembre 1999, Côte d’Ivoire. pp 3235. Kalanda K. 1981. La végétation du parc de la N’sele (Zaire). Bulletin du Jardin Botanique National de Belgique 51 (3-4): 347-357. Kalliola R. & Puhakka M. 1988. River dynamics and vegetation mosaicism: a case study of the River Kamajohka, northernmost Finland. Journal of Biogeography 15: 703-719. Katibath E.F. 1984. A brief history of riparian forests in the Central Valley of California. In: Warner R.E. & Hendrix K.M. (eds.), California Riparian Systems: Ecology, Conservation, and Productive management. University of California Press, Berkeley. pp 23-29. Keay R.W.J. 1949. An example of Sudan zone vegetation in Nigeria. Journal of Ecology 37: 335-64. Keay R.W.J. 1959. Derived savanna - derived from what? Bulletin IFAN série A 21: 427-38. Keay R.W.J. 1960. An example of Northern Guinean Zone vegetation in Nigeria. Nigerian Forestry Information Bulletin 1: 1-46. Keay R.W.J & Hepper F.N. (eds.). 1954-1972. Flora of West Tropical Africa, 2nd edition. Millbank, London.Vol. I, (828p), Vol. II (544p), Vol. III (574p). Keay R.W.J., Onochie C.F.A. & Standfield D.P. 1964. Nigerian trees. Federal Department of Forest Research, Ibadan, Nigeria. Vol 1 (334p), Vol 2 (495p). 176 References Keita S.M., Arnason J.T., Baum B.R., Marles R., Camara F. & Traoré A.K. 1999. Etude ethnopharmacologique traditionnelle de quelques plantes médecinales anthelminthiques de la Haute-Guinée (République de Guinée). Revue de Médecines et Pharmacopées Africaines 13: 49-64. Kellman M. & Tackaberry R. 1993. Disturbance and tree species coexistence in tropical riparian forest fragments. Global Ecology and Biogeography Letters 3: 1-9. Kellman M., Tackaberry R., Brokaw N. & Meave J. 1994. Tropical gallery forests. National Geographical Research and Exploration 10: 92-103. Kellman M, Tackaberry R. & Meave J. 1996. The consequences of prolongued fragmentation: lessons from tropical gallery forests. In: Schelhas J. & Greenberg R. (eds.), Forest patches in tropical landscapes. Island Press, Washington D.C. pp 37-58. Kellman M. & Meave J. 1997. Fire in the tropical gallery forests of Belize. Journal of Biogeography 24 (1): 23-34. Kent M. & Coker P. 1992. Vegetation description and analysis: A practical approach. John Wiley & Sons, England. 363 p. Kerharo J. & Adam J.G. 1974. La pharmacopée sénégalaise traditionnelle : Plantes médecinales et toxiques. Edition Vigot Frères, Paris, France. 1011p. Kingdon J. 1990. Island Africa: The evolution of Africa’s rare animals and plants. William Collins Sons & Co., London. 287p. Kish L. 1965. Survey sampling. John Wiley & Sons, New York. Knopf F.L. 1985. Significance of riparian vegetation to breeding birds across an altitudinal cline. USDA Forest Service, General Technical Report RM – 120, Rocky Mountain Forest and Range Experiment Station, Fort Collins, Colorado. pp 105-111. Knopf F.L., Johnson R.R., Rich T., Samson F.B. & Szaro R.C. 1988. Conservation of riparian ecosystems in the United States. Wilson Bulletin 100: 272-284. Knutson K.L. & Nae V.L. 1997. Management recommendations for Washington priority habitats: riparian. Washington Departement of Fish and Wildlife, Olympia, WA. Kokou K. 1998. Les mosaïques forestières au sud du Togo: Biodiversité, dynamique et activités humaines. Thèse Doctorat de l’Université de Montpellier II. Montpellier, France. 140p + Annexes. Kokou K. & Caballé G. 2000. Les îlots forestiers de la plaîne côtière togolaise. Bois et Forêts Tropicaux 263 (1): 39-50. Kokou K., Caballé G., Akpagana K. & Batawila K. 1999. Les îlots forestiers au sud du Togo: Dynamique et relations avec les végétations périphériques. Revue d’Ecologie (Terre et Vie) 54: 301-314. Kokou K., Couteron P., Martin A. & Caballé G. 2002. Taxonomic diversity of lianas and wines in forests fragments of southern Togo. Revue d’Ecologie (Terre Vie) 57: 3-17. Kouamé F.N. & Traoré D. 2001. Exploitation forestière et les lianes de forêt dense tropicale humide: cas de la forêt classée du Haut sassandra. In: Rapport d’activité 1999-2000 du Centre Suisse de Recherche Scientifique en Côte d’Ivoire (CRRS). Abidjan, Côte d’Ivoire. pp 20-21. Large A.R.G. & Petts G.E. 1994. Rehabilitation of rivers margins. In: Calow P. & Petts G.E. (eds.), The rivers handbook, hydrological and ecological principles. Blackwell Scientific Oxford. pp 401-418. Larue P., Bélanger L. & Huot J. 1995. Riparian edge effects on boreal balsam fir bird communities. Canadian Journal of Forest Resources 25: 555-566. Lawson G.W., Armstrong-Mensah K.O. & Hall J.B. 1970. A catena in tropical moist semi-deciduous forest near Kade, Ghana. Journal of Ecology 58: 371-398. 177 References Lebrun J. 1947. La végétation de la plaine alluviale au sud du lac Edouard, Exploration du Parc National Albert. Mission J. Lebrun. Institut des Parcs Nationaux du Congo Belge. 800p. Lebrun J. & Gilbert G. 1954. Une classification écologique des forêts du Congo. Publication I.N.E.A.C. Série Science 63: 89p. Lebrun J-P. & Stork L. 1991-1997. Enumération des plantes à fleur d’Afrique Tropicale. CIRATEMVT- Montpellier. Editions des Conservatoire et Jardin botaniques de Genève. Vol. I Publication hors-série no.7; 249p.; Vol. II Publication hors-série no.7a; 257p.; Vol. III Publication hors-série no.7b; 341p.; Vol. IV Publication hors-série no.7c; 712p. Leinard R.E., Wyman S.K. Sue Noggles R. & Volk W.P. 1999. Classification of riparian plant communities using the ecological site concept. VIth International Rangeland Congress Proceedings 1: 249-251. Lejoly J. 1994. Mise en place des travaux en vue des inventaires de biodiversité dans la forêt de Ngotto (République Centrafricaine). Projet Ecofac, Agreco-CTFT, Bruxelles. 109p. Lejoly J. & Mandango A. 1982. L’association arbustive ripicole à Alchornea cordifolia dans le hautZaire. Studies on Aquatic Vascular Plants, Royal Botanical Society of Belgium, Brussels. pp 257-265. Lejoly J., Sonké B. & van Essche K. 1996. Utilisation de la méthode du transect en vue de l’étude de la biodiversité dans la réserve de Faune du Dja (Cameroun). In: van der Maesen L.J.G., van der Burgt X.M. & van Medenback de Rooy J.M. (eds.), The biodiversity of African plants. Proceedings of the XIVth AETFAT congress. Wageningen, The Netherlands. pp 150-154. Léonard J. 1947. Contribution à l’étude des formations ripicoles arbustives et arborescentes de la région d’Eala. C.R. semaine agricole de Yangambi (du 26 février au 5 mai 1947). Publication I.N.E.A.C. Buxelles, 2ème partie 67: 863-877. Léonard J. 1952. Aperçu préliminaire des groupements végétaux pionniers dans la région de Yangambi (Congo Belge). Vegetatio 3: 279-297. Léonard J. 1965. Contribution à la subdivision phytogéographique de la région guinéo-congolaise d’après la répartition géographique d’Euphorbiacées d’Afrique tropicale. Webbia 19: 627649. Levaux P. 1990. Etude des relations faune-végétation sur transects au Parc National de la Pendjari au Bénin. Ecole Nationale du Génie Rural des Eaux et Forêts-Centre de Montpellier, France. 37p + annexes. Levine C.M. & Stromberg J.C. 2001. Effects of flooding on native and exotic plant seedlings: implications for restoring south-western riparian forests by manipulating water and sediment flows. Journal of Arid Environments 49: 111-131. Leyden B. 1985. Late quaternary aridity and Holocene moisture fluctuations in the Lake Valencia, Venezuela. Ecology 66: 1279-1295. Lock P.A. & Naiman R.J. 1998. Effects of stream size on bird community structure in coastal temperate forests of the Pacific Northwest, USA. Journal of Biogeography 25: 773-782. Loizeau P.A. 1992. La diversité: exemple des forêts denses humides amazoniennes. Saussurea 23: 4982. Lonard R.I., Judd F.W., Everitt J.H., Escobar D.E., Davis M.R., Crawford M.M. & Desai M.D. 2000. Evaluation of color-infrared photography for distinguishing annual changes in riparian forests vegetation of the lower Rio Grande in Texas. Forest Ecology and Management 128: 75-81. Lowrance R.R., Todd R.L. & Asmussen L.E. 1984. Nutrient cycling in an agricultural watershed-I phreatic movement. Journal of Environmental Quality 13: 22-27. Lowrance R.R., Altier L.S., Newbold J.D., Schnabel R.R., Groffman P.M., Denver J.M., Correll D.L., Gilliam J.W., Robinson J.L., Brinsfield R.B., Staver K.W., Lucas W. & Todd A.H. 1997. 178 References Water quality functions of riparian forest buffers in Chesapeak Bay watersheds. Environment Management 21: 687-712. Lykke A.M. & Goudiaby A. 1999. Structure, floristic composition, and conservation potentials of a remnant gallery forest at ‘Mare du Dragon’, Senegal. Nordic Journal of Botany 19: 561-574. Lykke A.M. 1996. How gallery forest turms into savanna: an example from Senegal. In: van der Maesen L.J.G., van der Burgt X.M. & van Medenbach de Rooy J.M. (eds.), The biodiversity of African plants. Proceedings of the XIVth AETFAT congress. Wageningen, The Netherlands. pp 150-154. MacDougall A. & Kellman M. 1992. The understorey light regime and patterns of tree seedlings in tropical riparian forest patches. Journal of Biogeography 19: 667-675. Machtans C.S., Villard M.A. & Hannon S.J. 1996. Use of riparian buffer strips as movement corridors by forest birds. Conservation Biology 10: 1366-1379. Magurran A.E. 1988. Ecological diversity and its measurement. Cambridge University Press, Cambridge. 179 p. Magurran A.E. & Henderson P.A. 2003. Explaining the excess of rare species in natural species abundance distributions. Nature 422: 714-716. Malanson G.P. 1993. Riparian lansdcapes. Cambridge University Press, Cambridge. Malanson G.P. & Armstrong M.P. 1996. Dispersal probability and forest diversity in a fragmented landscape. Ecological Modelling 87: 91-102. Maley J. 1989. 20.000 ans d’évolution des climats du Sahara Central aux savanes tropicales humides. In: Centre Culturel Français d’Abidjan (ed.), Les climats, Sud-Sahara, Sahel Nord. pp 34-39 et 42-50. Maley J. 1991. The African rain forest vegetation and paleoenvironments during late quaternary. Climatic Change 19: 79-98. Maley J. 1996. The African rain forest – main characteristics of changes in vegetation and climate from the Upper Cretaceous to the Quaternary. In: Alexander I.J., Swaine M.D. & Watling R. (eds.), Essays on the Ecology of the Guinea-Congo Rain Forest. Proceedings of the Royal Society of Edinburgh 104B: 31-73. Maley J. 1997. Middle to late Holocene changes in tropical Africa and the other continents: paleomonsoon and sea surface temperature variations. In: Dalfer H.N., Kukla G. & Weis H. (eds.), Third Millenium BC Climate Change and World Collapse. Nato ASI Series I 49: 611-640. Maley J. 2001. Si la forêt tropicale m'était contée. La destruction catastrophique des forêts d'Afrique centrale survenue il y a environ 2500 ans exerce encore une influence majeure sur la répartition actuelle des formations végétales. Canopée (Journal du Programme Européen ECOFAC) Libreville 19: 6-9. Maley J. 2002. A catastrophic destruction of African forests about 2,500 years still exerts a major influence on the present vegetation formations. IDS Bulletin 33 (1): 13-30. Maley J. & Chepstow-Lusty A. 2001. Elaeis guineensis Jacq. (oil palm) fluctuations in central Africa during the late holocene: climate or human driving forces for this pioneering species. Vegetation History and Archaeobotany 10: 117-120. Mandango A. 1982. Flore et végétation des îles du fleuve Zaïre dans la sous-région de la Tshopo (Haut-Zaire). Thèse Doctorat de l’Université de Kisangani. Zaire, 2 tomes. 425p. Mandango A. & Ndjele M. 1986. Végétation aquatique et semi-aquatique de l’île Kongolo à Kisangani (Zaire). Bulletin de la Société Royale Botanique de Belgique 119(1): 63-73 Mander Ü., Kuusemtes V., Lõhmus K. & Mauring T. 1997. Efficiency and dimensioning of riparian buffer zones in agricultural catchments. Ecological Engineering 8: 299-324. 179 References Mangenot G. 1955. Etude sur les forêts des plaines et plateaux de la Côte d’Ivoire. Etudes Eburnéennes IFAN 4 : 5-61. Manuwal D.A. 1983. Spring bird communities in the southern Washington Cascade range. In: Ruggiero L.F., Aubry K.B., Carey A.B. & Huff M.H. (eds.), Wildlife and vegetation of unmanaged Douglas-fir forests. USDA Forest Service General Technical Report PNW-GTR285. Pacific Northwest Research Station, Portland, Oregon. Martijena N.E. 1998. Soil properties and seedling establishment in soils from monodominant and high-diversity stands of the tropical deciduous forests of Mexico. Journal of Biogeography 25: 707-719. Martijena N.E. & Bullock S.H. 1994. Monospecific dominance of a tropical deciduous forests in Mexico. Journal of Biogeography 21: 63-74. Martin C. 1991. The rainforests of West Africa. Birkhauser Verlag Publishers. Basel. Switzerland. 235p. McClain M.E., Richey J.E. & Pimentel T.P. 1994. Groundwater nitrogen dynamics at the terrestriallotic interface of a small catchment in the Central Amazon Basin. Biogeochemistry 27:113127. McDowell W.H., Bowden W.B. & Asbury C.E. 1992. Riparian nitrogen dynamics in two geomorphological distinct tropical rain forest watersheds: subsurface solute patterns. Biogeochemistry 18: 77-99. McGarigal K. & McComb W.C. 1992. Streamside versus upslope breeding bird communities in the central Oregon Coast Range. Journal of Wildlife Management 56: 10-23. Meave J., Kellman M., MacDougall A. & Rosales J. 1991. Riparian habitats as tropical forest refugia. Global Ecology and Biogeography Letters 1: 69-76. Meave J. & Kellman M. 1994. Maintenance of rain forest diversity in riparian forests of tropical savannas: implications for species conservation during Pleistocene drought. Journal of Biogeography 21: 121-135. Medina A.L. 1990. Possible effects of residual development on streamflow, riparian plant communities, and fisheries on small mountain streams in Central Arizona. Forest Ecology and Management 33-34: 351-361. Medley M.E. 1992. Patterns of forest diversity along the Tana River, Kenya. Journal of Tropical Ecology 8: 358-371. Melick D.R. 1990. Ecology of rainforest and sclerophyllous communites in the Mitchell River National Park, Gippsland, Victoria. Proccedings of the Royal Society of Victoria 102: 71-87. Melick D.R. & Ashton D.H. 1991. The effects of natural disturbances on warm temperate rainforests in south-eastern Autralia. Austalian Journal of Ecology 39:1-30. Meurer M. 1994. Etudes sur le potentiel d’herbage dans les savanes du Nord Ouest du Bénin. Agriculture et Développement Rural 1(1): 37- 41. Minshall G.W., Petersen C.R.Jr. & Curtis N.F. 1985. Species richness in streams of different size from the same drainage basin. American Naturalist 125(1): 16-38. Molles Jr.M.C., Crawford C.S., Ellis L.M., Valett H.M. & Dahm C.N. 1998. Managed floods: restoration of riparian forest ecosystem structure and function along the Middle Rio Grande. BioScience 48: 749-756. Mondjannagni A. 1969. Contribution à l’étude des paysages végétaux du Bas-Bénin. Annales de l’Université d’Abidjan Série C 1(2): 191p. Monnier Y. 1990. La poussière et la cendre: Paysages, dynamique des formations et stratégies des sociétés en Afrique de l’Ouest. 2ème édition. ACCT, Paris, France. 264 p. Montgomery D.R. 1997. What’s best on the banks? Nature 388: 328-329. 180 References Mori S.A. & Boom B.M. 1987. The Lecythidaceae of a lowland Neotropical Forest. La fûmée Mountain, French Guiana. Memoire New York Botanical Garden 44: 9-29. Murray N.L. & Stauffer D.F. 1995. Nongame bird use of habitat in central Appalachian riparian forests. Journal of Wildlife Management 59: 78-88. Naiman R.J., Beechie T.J., Benda L.E., Berg D.R., Bison P.A., MacDonald L.H., O’Connor M.D., Olson P.L. & Stell E.A. 1992. Fundamental elements of ecologically healthy watershed. In: Naiman R.J. (ed.), Watershed management: balancing sustainability and environmental change. Springer-verlag, New York. pp 127-188. Naiman R.J. & Décamps H. 1997. The ecology of interface: riparian zones. Annual Review of Ecology and Systematics 28: 621-658. Naiman R.J. & Décamps H. (eds.) 1990. The ecology and management of aquatic-terrestrial ecotones. Man and the Biosphere series Vol. 4. UNESCO, Paris and Parthenon Publishing group, Carnforth. 316p. Naiman R.J., Décamps H. & Polleck M. 1993. The role of riparian corridors in maintaining regional biodiversity. Ecological Applications 3: 209-212. Naiman R.J., Melillo J.M., Lock M.A., Ford T.E. & Reice S.R. 1987. Longitudinal patterns of ecosystem process and community structure in a subarctic river continuum. Ecology 68: 11391156. Nakashizuka T., Zulkifli Yusop & Abdoul Rahim Nik 1992. Altitudinal zonation of forest communities in Selangor, Peninsular Malaysia. Journal of Tropical Forest Science 4 (3): 233244. National Research Council (NRC) 2002. Riparian areas: Functions and strategies for management. Summary. WSTB, CRZFSM. National Academy Press, Washington. 22p. Natta A.K. 1997. Contribution à l’étude et aménagement des pâturages naturels dans le Nord-Ouest de l’Atacora. Mémoire d’Ingénieur Agronome. DAGE/FSA/UNB, Abomey-Calavi, Bénin. 199p. Natta A.K. 2000. Tree species diversity assessment in riparian forests using remote sensing, geographical information systems and geostatistics (case study of Toui-Kilibo protected forest in Benin). MSc thesis, ITC Forest Science Division, Enschede, The Netherlands. 91p. Natta A.K. & Porembski S. (in press). Ouémé and Comoé: forest-savanna border relationships in two riparian ecosystems in West-Africa. Botanische JahrbĦcher 00: 00-00. Natta A.K., Sinsin B. & van der Maesen L.J.G. 2002. Riparian forests, a unique but endangered ecosystem in Benin. Notulae Florae Beninensis 4. Botanische JahrbĦcher 124: 55-69. Natta A.K., Sinsin B. & van der Maesen L.J.G. (in press). A phytosociological study of Riparian forests in Benin (West Africa). Belgian Journal of Botany 00:00-00. Natta A.K., Sinsin B. & van der Maesen L.J.G. A phytosociological study of the Garcinia livingstonei and Combretum acutum gallery forest plant community of North West Benin (West Africa) (Submitted to the Acta Botanica Gallica). Natta A.K. & van der Maesen L.J.G. Assessment of riparian forest fragments plant diversity in West African savanna region: an overview from Benin Republic. (Submitted to the Journal of Biogeography). Nierenberg T. & Hibbs D.E. 2000. A characterisation of unmanaged riparian areas in the central Coast range of western Oregon. Forest Ecology and Management 129: 195-206. Nilsson C.1992. Conservation management of riparian communities. In: Hansson L. (ed.), Ecological principles of nature conservation. Elsevier Applied Science, London. pp 352-372. Nilsson C., Ekbald A., Dynesius M.A., Backe S., Gardfjell M., Carlberg B., Hellqvist S & Jansson R. 1994. A comparison of species richness and traits of riparian plants between a main river channel and its tributaries. Journal of Ecology 82: 281-298. 181 References Nilsson C., Grelsson G., Johansson M. & Sperens U. 1989. Patterns of plant species richness along riverbanks. Ecology 70(1): 77-84. Nilsson C., Jansson R. & Zinko U. 1997. Long-term responses of river-margin vegetation to waterlevel regulation. Science 276: 798-800. Norris V.O.L. 1993. The use of buffer zones to protect water quality: a review. Water Resources Management 7: 257-272. Nye H. 1954. Some soil processes in humid tropics. 1. A field study of a catena in the West African Forest. Journal of Soil Science 5: 7-21. Oates J.F. 1996. Survey of Cercopithecus erythrogaster populations in the Dahomey Gap. African Primates 2(1): 9-11. Ohmart R.D. 1994. The effects of human-induced changes on the avifauna of western riparian habitats. Studies in Avian Biology 15: 273-285. Oldfield S., Lusty C. & Mackinven A. 1998. The World List of Threatened Trees. World Conservation Press, Cambridge, UK. 650p. Onochie C.F.A. 1979. The Nigerian rainforest ecosystem. In: Okali D.U.U (ed.), The Nigerian rainforest ecosystem. Proceedings of the Man and Biosphere Workshop on the Nigerian rainforest Ecosystem. University of Ibadan 24-26 January 1979. Ibadan, Nigeria. pp 1-37. Ott L.R. & Longnecker M. 2001. An introduction to Statistical methods and data analysis. Fifth edition. Thomson Learning Academic resource Center. Duxbury, USA. 1152p. Otto T.S. 1993. Biodiversity: A means for coping with environmental changes, and a basic property of life. Wildlife and Nature 9(3): 5-9. Ouattara N. 1999. Evolution du taux de germination de semences oléagineuses en fonction du mode et de la durée de conservation, Cas de Pentadesma butyracea Sabine (Lami). In: Ouédraogo A.S., Boffa J.-M. (eds.), Vers une approche régionale des ressources génétiques forestières en Afrique sub-saharienne. Actes du premier atelier régional de formation sur la conservation et l’utilisation durable des ressources génétiques forestières en Afrique de l’Ouest, Centrale et Madagascar, Centre National de Semences Forestières, Ouagadougou, Burkina-Faso 1627/03/98. pp 170-174. Pabst R.J. & Spies T.A. 1999. Structure and composition of unmanaged riparian forests in the coastal mountains of Oregon. Canadian Journal of Forest Resources 29: 1557-1573. Paradis G. 1975a. Observations sur les forêts marécageuses du Bas-Dahomey: localisation, principaux types et évolution au cours de quaternaire récent. Annales de l’Université d’Abidjan E 8(1): 281-315. Paradis G. 1975b. Physionomie, composition floristique et dynamisme des formations végétales d’une partie de la basse Vallée de l’Ouémé (Dahomey). Annales de l’Université d’Abidjan C 11: 55101. Paradis G. 1976. Contribution à l’étude de la flore et de la végétation littorale du Dahomey. Bulletin du Musée National d’Histoire Naturelle Paris 383: 33-68. Paradis G. 1979. Observation sur Laguncularia racemosa Gaertn. et Dalbergia ecasstaphyllum (L.) Taub. dans les successions secondaires de la mangrove du Bénin. Bulletin de l’Institut Francais d’Afrique Noire A 41(1): 92-102. Paradis G. 1980. Un cas particulier de zones dénudées dans les mangroves d’Afrique de l’Ouest: celles dues à l’extraction de sel. Bulletin du Musée National d’Histoire Naturelle Paris, 4, serie 2, section B 3: 227-261. Paradis G. 1981. Ecologie et géomorphologie littorale en climat sub-équatorial sec: la végétation côtière du Bas-Bénin occidental. Annales de l’Université d’Abidjan E 14: 7-63. Paradis G. 1983. A phytogeographic survey of Southern Benin. Bothalia 14 (3-4): 579-585. 182 References Paradis G. 1988. Etude comparative des végétations littorales du Bénin et de la Côte-d’Ivoire. Thèse de Doctorat es Sciences. Université de Bordeaux III, Bordeaux, France. 295p. Paradis G. & Adjanohoun E. 1974. L’impact de la fabrication du sel sur la végétation de mangrove et la géomorphologie dans le Bas-Dahomey. Annales de l’Université d’Abidjan E 7: 599-612. Paradis G., de Souza S. & Houngnon P. 1978. Les stations à Lophira lanceolata dans la mosaïque forêt-savane du Sud-Bénin (ex Sud-Dahomey). Bulletin du Musée National d’Histoire Naturelle Paris, Botanique 521: 39-58. Paradis G. & Houngnon P. 1977. La végétation de l’aire classée de la Lama dans la mosaïque forêtsavane du Sud-Bénin (ex Sud-Dahomey). Bulletin du Musée National d’Histoire Naturelle Paris 503: 169-198. Paradis G. & Rabier J. 1979. La végétation de la lagune de Porto-Novo (Bénin) avant la fermeture du chenal de Cotonou en 1978. Annales de l’Université d’Abidjan E 12: 7-28. Parren M.P.E. 2003. Lianas and logging in West Africa. Tropenbos-Cameroon Series 6. PhD thesis, Wageningen University, The Netherlands.168p. Pauwels L. 1993. Nzayilu N’Ti. Guide des arbres et arbustres de la région de Kinshasa - Brazaville. Jardin Botanique National de Belgique, Meise, Belgique. 495p. Peterjohn W.T. & Correl D.L. 1984. Nutrient dynamics in an agricultural watershed: observation on the role of a riparian forest. Ecology 65(5): 1466-1475. Peterson D.L. & Rolfe G.L. 1982. Nutrient dynamics and decomposition of litterfall in floodplain and upland forest of Central Illinois. Forest Science 28: 667-681. PGRN/IUCN. 1994. Preparation of a strategy for conservation and management of protected areas: Ecology and biodiversity components. MDR, Cotonou, Benin. 87p + annexes. Piegay H. 1997. Riparian forests, a structural compartment of the intra-alpine and piedmont river hydrosystem. Houille-Blanche 52(1-2): 13-18. Piegay H. & Bravard J-P. 1997. Response of a Mediterranean riparian forests to a 1 in 400 year flood, Ouveze River, Drome-Vaucluse, France. Earth Surface Processes and Landforms 22(1): 3143. Pinay G. & Décamps H. 1988. The role of riparian woods in regulating nitrogen fluxes between alluvial aquifer and surface water: a conceptual model. Regulated Rivers 2: 507-516. Poilecot P., Bonfou K., Dosso H., Lauginie F., N’Dri K., Michel N. & Sangaré Y. 1991. Un écosystème de savane soudanienne: Le parc de la Comoé, Côte d’Ivoire. Projet UNESCO/PNUD, No. IVC/87/007. Note technique 2 MAB. Paris, France. pp 1-163. Polansky C. 1994. Sommaire de l’inventaire quantitatif de la forêt classée de Toui-Kilibo. PGRN/VAF. DFRN/MDR, Cotonou, Bénin. 68 p. Popma J., Bongers F. & Meave J. 1988. Patterns in the vertical structure of the tropical lowland rainforest at Los Tuxtlas, Mexico. Vegetatio 74: 81-91. Porembski S. 2001. Phytodiversity and structure of the Comoé River Gallery forest (NE Ivory-Coast). Life forms and dynamics in Tropical Forests. Dissertationes Botanicae 346:1-10. PRRF 1998. Plan d’aménagement participatif de la forêt classée de Pénéssoulou. Projet de Restauration des Ressources Forestières dans la région de Bassila, PRRF/DFRN/MAEP – GTZ/Luso - Consult. Bassila, Bénin. 86 p + cartes. Projet Flore du Bénin (PFB). 1997. Projet d’Aménagement d’un Herbier National et de Rédaction de la flore Analytique du Bénin. Wageningen et Cotonou. 25p. Pulliam H.R., Dunning J.B. & Lui J. 1992. Populations dynamics in complex landscapes: a case study. Ecological Applications 2: 165-177. Purba K. & Sumarua E. 1987. Chemical analysis of twenty seven wood species from West-Java. Jurnal Penelitian Hasil Hutan 4: 23-29. 183 References Rachman O., Balfas J. & Sifat P.K. 1987. Machining properties of wood species from West-Java. Jurnal Penelitian Hasil Hutan 4: 54-64. Raunkiaer C. 1934. The life forms of plants and statistical plant geography. Clarendron Press, London. 632p. Raynaud J. & Georgy G. 1969. Nature et chasse au Dahomey: Manuel pratique. Secrétariat d’Etat aux Affaires Etrangères pour la Coopération. Paris, France. 320p. Reid W.V. 1998. Biodiversity hotspots. Tree 13(7): 275-280. Reitsma J.M. 1988. Végétation forestière du Gabon. Technical series 1. Tropenbos foundation, Ede, The Netherlands. 142p. République du Bénin (RB). 1993. Loi Nº 93-009 du 2 juillet 1993 portant régime des forêts en République du Bénin. Présidence de la République, Cotonou, Bénin. 26p. Richards P.W. 1939. Ecological studies on the rain forest of southern Nigeria. I. The structure and floristic composition of the primary forest. Journal of Ecology 27: 1-61. Richards P.W. 1952. The tropical rain forest, an ecological study. Cambridge University Press, Cambridge. 450p. Richards P.W. 1963. Ecological notes on West African vegetation. II. Lowland forest of the Southern Bakundu Forest Reserve. Journal of Ecology 51: 123-149. Robinson C.T., Tockner K & Ward J.V. 2002. The fauna of dynamic riverine landscape. Freshwater Biology 47: 661-677. Roché J. 1993. The use of historical data in the ecological zonation of rivers: the case of the ‘tern zone’. Vie et Milieu 43: 27-41. Roggeri H. 1995. Zones humides tropicales d'eau douce. Guide des connaissances actuelles et de la gestion durable. WIW, Union Européenne, Université de Leiden. Kluwer Academic Publishers, Pays-Bas. 385p. Roods S.B. & Mahoney J.M. 1990. Collapse of riparian poplar forest downstream from dams in western prairies: probable causes and prospects for mitigation. Environmental Management 14: 451-464. Roods S.B., Mahoney J.M., Reid D.E. & Zilm L. 1995. Instream flows and the decline of riparian cottonwoods along the St. Mary River, Alberta. Canadian Journal of Botany 73:1250-1260. Rottenborn S.C. 1999. Predicting the impacts of urbanisation on riparian bird communities. Biological conservation 88: 289-299. Russell-Smith J. 1991. Classification, species richness, and environmental relations of monsoon rain forests in northern Australia. Journal of Vegetation Science 2: 259-278. Saab V. 1999. Importance of spatial scale to habitat use by breeding birds in riparian forests: a hierarchical analysis. Ecological Applications 9: 135-151. Sagers C.L. & Lyon J. 1997. Gradient analysis in a riparian landscape: contrasts among forests layers. Forest Ecology and Management 96: 13-26. Sala O.E., Chapin I.F.S., Armesto J.J., Berlow E., Bloomfield J., Dirzo R., Huber Sanwald E., Huenneke L.F., Jackon R.B., Kinzig A., Leemans R., Lodge D.M., Mooney H.A., Oesterheld M., Poff N.L., Sykes M.T., Walker B.H., Walker M. & Wall D.H. 2000. Global biodiversity scenarios for the year 2100. Science 287: 1770-1774. Salley M.N. & Nieng F.S. 1997. Ouvrage sur l'aménagement durable des forêts. Etude FAO des forêts n° 122. FAO Rome, Italie. 305p. Salo J., Kalliola R., Häkkinen I., Mäkinen Y., Niemelä P., Puhakka M. & Coley P.D. 1986. River dynamics and the diversity of Amazon lowland forest. Nature 322: 254-258. 184 References Sayer J.A. & Wegge P. 1992. Biological conservation issues in forest management. In: Blockhus J.M., Dillenbeck M., Sayer J.A. & Wegge P. (eds.), Conserving biological diversity in managed tropical forests. Workshop held at the IUCN General Assembly at Perth, Australia, 30 november – 01 december 1990. pp 1-4. Schmitz A. 1963. Aperçu sur les groupements végétaux du Katanga. Bulletin Société Royale Botanique de Belgique 96(2): 233-347. Schmitz A. 1971. La végétation de la plaine de Lubumbashi (Haut-Katanga). Publications INEAC, R.D.C (Zaire). Série Scientifique no. 113. 388p. Schmitz A. 1988. Révision des groupements végétaux décrits du Zaire, du Rwanda et du Burundi. Musée Royal de l’Afrique Centrale, Tervuren, Belgique. Annales Sciences Economiques 17: 315p. Schnabel R.R., Cornish L.F., Stout W.L. & Shaffer J.A. 1996. Denitrification in grassed and a wooded, valley and ridge, riparian ecotone. Journal of Environmental Quality 25: 1230-1235. Schnell R. 1952. Contribution à l’étude phytosociologique et phytogéographique de l’Afrique occidentale. Les groupements et les unités géobotaniques de la région Guinéenne. Mémoire IFAN, Dakar 18:145-234. Schnell R. 1971. Intoduction à la phytogéographie des pays tropicaux. Volume II: Milieux, les groupements végétaux. Edition Gauthier-Villars, Paris, France. pp 503-951. Schnell R. 1976. La flore et la végétation de l’Afrique tropicale. Volume III, 1ère partie. Edition Gauthier-Villars, Paris, France. 460 p. Schreckenberg K. 1996. Forests, field and markets: A study of indigenous Tree Products in the Woody savannas of the Bassila region. PhD thesis. Department of Geography, School of African and Oriental studies. University of London, UK. 326p. Sedell J.R., Bisson P.A., Swanson F.J. & Gregory S.V. 1988. What we know about trees that fall into streams and rivers. In: Maser C., Tarrant R.F. & Trappe J.M. (eds.), From the Forest to Sea: a story of fallen trees. General Technical Report PNW-GTR-229. USDA Forest Service Pacific Northwest Resource Station, Portland, OR. pp 47-81. Shure D.J., Gottschalk M.R. & Parsons K.A. 1986. Litter decomposition processes in a floodplain forest. American Midland Naturalist 115: 314-327. Simberloff D. & Gotelli N. 1984. Effects of insularisation on plant species richness in the prairieforest ecotone. Biological Conservation 29: 27-46. Sinadouwirou Th.A. 1997. Forêts rémanentes et conservation de la biodiversité: Etude écologique de quelques forêts dans le département de l’Atlantique au Sud-Bénin. Mémoire d’Ingénieur Agronome. DAGE/FSA-UNB, Abomey-Calavi, Bénin. 132p. Sinadouwirou Th.A. 2000. Produits forestiers non ligneux et développement durable: structure des peuplements naturels et importance socio-économique du Pentadesma butyracea dans la région de Bassila au Bénin. Mémoire de Master FASA/Université de Dschang, CRESA ForêtBois, Yaoundé, Cameroon. 78 p. Sinsin B. 1991. Exploitation des pâturages dans un système traditionnel d’élevage bovin dans le périmètre Nikki-Kalalé. IVth International Rangeland Congress. Montpellier, France 2: 660662. Sinsin B., Daouda I., Ahokpè E., Téhou A., Toko I., Coubéou P. & Fandohan S. 1997. Faune sauvage des forêts classées de Goungoun de la Sota et des environs. MDR/DFRN. Projet UNSO. Laboratoire d’Ecologie Appliquée (LEA/FSA/UNB). Cotonou, Bénin. 158p. Sinsin B., Eyog-Matig O., Sinadouwirou Th.A. & Assogbadjo A. 2002a. Caractérisation écologique des essences fourragères Khaya senegalensis (Desr.) et Afzelia africana Sm. suivant les gradients de latitude et de station au Bénin. In: Eyog-Matig O., Gaoué O.G. & Obel-Lawson E. (eds.), Development of appropriate conservation strategies for African forest trees 185 References identified as priority species by SAFORGEN member countries. IPGRI Nairobi, Kenya. pp 15-50. Sinsin B., Nobimé G., Téhou A., Bekhuis P. & Tchibozo S. 2002b. Past and present distribution of red-bellied monkey Cercopithecus erythrogaster erythrogaster in Benin. Folia primatologia 73: 116-123. Skagen S.K., Melcher C.P., Howe W.H. & Knopf F.L. 1998. Comparative use of riparian corridors and oases by migrating birds in southeast Arizona. Conservation Biology 12: 896-909. Skoglund J. 1990. Seed dispersing agents in two regularly flooded river sites. Canadian Journal of Botany 68: 754-760. Smith S.D., Willington A.B., Nachlinger J.L. & Fox C.A. 1991. Functional responses of riparian vegetation to streamflow diversion in the eastern Sierra Nevada. Ecological Applications 1: 89-97. Smith R.J. & Schaefer J.M. 1992. Avian characteristics of urban riparian strip corridor. Wilson Bulletin 104: 732-738. Snedecor G.W. & Cochran W.G. 1989. Statistical methods. 8th edition. Iowa State University Press, Iowa Ames. 503p. Sokpon N. 1995. Recherches écologiques sur la forêt dense semi-décidue de Pobè au Sud-Est du Bénin. Groupements végétaux, structure, régénération naturelle et chute de litière. Thèse Doctorat en Sciences Agronomiques. Université Libre de Bruxelles, Belgique. 350 p. Sokpon N. & Adjakidjè V. 1999. Inventaire et caractérisation des espèces et formations végétales des zones humides du Sud-Bénin (Rapport provisoire). Projet d’Aménagement des Zones Humides (PAZH). Cotonou, Bénin. 35p. Sokpon N. & Ouinsavi C. 2002. Utilisations du Khaya senegalensis en médecine traditionnelle au Benin. Revue de Médecines et Pharmacopées Africaines 16: 9-19. Sokpon N., Sinadouwirou Th., Gbaguidi F. & Biaou H. 2001. Aperçu sur les forêts édaphiques hygrophiles du Bénin. Belgian Journal of Botany 134(1): 79-93. Sonké B. & Lejoly J. 1998. Biodiversity study in Dja fauna reserve (Cameroon): using the transect method. In: Huxley C.R., Lock J.M. & Cutler D.F. (eds.), Chorology, Taxonomy and Ecology of the Floras of Africa and Madagascar. Royal Botanic Gardens, Kew. pp 171-179. Sounon B.B. 2001. Facteurs pédologiques, biologiques et socio-économiques de mise en jachère des terres dans le secteur de Wari-Maro - Igbomakoro. Mémoire DEA, Option Gestion de l’environnement. FLASH-UAC, Abomey-Calavi, Bénin. 89p. Sowunmi M.A. 1986. Change of vegetation with time. In: Lawson G.W. (ed.), Plant Ecology in West Africa. pp 273-307. Sowunmi M.A. 1999. The significance of oil palm (Elaeis guineensis Jacq.) in the late Holocene environments of west and west central Africa: a further consideration. Vegetation History and Archeobotany 8: 199-210. Spackman S.C. & Hughes J.W. 1995. Assessment of stream corridor width for biological conservation: species richness and distribution along mid-order streams in Vermont, USA. Biological Conservation 71:325-332. Sparovek G., Ranieri S.B.L., Gassner A., de Maria I.C., Schnug E.E., dos Santos R.F. & Joubert A. 2002. A conceptual framework for the definition of the optimal width of riparian forests. Agriculture, Ecosystems and Environment 90: 169-175. Stauffer D.F. & Best L.B. 1980. Habitat selection by birds of riparian communities: evaluating effects of habitat alterations. Journal of Wildlife Management 44: 1-15. STATISTICA® 5.1. 1998. Logiciel Statisitca. Editions 1998. Maisons Alfort, Paris Statsoft, France. 186 References Sterck F.J. & Bongers F. 2001. Crown development in tropical rain forests trees: patterns with tree height and light availability. Journal of Ecology 89: 1-13. Stromberg J.C. & Patten D.T. 1990. Variation in seed size of a Southwestern riparian tree, Arizona walnut (Juglans major). American Midland Naturalist 124: 269-277. Stromberg J.C., Richter B.D., Patten D.T. & Wolden L.G. 1993a. Response of a Sonoran riparian forest to a 10-year return flood. Great Basin Naturalist 53: 118-130. Stromberg J.C., Wilkins S.D. & Tress J.A. 1993b. Vegetation-hydrology models as management tools for velvet mesquite (Prosopis velutina) riparian ecosystems. Ecological Applications 3: 307314. Stromberg J.C., Tiller R. & Richter B. 1996. Effects of groundwater decline on riparian vegetation of semiarid regions: The San Pedro, Arizona. Ecological Applications 6(1): 113-131. Stromberg J.C. 1997. Growth and survivorship of Fremont cottonwood, Gooding willow, and salt cedar seedlings after large floods in central Arizona. Great Basin Naturalist 57: 198-208. Stromberg J.C. 2001. Biotic integrity of Platanus wrightii riparian forests in Arizona: first approximation. Forest Ecology and Management 142: 251-266. Stutz De Ortega L.C. 1987. Etudes floristiques de divers stades secondaires des formations forestières du Haut Parana (Paraguay Oriental). Structure, composition floristique et régénération naturelle: comparaison entre la forêt primaire et la forêt sélectivement exploitée. Candollea 42(1): 205-262. Swaine M.D. & Hall J.B. 1976. An application of ordination to the identification of forest types. Vegetatio 32(2): 83-86. Swaine M.D., Hall J.B. & Alexander I.J. 1987. Tree population dynamics at Kade, Ghana (19681982). Journal of Tropical Ecology 3: 331-345. Swaminathan S.M. 1990. Foreword. In: McNeely J.A., Miller K.M., Reid W.V., Mittermeier R.A. & Werner T. (eds.), Conserving the world’s biological diversity. IUCN, Gland Switzerland; WRI, CI, WWF- US and the Word Bank, Washington D.C. 155p. Swanson F.J. & Lienkamper G.W. 1982. Interaction among fluvial process, forest vegetation and aquatic ecosystems, South Fork Hoh River, Olympic National Park. In: Starkey E.E., Franklin J.F. & Mattews J.W. (eds.), Ecological research in the national parks of the Pacific Northwest. Oregon State University, Corvalis, Oregon. pp 30-35. Swanson F.J., Gregory S.V., Sedell J.R. & Campbell H.G. 1982. Land-water interactions: the riparian zone. In: Edmonds R.L. (ed.), Analysis of coniferous forest ecosystems in the western United States. Hutchinson Ross Publishing Company. Stroudsburg, Pennsylvania. pp 267-291. Swanson F.J., Johnson S.L., Gregory S.V. & Acker A.S. 1998. Flood disturbance in a forested mountain landscape. BioScience 48: 681-689. Szaro R.C. 1980. Factors influencing bird populations in southwestern riparian forests. Management of western forests and grassland for nongame birds. In: USDA Forest Service General Technical Report INT-86. Intermountain Forest and range Experiment Station, Ogden, Utah. pp 403-418. Szaro R.C. & Jakle M.D. 1985. Avian use of a desert riparian island and its adjacent shrub habitat. Condor 87: 511-519. Szaro R.C. 1990. Southwestern riparian forests plant communities: site characteristics, tree species distributions, and size-class structures. Forest Ecology and Management 33-34: 315-334. Tabacchi E., Planty-Tabacchi A-M. & Décamps O. 1990. Continuity and discontinuity of the riparian vegetation along a fluvial corridor. Landscape Ecology 5(1): 9-20. Taylor J.R., Cardonne M.A. & Mitsch W.J. 1990. Bottomland hardwood forests: their functions and values. In: Gosselink J.P., Lee L.C. & Muir T.A. (eds.), Ecological Processes and Cumulative 187 References impacts: Illustrated by Bottomland Hardwood Wetlands Ecosystems. Lewis Publishers, Chelsea, MI. pp 13-86. Tenté A.B.H. 2000. Dynamique actuelle de l'occupation du sol dans le massif de l'Atacora: Secteur Perma-Toukountouna. Mémoire DEA, Spécialité Gestion des Ressources Naturelles, Aménagement du Territoire et Politique Environnementale. FLASH/UAC, Abomey-Calavi, Bénin. 83p. Ter Braak C.J.F. 1987. The analysis of vegetation-environment relationships by canonical correspondence analysis. Vegetatio 69: 69-77. Ter Braak C.J.F. 1995. Ordination, Chapter 5. In: Jongman R.H.G., Ter Braak C.J.F. & Van Tongeren O.F.R. (eds), Data Analysis in Community and Landscape Ecology. Cambridge University Press, Cambridge, UK. pp 91-173. Ter Braak C.J.F. & Smilauer P. 1998. CANOCO reference manual and User’s guide to Canoco for Windows. Software for Canonical Community Ordination (version 4). Centre for Biometry Wageningen, CPRO-DLO. Wageningen, the Netherlands. 352p. Texier H., Colleuil B., Profizi J-P. & Dossou C. 1980. Le Lac Nokoué, environnement du domaine margino-littoral sud-béninois. Bulletin de l’Institut Géologique du Bassin d’Aquitaine Bordeaux 28: 115-142. Thomas D.H.L. 1996. Dam construction and ecological change in the riparian forest of the HadejiaJama'are floodplain, Nigeria. Land Degradation and Development 7(4): 279-295. Thompson S.K. 1992. Sampling. John Wiley, New York. 343p. Thompson S.K. & Seber G.A.F. 1996. Adaptive sampling. John Wiley, New York. Tilman D. 1988. Plant strategies and the dynamics and structure of plant communities. Princeton University Press, Princeton. 360p. Tongeren O. & Prentice C. 1986. A spatial simulation for vegetation dynamics. Vegetatio 65: 163173. Touré S.F. 1991. Arbres et arbustes fourragers dans l’alimentation des ruminants en zones sahéliennes et soudaniennes. Valeur nutritive espèces appétées (Sénégal). IVth International Rangeland Congress. Montpellier, France 2: 670-673. Tossou G.M. 2002. Recherche palynologique sur la végétation holocène du sud-Bénin (Afrique de l’Ouest). Thèse de Doctorat de l’Université de Lomé, Faculté des Sciences, Togo. 136p. Trimble S.W. 1999. Decreased rates of alluvial sediment storage in the Coon creek basin, Wisconsin, 1975-1993. Science 285: 1244-1246. Trochain J.L. 1957. Accord interafricain sur la définition des types de végétation de l’Afrique Tropicale. Bulletin de l’Institut d’Etude Centrafricaines, nouvelle série no. 13-14: 55-93. Tuani G.K., Cobbinah J.R. & Agbodaze P.K. 1994. Bioactivity and phytotechemical studies on extractives from some Ghanaian plants. Ghana Journal of Forestry 1: 44 - 48. UNCED. 1992. Biological Diversity Convention. United Nations Conference on Environment and Development. Rio de Janeiro June 1992, Brazil. van Groenewoud H. 1975. Theoretical considerations on the covariation of plant species along ecological gradients with regard to multivariate analysis. Journal of Ecology 63: 837-47. van der Maarel E. 1990. Ecotones and ecoclines are different. Journal of Vegetation Science 1: 135138. van Etten E.J.B. 1999. Vegetation maps and geographic information systems as tools for the management and planning of arid lands. VIth International Rangeland Congress Proceedings 1: 251-253. van Rompaey R.S.A.R. 1993. Forest gradients in West Africa: a spatial gradient analysis. PhD Thesis. Wageningen University, The Netherlands. 138p. 188 References Varty N. 1990. Ecology of small mammals in the Riverine forests of the Jubba Valley, Southern Somalia. Journal of Tropical Ecology 6: 179-189. Vivien J. & Faure J.J. 1985. Arbres des forêts d’Afrique Centrale. Ministère des Relations Extérieures de la Coopération et du Développement. Paris, France. 526 p. Vought L.B.M., Dahl J., Pedersen C.L. & Lacoursiére J.O. 1994. Nutrient retention in riparian ecotones. Ambio 23: 324-348. Wakentin I.G. & Reed J.M. 1999. Effects of habitat type and degradation on avian species richness in great Basin riparian habitats. Great Basin Naturalist 59: 205-212. Ward J.V., Tockner K., Arscott D.B. & Claret C. 2002. Riverine landscape diversity. Freshwater Biology 47: 517-539. Warner R.E. 1979. California riparian study program. Background information and proposed study design. California Resources Agency, Department of Fish and Game, Sacramento, California (USA). 177p. Warner R.E. & Katibath E.F. 1980. Measurement techniques for inventorying riparian systems. An International Workshop, November 30-December 6, 1980. La Paz, Mexico. pp 487-495. Weller D.E., Correll D.L. & Jordan T.E. 1994. Denitrification in riparian forests receiving agricultural discharges. In: Mitsch W.J. (ed.), Global wetlands: Old Worlds and New. Elsevier, Amsterdam. pp 117-132. Whitaker D.M. & Montevecchi W.A. 1997. Breeding bird assemblages associated with riparian, interior forest, and non riparian edge habitats in a balsam fir ecosystems. Canadian Journal of Forest Resources 27: 1159-1167. White F. 1979. The Guineo-Congolian region and its relationships to other phytocoria. Bulletin du Jardin Botanique National de Belgique 49: 11-55. White F. 1986. La végétation de l’Afrique. Mémoire accompagnant la carte de végétation de l’Afrique. UNESCO/AETFAT/UNSO. 384p. Whitmore T.C. 1990. An introduction to tropical rain forests. Clarendon Press, Oxford. 225p. Whittaker R.H. 1972. Evolution and measurement of species diversity. Taxon 21: 213-251. Whittaker R.H. 1975. Communities and ecosystems. 2nd edition. Macmillan Comp, New York. 385p. Willems H.P.L., Rotelli M. D., Berry D.F., Smith E.P., Reneau R.B. & Mostaghimi S. 1997. Nitrate removal in riparian wetlands soils: effects of flow rate, temperature, nitrate concentration and soil depth. Water Resources 31(4): 841-849. Williams P.H., Humphries C.J. & Vane-Wright R.I. 1991. Measuring biodiversity: Taxonomic relatedness for conservation priorities. Australian Systematic Botany 4: 665-79. Williamson M. 1975. The design of wildlife reserves. Nature 256: 519. Woinarski J.C.Z., Brock C., Armstrong M., Hempel C., Cheal D. & Brennan K. 2000. Bird distribution in riparian vegetation in the extensive natural landscape of Australia’s tropical savanna: a broad-scale survey and analysis of a distributional data base. Journal of Biogeography 27: 843-868. Wotto J. 2001. Dynamique des pâturages naturels dans le système agraire de la sous-préfecture de Péhunco. Mémoire de Maîtrise en Géographie. FLASH-UNB, Abomey-Calavi, Bénin. 116p. Xiang W.N. 1993. Application of a GIS-based stream buffer generation model to environmental policy evaluation. Environment Management 17: 817-827. Yédomonhan H. 2002. Contribution à l’étude des formations saxicoles au centre méridional du Bénin. Mémoire de Maîtrise en Géographie. FLASH/UAC, Abomey-Calavi, Bénin. 115p. Zepernick B. & Timler F.K. 1984. Guttiferae. In: Brunel J.F., Hiepko P., Scholz H. (eds), Flore analytique du Togo, Phanerogames, Englera 4. GTZ, Berlin. pp 216-217. 189 ANNEX LIST OF RIPARIAN FORESTS PLANT SPECIES OF BENIN A.K. Natta Riparian forests plant species of Benin List of riparian forests plant species of Benin no. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 LF PT Lnph Lnph Lnph Ch Lmph mPh mph mPh mph mph Th Th Lmph He He Lnph Lmph Lnph Lnph Ch Lmph Th Th Ch Ep Ch Th Ge Ge mPh Th LmPh LmPh LmPh mPh mPh mph mPh Lmph mph mph Th Th Th Th Th mPh Ge Ge Ge Ge Lmph Lmph Lmph mph Ge Ge TA TA Pan TA TA S SZ SZ SZ _ Pan Cosmo GC SG Pan GC GC GC G SZ TA TA GC Pan GC PRA Pan GC _ SZ Pan GC GC GC PRA GC GC GC GC GC GC TA Pan Cosmo Pan Pan S G G G G GC GC PRA Pan G SZ Families Papilionaceae Papilionaceae Papilionaceae Malvaceae Mimosaceae Mimosaceae Mimosaceae Mimosaceae Mimosaceae Mimosaceae Euphorbiaceae Amaranthaceae Malpighiaceae Poaceae Poaceae Passifloraceae Passifloraceae Passifloraceae Passifloraceae Papilionaceae Cucurbitaceae Asteraceae Asteraceae Adiantaceae Orchidaceae Papilionnaceae Papilionnaceae Zingiberaceae Zingiberaceae Caesalpiniaceae Asteraceae Apocynaceae Apocynaceae Apocynaceae Mimosaceae Mimosaceae Mimosaceae Mimosaceae Euphorbiaceae Sapindaceae Sapindaceae Poaceae Amaranthaceae Amaranthaceae Amaranthaceae Amaranthaceae Papilionnaceae Araceae Araceae Araceae Araceae Vitaceae Vitaceae Vitaceae Anacardiaceae Araceae Araceae Plant list Abrus canescens Welw. ex Wight & Arn. Abrus fruticulosus Wall. ex Wight & Arn syn. A. pulchellus Wall. ex Thwaites Abrus precatorius L. Abutilon mauritianum (Jacq.) Medic. Acacia erythrocalyx Brenan syn. Acacia pennata (L.) Willd. Acacia gourmaensis A. Chev. Acacia macrostachya Reichenb. ex DC. Acacia polyacantha Willd. subsp. campylacantha (Hochst. ex A. Rich.) Brenan Acacia sieberiana DC. var. sieberiana Acacia sp. Acalypha ciliata L. Achyranthes aspera L. Acridocarpus alternifolius (Schum. & Thonn.) Acroceras amplectens Stapf. Acroceras zizanoides (Kunth.) Dandy Adenia cissampeloides (Planch. ex Hook) Harms Adenia lobata (Jacq.) Engl. Adenia rumicifolia Engl. & Harms var. miegei Adenia tennuispira (Stapf.) Engl. Adenodolichos paniculatus (Hua) Hutch. & Dalz. syn. Dolichos paniculatus Hua Adenopus brevifolius Benth. Adenostemma caffrum DC. var. caffrum Adenostemma perrottetii DC. Adiantum philippensis L. Aerangis biloba (Lindl.) Schltr. excl. Aeschynomene afraspera J. Leonard Aeschynomene indica L. Aframomum sceptrum (Oliv. & Hanb.) K. Schum. Aframomum sp. Afzelia africana Smith ex Pers. Ageratum conyzoides L. subsp. conyzoides Alafia barteri Oliv. Alafia benthamii (Baill. ex Stapf) Stapf var. benthamii Alafia scandens (Thonning) De Wild. syn A. landolphioides Albizia coriaria Welw. ex Oliv. Albizia ferruginea (Guill. & Perr.) Benth. Albizia glaberrima (Schum. & Thonn.) Benth. Albizia zygia (DC.) J. F. Macbr. Alchornea cordifolia (Schum. & Thonn.) Müll. Arg. Allophyllus spicatus (Poir.) Radlk. Allophyllus africanus P. Beauv. syn A. cobbe (L.) Raeusch. Alloteropsis paniculata (Benth.) Stapf. syn. Urochloa paniculata Benth. Alternanthera sessilis (L.) DC. Amaranthus hybridus L. Amaranthus spinosus L. Amaranthus viridis L. Amblygonocarpus andongensis (Welw. ex Oliv.) Exell & Torre syn. Tetrapleura andongensis Amorphophallus aphyllus (Hook.) Hutch. Amorphophallus dracontioides (Engl.) N. E. Br. Amorphophallus flavovirens N. E. Br. Amorphophallus johnsonii N. E. Br. Ampelocissus bombycina Planch. Ampelocissus leonensis (Hook. f.) Planch. Ampelocissus multistriata (Baker) Planch. syn. A. pentaphylla (Guill. & Perr.) Gilg. & Brandt Anacardium occidentale L. Anchomanes difformis (Blume) Engl. Anchomanes welwitschii Rendle 193 Riparian forests plant species of Benin 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 Lmph mph He He He He He He He He He Lnph Lmph nph mPh Ge Th mPh mPh mPh MPh mph mph mph LGe LGe Lnph Ge(Fern) Lnph Th Th Th Th Ch Ge(Fern) Th Th He mPh mph He Lnph Th mPh Th Th Ch mPh mph Th Th Th Th Ge(Fern) mPh mPh He Th He Th GC SZ SG SG SG GC _ SG GC TA GC Pan G PRA PRA _ PRA TA G GC GC GC TA GC SG AA GC TA SZ TA GC G G SG GC G Pan Pan Paleo SZ Pan SG TA TA Pan Pan TA Pan GC SG SG Pan Cosmo PRA SZ SZ TA Paleo SZ Paleo Apocynaceae Ancylobotrys scandens (Schumach. & Thonn.) Pichon Papilionaceae Andira inermis (Wright) DC. subsp. rooseveltii (De Wild.) Gillett ex Polhill Poaceae Andropogon gayanus Kunth. var. bisquamulatus (Hochst.) Hack. Poaceae Andropogon gayanus Kunth. var. gayanus Poaceae Andropogon gayanus Kunth. var. polycladus (H.) W.D. Clayton syn. A. gayanus var. squamulatus Poaceae Andropogon macrophyllus Stapf. Poaceae Andropogon sp. Poaceae Andropogon tectorum Schum. & Thonn. Commelinaceae Aneilema beninense (P. Beauv.) Kunth Commelinaceae Aneilema dispermum Brenan Commelinaceae Aneilema umbrosum (Vahl) Kunth Convolvulaceae Aniseia martinicensis (Jacq.) Choisy Asclepiadaceae Anisopus manii N.E.Br. Annonaceae Annona senegalensis Pers. Combretaceae Anogeissus leiocarpa (DC.) Guill. & Perr. syn. A. leiocarpus (DC.) Guill. & Perr. Liliaceae Anthericum sp. Melastomataceae Antherotoma naudini Hook. F. Loganiaceae Anthocleista djalonensis A. Chev. Loganiaceae Anthocleista nobilis G. Don Loganiaceae Anthocleista vogelii Planch. Moraceae Antiaris toxicaria Lesch. var africana A. Chev. syn. A. africana Engl. Euphorbiaceae Antidesma membranaceum Müll. Arg. Euphorbiaceae Antidesma venosum Tul. Sapindaceae Aphania senegalensis (Juss. ex Poir.) Radlk. syn. Lepisanthes senegalensis (Juss. ex Poir.) Leenh Aristolochiaceae Aristolochia albida Duchartre Aristolochiaceae Aristolochia ringens Vahl. Annonaceae Artabotrys velutinus Sc. Elliot Davalliaceae Artopteris orientalis (Gmel.) Posth. Liliaceae Asparagus africanus Lams Asteraceae Aspilia africana (Pers.) C.D. Adams Asteraceae Aspilia angustifolia Oliv. &Hiern. Asteraceae Aspilia bussei (Schum. & Thonn.) Oliv. & Hiern Asteraceae Aspilia paludosa Berhaut Asteraceae Aspilia rudis C. D. Adams subsp. fontinaloides Adams Aspleniaceae Asplenium diplazorum Hieron. Acanthaceae Asystasia calycina Benth. Acanthaceae Asystasia gangetica (L.) T. Anders Poaceae Axonopus flexuosus (Peter) C.E. Hubbard ex Troupin Meliaceae Azadirachta indica A. Juss. Balanitaceae Balanites aegyptiaca (L.) Del. Poaceae Bambusa vulgaris Schrader ex Wendel. Acanthaceae Barleria oenotheroides Dum. Lamiaceae Basilicum ploystachion (L.) Moench. Caesalpiniaceae Berlinia grandiflora (Vahl) Hutch. & Dalz. Asteraceae Bidens pilosa L. Oxalidaceae Biophytum umbraculum Welw. syn. B. petersianum Klotszch Acanthaceae Blepharis maderaspatensis (L.) Heyne ex Roth. Sapindaceae Blighia sapida Koenig Sapindaceae Blighia unijugata Baker Asteraceae Blumea viscosa (Mill.) Badillo syn. B. aurita DC. Nyctaginaceae Boerhavia coccinea Mill. Nyctaginaceae Boerhavia diffusa L. Nyctaginaceae Boerhavia erecta L. Lomariopsidaceae Bolbitis heudelotii (Bory ex Fee) Aston Bombacaceae Bombax costatum Pellegr. & Vuill. Arecaceae Borassus aethiopum Mart. Poaceae Brachiaria brizantha (Hochst. ex A. Rich.) Stapf Poaceae Brachiaria deflexa (Schumach.) Robyns Poaceae Brachiaria jubata (Figari & De Notaris) Stapf. Poaceae Brachiaria lata (Schumach.) C.E. Hubbard 194 Riparian forests plant species of Benin 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 He Th mph mph mPh mph mPh Th Ge mph nph nph nph Ep(fern) Ep(fern) nph nph mph Lmph Lmph Th Th Lnph Lnph Lnph Lnph mph Th nph Lnph Lmph Lnph Th Lnph Lnph mph mph Ch mPh Par MPh Th LTh LTh mPh Th Lnph Th Ge Ge Ge nph Lnph Th nph mph He He He He _ SZ GC SZ TA SC SZ Pan _ SZ Paleo Paleo GC GC GC GC SG GC GC Pan Pan Pan SG GC SG GC TA GC S SZ G G Pan AA Pan Pan SZ _ GC Pan Pan Paleo GC Paleo SZ G AA TA G _ G Pan GC SZ GC SZ PRA GC GC PRA Poaceae Poaceae Euphorbiaceae Euphorbiaceae Euphorbiaceae Euphorbiaceae Rubiaceae Cyperaceae Cyperaceae Caesalpiniaceae Capparidaceae Papilionaceae Apocynaceae Orchidaceae Orchidaceae Ochnaceae Ochnaceae Ochnaceae Celastraceae Papilionaceae Gentinaceae Gentinaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Euphorbiaceae Capparidaceae Capparidaceae Capparidaceae Capparidaceae Solanaceae Sapindaceae Sapindaceae Caricaceae Caesalpiniaceae Caesalpiniaceae Rhizophoraceae Lauraceae Bombacaceae Amaranthaceae Amaranthaceae Amaranthaceae Ulmaceae Poaceae Papilionaceae Adiantaceae Asclepiadaceae Asclepiadaceae Asclepiadaceae Caesalpiniaceae Menispermaceae Poaceae Rubiaceae Oleaceae Poaceae Liliaceae Liliaceae Liliaceae Brachiaria sp. Brachiaria villosa (Lam.) A. camus syn. B. distichophylla (Trin.) Stapf. Bridelia atroviridis Müll. Arg. Bridelia ferruginea Benth. Bridelia micrantha (Hochst.) Baill. Bridelia scleroneura Müll. Arg. Brenadia salicina (Vahl) Hepper & Wood syn. Adina microcephala (Del.) Hiern Bulbostylis barbata (Rottb.) C.B.Cl. Bulbostylis sp. Burkea africana Hook. Cadaba farinosa Forsk. Cajanus cajan (L.) Millsp. Callichilia barteri (Hook. f.) Stapf syn. Hedranthera brateri (Hook. f.) pichon Calyptrochilum christyanum (Rchb. f.) Summerh. Calyptrochilum emarginatum (Sw.) Schltr. Campylospermum flavum (Schumach.&Thonn. ex Stapf) Farron syn. Ouratea flava Campylospermum glaberrimum (P. Beauv.) Farron syn. Ouratea glaberrima Campylospermum reticulatum (P.Beauv) Farron var. reticulatus syn. Ouratea reticulata Campylostemon warneckeanum Loes. ex Fritsch Canavalia ensiformis (L.) DC. Canscora decussata (Roxb.) Roem. & Schult. Canscora diffusa (Vahl) R. Br. ex Roem. & Schult Canthium cornelia Cham. & Schlecht. Canthium henriquesianum (K. Schum.) G. Tayl. Canthium horizontale (Schum. & Thonn.) Hiern Canthium setosum Hiern Canthium vulgare (K. Schum.) Bullock. F.R. Caperonia senegalensis Müll. Arg. Capparis decidua (Forsk.) Edgew. Capparis sepiaria L. var. fischeri (Pax) De Wolf syn. C. corymbosa Lam. Capparis thonningii Schum. syn. C. brassii DC. Capparis viminea Hook. f. & Th. ex. Oliv. var. viminea Capsicum frutescens L. Cardiospermum grandiflorum Swartz Cardiospermum halicacabum L. Carica papaya L. Cassia sieberiana DC. Cassia sp. Cassipourea congoensis R. Br. ex DC. Cassytha filiformis Linn. Ceiba pentandra (L.) Gaertn. Celosia argentea L. Celosia isertii C. Towns. syn. C. laxa Schum. & Thonn. Celosia trigyna L. Celtis toka (Forssk) Hepper & Wood syn. C. integrifolia Centosteca latifolia (Osb.) Trin. syn. C. lappacea (L.) Desv. Centrosema pubescens Benth. Ceratopterix cornuta (P. Beauv.) Lepr. Ceropegia fusiformis N. E. Br. Ceropegia sp. Ceropegia yorubana Schlechter Chamaecrista mimosoides (L.) Greene syn. Cassia mimosoides L. Chasmanthera dependens Hochst. Chasmopodium caudatum (Hack.) Stapf Chassalia kolly (Schumach.) Hepper Chionanthus niloticus (Oliv.) Stearn syn. Linociera nilotica Oliv. Chloris gayana Kunth Chlorophytum alismifolium Baker Chlorophytum andongense Baker Chlorophytum blepharophyllum Schweinf. ex. Baker 195 Riparian forests plant species of Benin 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 He Ge Ge Ch mPh Lnph Lnph Lnph Lnph Lnph Lnph Lnph Lnph Lnph Lmph LTh LGe LTh mph nph mPh Th Th Lnph Lnph Lnph Lnph nph MPh mPh mph Lmph mph mph mph mph Lnph mph LmPh mph LmPh Lmph Lnph He He He He Th He He Lmph Th Th Th mph mph nph Ge mph Lmph Liliaceae Chlorophytum macrophyllum (A. Rich.) Aschers. GC Liliaceae Chlorophytum sp. _ Liliaceae Chlorophytum togoense Engl. SG Asteraceae Chromolaena odorata (L.) R.M.King & H. Robinson Pan Chrysobalanaceae Chrysobalanus icaco L. subsp. atacorensis SC Menispermaceae Cissampelos mucronata A. Rich. SZ Menispermaceae Cissampelos owariensis P. Beauv. ex DC. GC Vitaceae Cissus aralioides (Welw. ex Baker) Planch. TA Cissus cymosa Schum. & Thonn. PRA Vitaceae Vitaceae Cissus glaucophylla Hook. G Vitaceae Cissus gracilis Guill. & Perr. TA Vitaceae Cissus kouandeensis A. Chev. S Vitaceae Cissus palmatifida (Baker) Planch. SG Cissus petiolata Hook. f. PRA Vitaceae Cissus populnea Guill. & Perr. var. populnea PRA Vitaceae Vitaceae Cissus rubiginosa (Welw. ex Baker) Planch TA Cissus rufescens Guill. & Perr. PRA Vitaceae Vitaceae Cissus sp. _ Aurantiaceae Citrus limonum Risso Pan Rutaceae Clausena anisata (Willd.) Benth. TA Annonaceae Cleistopholis patens (Benth.) Engl. & Diels GC Cappariadaceae Cleome rutidosperma DC. syn. C. ciliata Schum. & Thonn. S Cappariadaceae Cleome viscosa L. SG Clerodendrum capitatum (Willd.) Schum. & Thonn. PRA Verbenaceae Verbenaceae Clerodendrum dusenii Gurke GC Verbenaceae Clerodendrum polycephalum Baker GC Coccinia grandis L. Paleo Cucurbitaceae Cochlospermaceae Cochlospermum planchonii Hook. f. SG Sterculiaceae Cola gigantea A. Chev. GC Sterculiaceae Cola laurifolia Mast. GC Sterculiaceae Cola millenii K. Schum. G Combretaceae Combretum acutum Laws. S Combretaceae Combretum adenogonium Steud. ex A. Rich. syn. C.ghasalense Engl. & Diels SG Combretaceae Combretum collinum Fresen subsp. collinum SG Combretaceae Combretum collinum Fresen subsp. hypopilinum S Combretaceae Combretum glutinosum Perr. ex DC. SZ Combretaceae Combretum lecardii Engl. & Diels G Combretaceae Combretum molle R. Br. ex G. Don SZ Combretaceae Combretum mucronatum Schum. & Thonn. syn. C. smeathmannii G. Don GC Combretaceae Combretum nigricans Lepr. ex Guill. & Perr. SG Combretaceae Combretum paniculatum Vent. TA Combretaceae Combretum racemosum P. Beauv. GC Combretaceae Combretum tomentosum G. Don G Commelinaceae Commelina benghalensis L. var. bengalensis GC Commelinaceae Commelina diffusa Bum. f. subsp. diffusa Pan Cosmo Commelinaceae Commelina erecta L. subsp. erecta Commelinaceae Commelina erecta subsp livingstonei (C.B. Cl.) J. K. Morton TA PRA Commelinaceae Commelina nigritana Benth. Commelinaceae Commelina sp. _ Commelinaceae Commelina thomasii Hutch. G Connaraceae Connarus africanus Lam. SG Tiliaceae Corchorus aestuans L. Pan Tiliaceae Corchorus fascicularis Lam. Pan Tiliaceae Corchorus olitorius L. Pan Boraginaceae Cordia africana Lam. TA Cordia myxa L. Paleo Boraginaceae Boraginaceae Cordia senegalensis Juss. GC Zingiberaceae Costus afer Ker-Gawl. TA Crateva adansonii DC. subsp. adansonii syn. C. religiosa auct. Afr. Fl. non Forst. f. Paleo Cappariadaceae Rubiaceae Cremaspora triflora (Thonn.) K. Schum. TA 196 Riparian forests plant species of Benin 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 Ge Ge Ge mph Th Ch Th Th Th Ch Ch Ch Th Th Lmph Th Lnph Lmph Ge mph nph Th Ch mPh Ge Ge Ge Ge Th LGe Th Lmph Lmph Lmph Lmph LmPh mPh mPh mph Ch Ch nph Th Th Ch Th LTh Lnph Ch mph mPh mPh mph nph Th Th Th Th Th Lnph SG GC SG SZ Pan TA SC S PRA PRA Pan Pan _ Pan SG Pan PRA G PRA SG GC Pan GC GC Pan Pan Pan _ Pan _ Pan GC PRA G GC GC SG Pan G AA AA GC TA PRA GC _ _ Pan Pan SZ GC GC G TA TA TA Pan Pan _ _ Amaryllidaceae Amaryllidaceae Amaryllidaceae Rubiaceae Papilionaceae Papilionaceae Papilionaceae Papilionaceae Papilionaceae Papilionaceae Papilionaceae Papilionaceae Papilionaceae Euphorbiaceae Asclepiadaceae Poaceae Cucurbitaceae Araceae Hypoxidaceae Araliaceae Amaranthaceae Amaranthaceae Thelypteridaceae Caesalpiniaceae Cyperaceae Cyperaceae Cyperaceae Cyperaceae Cyperaceae Vitaceae Poaceae Papilionaceae Papilionaceae Papilionaceae Papilionaceae Papilionaceae Caesalpiniaceae Caesalpiniaceae Annonaceae Papilionaceae Papilionaceae Papilionaceae Papilionaceae Papilionaceae Papilionaceae Papilionaceae Papilionaceae Papilionaceae Papilionaceae Caesalpiniaceae Caesalpiniaceae Caesalpiniaceae Dichapetalaceae Mimosaceae Poaceae Poaceae Poaceae Poaceae Poaceae Dilleniaceae Crinum glaucum A. Chev. Crinum jagus (Thomps.) Dandy Crinum zeylanicum (L.) L. syn. C. ornatum (Alt.) Bury Crossopteryx febrifuga (G. Don) Benth. Crotalaria calycina Schrank Crotalaria cephalotes Steud. ex A. Rich. Crotalaria comosa Baker Crotalaria deightonii Hepper Crotalaria hyssopifolia Klotszch Crotalaria lachnosema Stapf Crotalaria pallida var. mucronata Desv. Crotalaria retusa L. Crotalaria sp. Croton lobatus L. Cryptolepis sanguinolenta (Lind.) Schltr. Ctenium newtonii Hack Cucumis metuliferus E. Mey. ex Naudin Culcasia scandens P. Beauv. Curculigo pilosa (Schum. & Thonn.) Engl. Cussonia arborea Hochst. ex A. Rich. syn. C. barteri Seem. & C. kirkii Seem. Cyathula achyranthoides (H. B. et K.) Moq Cyathula prostrata (L.) Blume Cyclosorus striatus Cop. Cynometra megalophylla Harms Cyperus difformis L. Cyperus distans L.f. Cyperus haspan L. Cyperus sp. Cyperus sphacelatus Rottb. Cyphostema sp. Dactyloctenium aegyptium (L.) P.Beauv Dalbergia dalzielii Baker f. ex. Hutch. & Dalz. Dalbergia lactea Vatke Dalbergia rufa G. Don Dalbergia saxatilis Hook. f. var. saxatilis Dalbergiella welwitschii (Baker) Baker f. Daniellia oliveri (Rolfe) Hutch. & Dalz. Delonix regia (Boj. ex Hook.) Raf. Dennettia tripetala Baker f. Desmodium adscendens (Sw.) DC var. adscendens Desmodium adscendens (Sw.) DC var. robustrum Schubert Desmodium gangeticum (L.) DC. Desmodium hirtum Guill. & Perr. Desmodium ramosissimum G. Don Desmodium salicifolium (Poir.) DC. var. salicifolium Desmodium sp1. Desmodium sp2. Desmodium triflorum (L.) DC. Desmodium velutinum (Willd.) DC. Detarium microcarpum Harms Detarium senegalense J. F. Gmelin Dialium guineense Willd. Dichapetalum madagascariense Poir var. madagascariense syn. D. guineense Dichrostachys cinerea (L.) Wight & Arn Digitaria argyrotricha (Anderss.) Chiov. Digitaria ciliaris (Retz.) Koel. Digitaria gayana (Kunth) Stapf ex A. Chev. Digitaria horizontalis Willd. ssp. proximus (Hocht. ex A.Rich.) Maire & Weiler Digitaria sp. Dillenia sp. 197 Riparian forests plant species of Benin 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 Lmph LTh LGe LGe LGe LGe LGe LGe LGe LGe LGe LGe LGe LGe Lnph mPh mPh mph Ge(fern) Ch Ge Ge Ge Ch mPh mph mph mph Ch Th Th Th mPh mPh Ch Th Th He He Th Th mph mph Lmph He Th Th nph mPh Ch Ch mPh mph mPh mPh mph mph Th Ge Ge Pan Pan SG Pan SZ SZ SG GC SG GC SG GC _ GC _ GC SZ GC GC GC SG S S _ GC TA G GC G TA Pan Pan SZ GC Pan SG Pan SZ GC SG PRA TA PRA GC Paleo _ Paleo GC GC PRA Pan GC SG GC GC TA GC Pan PRA PRA Papilionaceae Rubiaceae Dioscoraceae Dioscoraceae Dioscoraceae Dioscoraceae Dioscoraceae Dioscoraceae Dioscoraceae Dioscoraceae Dioscoraceae Dioscoraceae Dioscoraceae Dioscoraceae Menispermaceae Ebenaceae Ebenaceae Ebenaceae Athyriaceae Umbelliferae Melastomataceae Melastomataceae Melastomataceae Melastomataceae Sterculiaceae Sterculiaceae Agavaceae Euphorbiaceae Acanthaceae Acanthaceae Poaceae Asteraceae Meliaceae Arecaceae Asteraceae Asteraceae Poaceae Poaceae Acanthaceae Lamiaceae Lamiaceae Mimosaceae Mimosaceae Mimosaceae Poaceae Poaceae Poaceae Acanthaceae Sapindaceae Papilionaceae Papilionaceae Papilionaceae Papilionaceae Caesalpiniaceae Caesalpiniaceae Erytroxylaceae Capparidaceae Poaceae Orchidaceae Orchiadaceae Dioclea reflexa Hook. f. Diodia sarmentosa Sw. syn. D. scandens auctt Dioscorea abyssinica Hochst. ex Kunth. Dioscorea bulbifera L. var. bulbifera Dioscorea dumetorum (Kunth) Pax Dioscorea hirtiflora Benth. subsp. hirtiflora Dioscorea lecardii De Wild. Dioscorea multiflora Pax syn. D. minutiflora Engl. Dioscorea odoratissima Pax syn. D. praehensilis Benth. Dioscorea preussii Pax Dioscorea sagittifolia Pax Dioscorea sansibarensis Pax Dioscorea sp. Dioscorea togoensis Kunth syn. D. callei A. Chev. ex De Wild. Dioscoreophyllum sp. Diospyros abyssinica (Hiern) White Diospyros mespiliformis Hochst. ex A. DC. Diospyros monbuttensis Gürke Diplazium sammatii (Kuhn) C. Chr. Diplolophium africanum Turcz. Dissotis anthennima (Sm.) Triana Dissotis grandiflora (Sm.) Benth. var. lambii (Hutch.) Keay Dissotis irvingiana Hook. syn. D. senegambiensis (Guill. & Rev.) triana Dissotis sp. Dombeya ledermanii Engl. Dombeya quinqueseta (Del.) Exell Dracaena arborea Baker Drypetes floribunda (Müll. Arg.) Hutch. Dyschoriste heudelotiana (Nees) O. Ktze. Dyschoriste perrottetii (Nees) O. Kuntze Echinochloa colona (L.) Link. syn. E. colonum (L.) Link. Eclipta prostata (L.) L. Ekebergia capensis Sparrm. syn. E. senegalensis A. Juss. Elaeis guineensis Jacq. Elephantopus mollis Kunth Elephantopus senegalensis (Klatt) Oliv. & Hiern Eleusine indica Gaertn. Elymandra androphila (Stapf) Stapf Elytraria marginata Vahl Englerastrum gracillimum Th. C. E. Fries Englerastrum schweinfurthii Briq. Entada abyssinica Steud. ex A. Rich. Entada africana Guill. & Perr. Entada mannii (Oliv.) Tisserant Eragrostis atrovirens (Desf.) Trin. ex Steud. (Hochst. ex A. Rich.) Maire & Weiler Eragrostis sp. Eragrostis tremula Horst. & Steud. Eremomastax speciosa (Hochst.) Cuf. Eriocoelum kerstingii Gilg. ex Engl. var. kerstingii Eriosema glomeratum (Guill & Perr) Hook.f. syn. Rhynchosia glomerata Guill. & Perr. Eriosema psoraleoides (Lam.) G. Don syn. Crotaliaria psoralioides Lam. Erythrina excelsa Baker Erythrina senegalensis A. DC. Erythrophleum africanum (Wel. ex Benth.) Harms Erythrophleum suaveolens (Guill. & Pierr.) Brenan syn. E. guineense G. Don Erythroxylum emarginatum Thonn. Euadenia trifoliolata (Schum. & Thonn.) Oliv. Euclasta condylotricha (Steud.) Stapf (Hochst. ex A. Rich.) Maire & Weiler Eulophia guineensis Lindl. Eulophia quartiniana A. Rich. 198 Riparian forests plant species of Benin 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 Th Th Th Th Th Th Th Th Th Ch nph nph mph mPh mph mph mPh mPh mPh mph mPh mph MPh mPh mph mph mph mPh mph mph mPh mph He He Th Lmph mph Lnph He Th nph Th Ch Th Th mPh mPh mph mPh mph nph mph nph nph LTh LTh Th Ch Th LGe AA Pan Paleo AA Pan Pan AA Pan S SZ GC SZ SC GC GC TA SZ SZ S GC SC SG GC _ SG SZ TA G TA GC GC Pan SG _ GC GC GC GC SZ Pan TA TA _ PRA GC GC GC SZ SZ SG GC _ SG GC Pan Paleo _ Pan GC Asteraceae Asteraceae Euphorbiaceae Euphorbiaceae Euphorbiaceae Euphorbiaceae Euphorbiaceae Euphorbiaceae Euphorbiaceae Euphorbiaceae Rubiaceae Moraceae Moraceae Moraceae Moraceae Moraceae Moraceae Moraceae Moraceae Moraceae Moraceae Moraceae Moraceae Moraceae Moraceae Moraceae Moraceae Moraceae Moraceae Moraceae Moraceae Moraceae Cyperaceae Cyperaceae Cyperaceae Malpighiaceae Flacourtiaceae Flagellariaceae Commelinaceae Commelinaceae Euphorbiaceae Urticaceae Urticaceae Cyperaceae Cyperaceae Apocynaceae Apocynaceae Rubiaceae Clusiaceae Clusiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Molluginaceae Iridaceae Molluginaceae Liliaceae Eupatorium microstemon Cass. Eupatorium triplinerve Vahl. Euphorbia aegyptiaca Boiss. Euphorbia glomerifera (Millesp.) Wheeler Euphorbia heterophylla L. Euphorbia hirta L. Euphorbia hyssopifolia L. Euphorbia sp. Euphorbia thymifolia L. Excoecaria grahamii Stapf. syn. Sapium grahamii (Stapf) Prain Feretia apodanthera Del. Ficus asperifolia Miq. Ficus capreifolia Del. Ficus congensis Engl. syn. Ficus trichopoda Baker Ficus cordata Thunb. Warb. Ficus dicranostyla Mildbr. Ficus exasperata Vahl Ficus glumosa Del. Ficus ingens (Miq.) Miq. Ficus kerstingii Hutch. syn. Ficus abutilifolia (Miq.) Miq. Ficus lyrata Warb. Ficus ovata Vahl Ficus platyphylla Del. Ficus polita Vahl Ficus sp. Ficus sur Forssk. syn. Ficus capensis Thunb. Ficus sycomorus L. subsp. gnaphalocarpa (Miq.) Berg Ficus sycomorus L. var psychomorus Ficus tesselata Warb. Ficus thonningii Blume syn. F. basarensis Mildbr. & Burret Ficus vogeliana (Miq.) Miq. Ficus vogelii (Miq.) Miq. Fimbristylis dichotoma (L.) Vahl subsp. dichotoma Fimbristylis hispidula (Vahl.) Kunth Fimbristylis sp. Flabellaria paniculata Cav. Flacourtia flavescens Willd. Flagellaria guineensis Schumacher Floscopa africana (P. Beauv.) C. B. Clarke Floscopa flavida C. B. Cl. Flueggea virosa (Roxb ex Willd) Voigt subsp.virosa syn. Securinega virosa (Roxb ex Willd) Fluerya aestuans (L.) ex Miq. Fluerya ovalifolia (Schum. & Thonn.) Dandy Fuirena sp. Fuirena stricta Steud. Funtumia africana (Benth.) Stapf Funtumia elastica (Preuss) Stapf Gaertnera paniculata Benth. Garcinia livingstonei T. Anders. Garcinia ovalifolia Oliv. Gardenia erubescens Stapf & Hutch. Gardenia imperialis K. Schum. Gardenia sp. Gardenia ternifolia Schum. & Thonn. syn. G. triacantha DC. Geophila obvallata (Schumach.) F. Didr. Geophila reniformis D. Don syn. G. repens (L.) I.M. Johnston Gisekia pharnacioides L. Gladiolus sp. Glinus oppositifolus (L.) Aug. DC. Gloriosa superba L. syn. G. simplex L. 199 Riparian forests plant species of Benin 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 mph mPh Lmph Ep Lmph nph nph nph Lnph Ge Ge MPh Th Ch Th mPh mph Th Th nph Ch Ch Lmph Ch Ch Ch Ch nph Th Lmph mPh nph Ch Th Th Th mph nph Th He Th Th Th Ge Ge Th Th Th He Th Th Ch Th Lnph Ch Th Ch Th Th Th TA Pan TA PRA GC Paleo SZ Paleo Pan PRA TA GC Pan Paleo SG GC SZ TA TA G Paleo Paleo Pan PRA TA Pan PRA _ Pan GC TA PRA Pan PRA Pan _ SZ PRA S Pan _ GC _ GC GC AA Paleo Pan Pan SZ GC SG GC GC GC PRA TA S _ Pan Tiliaceae Verbenaceae Asclepiadaceae Orchidaceae Tiliaceae Tiliaceae Tiliaceae Tiliaceae Asclepiadaceae Orchidaceae Amaryllidaceae Simaroubaceae Boraginaceae Boraginaceae Compositeae Annonaceae Annonaceae Malvaceae Malvaceae Malvaceae Malvaceae Malvaceae Malvaceae Malvaceae Malvaceae Malvaceae Malvaceae Malvaceae Malvaceae Hippocrateaceaae Apocynaceae Lamiaceae Violaceae Hippocrateaceaae Acanthaceae Acanthaceae Euphorbiaceae Rubiaceae Poaceae Poaceae Poaceae Acanthaceae Acanthaceae Marantaceae Marantaceae Lamiaceae Lamiaceae Lamiaceae Poaceae Papilionaceae Papilionaceae Papilionaceae Papilionaceae Papilionaceae Papilionaceae Papilionaceae Papilionaceae Papilionaceae Papilionaceae Papilionaceae Glyphaea brevis (Spreng.) Monachino Gmelina arborea Roxb. Gongronema latifolium Benth. Graphorkis lurida (Sw.) O. Ktze. Grewia carpinifolia Juss. Grewia flavescens Juss. Grewia mollis Juss. var. mollis syn. G. pubescens P. Beauv. Grewia ternax (Forssk.) Gymnema sylvestre (Retz.) Schultes Habenaria laurantii De Wild. Haementhus multiforus Martyn Hannoa klaineana Pierre & Engl. syn. Quassia undulata (Guill. & Perr.) D. Dietr. Heliotropium indicum L. Heliotropium strigosum Willd. Herderia truncata Cass. Hexalobus crispiflorus A. Rich. Hexalobus monopetalus (A. Rich.) Engl. & Diels var. monopetalus Hibiscus asper Hook. F. Hibiscus esculentus L. Hibiscus grewioides Baker. F. Hibiscus lunariifolis Willd. Hibiscus micranthus L. Hibiscus panduriformis Burm. f. Hibiscus physaloides Guill. & Perr. Hibiscus rostellatus Guill. & Perr. Hibiscus sabdariffa L. Hibiscus sidiformis Baill. Hibiscus sp. Hibiscus surratensis L. Hippocratea welwitchii Oliv. syn. Simirestis welwitchii (Oliv.) Halle Holarrhena floribunda (G. Don) Dur. & Schinz Hoslundia opposita Vahl Hybanthus enneaspermus (L.) F. Muell. Hydrolea glabra Schum. & Thonn. Hygrophila auriculata (Schumach.) Heine Hygrophila sp. Hymenocardia acida Tul. var. acida Hymenodiction floribundum (Steud. & Hochst.) B.L.Rob. Hyparrhenia involucrata Stapf Hyparrhenia rufa (Nees) Stapf Hyparrhenia sp. Hypoestes cancellata Nees Hypoestes sp. Hypselodelphys violacea (Ridley) Milne. Redh. Hypselodelphys poggeana (K. Schum.) Milne-Redh. Hyptis lanceolata Poir Hyptis spicigera Lam. Hyptis suaveolens Poit. Imperata cylindrica (L.) Raeuschel Indigofera dendroides Jacq. Indigofera geminata Baker Indigofera hirsuta L. var hirsuta Indigofera longicalyx Gillett Indigofera macrophylla Schum. Indigofera nigritana Hook. f. Indigofera polysphaera Baker Indigofera rhynchocarpa Baker Indigofera secundiflora Poir. Indigofera sp. Indigofera subulata Vahl ex Poir. 200 Riparian forests plant species of Benin 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 Ch Lnph LGe Lnph Lnph Lnph Lnph Lnph LGe Lnph Lnph mPh mPh mPh mPh Ch mPh mph Lmph Lmph Lnph Lnph Lmph Th Ge Lmph Lnph Lmph MPh mPh mPh Ch Ge Ge Ge Ge Th Th LmPh LMPh Lmph Lmph mPh mPh mPh mPh mph Ch He Lnph Lmph Lnph Th LmPh Lmph mph mph Th Th Ge TA S Pan SG Pan Pan TA Pan Pan PRA _ GC PRA SZ SZ G GC PRA PRA SG TA G GC TA TA GC PRA GC GC SZ GC PRA PRA AA Cosmo _ G Pan GC PRA _ G PRA PRA SZ G GC TA Pan TA Paleo PRA TA GC _ Pan GC GC _ TA Papilionaceae Convolvulaceae Convolvulaceae Convolvulaceae Convolvulaceae Convolvulaceae Convolvulaceae Convolvulaceae Convolvulaceae Convolvulaceae Convolvulaceae Simaroubaceae Simaroubaceae Caesalpiniaceae Caesalpiniaceae Lamiaceae Annonaceae Rubiaceae Oleaceae Oleaceae Oleaceae Oleaceae Connaraceae Acanthaceae Zingiberaceae Rubiaceae Rubiaceae Rubiaceae Meliaceae Meliaceae Bignoniaceae Asteraceae Cyperaceae Cyperaceae Cyperaceae Cyperaceae Compositeae Asteraceae Apocynaceae Apocynaceae Apocynaceae Apocynaceae Anacardiaceae Anacardiaceae Anacardiaceae Anacardiaceae Sapindaceae Leeaceae Poaceae Convolvulaceae Asclepiadaceae Asclepiadaceae Poaceae Papilionaceae Papilionaceae Mimosaceae Gentianaceae Scrophulariaceae Scrophulariaceae Cyperaceae Indigofera trita L.f. Ipomoea acanthocarpa (Hochst. ex Choisy) Aschers. & Schweinf. Ipomoea alba L. Ipomoea argentaurata Forssk. Ipomoea asarifolia (Desr.) Roen. & Schult. Ipomoea eriocarpa R. Br. Ipomoea involucrata P. Beauv. Ipomoea mauritiana Hall. f. Ipomoea muricata (L.) Jacq. Ipomoea rubens Choisy Ipomoea sp. Irvingia gabonensis (Aubry-Lecomte ex O'Rorke) Baill. Irvingia smithii Hook. F. Isoberlinia doka Craib & Stapf. Isoberlinia tomentosa (Harms) Craib & Stapf. Isodictyophorus reticularus (A. Chev.) J. K. Morton Isolona thonneri (De wild. Th. Dur.) Engl. & Diels Ixora brachypoda DC. Jasminum dichotomum Vahl Jasminum obtusifolium Baker Jasminum pauciflorum Benth. Jasminum preussii Engl. & Knobl. Jaundea pinnata (P. Beauv.) Schellenb. Justicia anselliana (Nees) T. Anderson Kaempferia aethiopica (Solms-Laub.) Benth. Keetia hispida (Benth.) Bridson Keetia multiflora (Schum.& Thonn.) Bridson syn. Canthium multiflorum (S.&T.) Hiern Keetia venosa (Oliv.) Bridson syn. Canthium venosum (Oliv.) Hiern Khaya grandifoliola C. DC. Khaya senegalensis (Desv.) A. Juss. Kigelia africana (Lam.) Benth. Kinghamia macrocephala (Oliv. & Hiern.) Jeffrey Kyllinga erecta Schumach. var. erecta Kyllinga odorata Vahl Kyllinga pumila Michx. Kyllinga sp. Laggera alata (D.Don) Sch. Bip. ex Oliv. Laggera aurita (L.f.) Benth ex C.B. Clarke syn. Blumea aurita (L.f.) DC. Landolphia hirsuta (Hua) Pichon Landolphia owariensis P. Beauv. Landolphia sp. Landolphia togolana (Hallier f.) Pichon Lannea acida A. Rich. Lannea kerstingii Engl. & K. Krause syn. Lannea barteri (Oliv.) Engl. Lannea microcarpa Engl. & K. Krause Lannea nigritana (Sc. Elliot) Keay var. nigritana Lecaniodiscus cupanioides Planch. Leea guineensis G. Don Leersia hexandra Sw. Lepistemon owariense (P. Beauv.) Hall. f. Leptadenia arborea (Forsk.) Schweinf. Leptadenia hastata (Pers.) Decne. Leptochloa caerulescens Steud. Leptoderris brachyptera (Benth.) Dunn. Leptoderris sp. Leucaena leucocephala Benth. Lindackeria dentata (Oliv.) Gilg. Lindernia diffusa (L.) Wettst. Lindernia sp. Lipocarpha albiceps Ridl. 201 Riparian forests plant species of Benin 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 Ge Ge nph Th LmPh LmPh LmPh mph mPh mph He He Th He He Th Th Th Th Ch Lnph Th He(Fern) nph mph mph mPh Ge mPh mPh mph mph mPh Ge Ge mph He Th Ge Ge He He mph Th mph mph He Th Ch Ch Th mph Lnph Lnph Ge Lmph mph Ch Lmph MPh TA TA SG PRA TA _ GC PRA PRA GC TA _ S TA _ PRA TA TA _ TA Pan Cosmo Pan SG TA PRA Pan Pan G TA SG SG G G _ TA Pan Paleo AA TA GC _ GC PRA SZ PRA GC TA G Pan _ G AA Pan _ G GC Paleo Pan G Cyperaceae Cyperaceae Verbenaceae Campanulaceae Hippocrateaceae Hippocrateaceae Papilionaceae Papilionaceae Papilionaceae Ochnaceae Poaceae Poaceae Poaceae Poaceae Poaceae Onagraceae Onagraceae Onagraceae Onagraceae Onagraceae Cucurbitaceae Solanaceae Lycopodiaceae Rubiaceae Capparidaceae Euphorbiaceae Anacardiaceae Euphorbiaceae Euphorbiaceae Sapotaceae Chrysobalanaceae Chrysobalanaceae Chrysobalanaceae Maranthaceae Maranthaceae Euphorbiaceae Cyperaceae Cyperaceae Cyperaceae Cyperaceae Cyperaceae Cyperaceae Bignoniaceae Marsileaceae Celastraceae Celastraceae Marantaceae Asteraceae Melastomataceae Sterculiaceae Sterculiaceae Melastomataceae Convolvulaceae Convolvulaceae Eriocaulaceae Caesalpiniaceae Euphorbiaceae Asteraceae Asteraceae Moraceae Lipocarpha atra Ridl. Lipocarpha prieuriana Steud. Lippia multiflora Moldenke Lobelia sapinii De Wild. Loeseneriella africana Willd. var. africana syn. Hippocratea africana (Willd.) Loes. Loeseneriella sp. Lonchocarpus cyanescens (Schum. & Thonn.) Benth. Lonchocarpus laxiflorus Guill. & Perr. syn. Philenoptera laxiflora (Guill. & Perr.) G. Roberty Lonchocarpus sericeus (Poir.) H. B. & K. Lophira lanceolata Van Tiegh. ex Keay Loudetia phragmitoides (Peter) C. E. Hubb. Loudetia sp. Loudetia togoensis (Pilg.) C. E. Hubb. Loudetiopsis ambiens (K. Schum.) Conert Loudetiopsis sp. Ludwigia abyssinica A. Rich. syn. Jussiaea abyssinica (A. Rich.) Dandy & Brenan Ludwigia hyssopifolia (G. Don) Exell syn. Jussiaea linifolia Vahl Ludwigia octovalvis (Jacq.) Raven Ludwigia sp. Ludwigia stenorraphe (Brenan) Hara subsp. stenorraphe Luffa cylindrica (L.) M. J. Roem syn. Luffa aegyptiaca Mill Lycopersicon esculentum Mill syn. Solanum lycopersicon L. Lycopodium cernuum L. Macrosphyra longistyla (DC.) Hiern Maerua angolensis DC. Mallotus oppositifolius (Geisel.) Müell. Arg. var. oppositifolius Mangifera indica L. Manihot esculenta Crantz Manihot glaziovii Müll. Arg. Manilkara obovata (Sabine & G. Don) syn. M. multinervis (Baker) Dubard Maranthes kerstingii (Engl.) Prance Maranthes polyandra (Benth.) Prance Maranthes robusta (Oliv.) Prance syn. Parinari robusta Oliv. Maranthochloa purpurea (Ridley) Milne. Redh. Maranthochloa sp. Margaritaria discoidea (Baill.) Webster Mariscus cylindristachyus Steud. syn. M. alternifolius auct. Mariscus dubius (Rottb.) C.E.C. Fisch. Mariscus flabelliformis Kunth Mariscus longibracteatus Cherm. Mariscus soyauxii (Boech.) C.B. Clarke Mariscus sp. Markhamia tomentosa (Benth.) K. Schum. ex Engl. Marsilea crenulata Desv. Maytenus heterophylla (Eckl. & Zeyh.) Robson Maytenus undatus (Thunb.) Blakelock Megaphrynium macrostachyum (Benth.) Milne-Redh. Melanthera scandens (Schum. & Thonn.) Rob. Melastomastrum segregatum (Benth) A.&R.Fern. syn. Dissotis segregata (Benth) H.f. Melochia corchorifolia Linn. Melochia sp. Memecylon afzelii G. Don var. afzelii Merremia cissoides (Lam.) Hallier f. Merremia hederacea Burm. f. Mesanthemum sp. Mezoneuron benthanianum Baill. Microdesmis puberula Hook. f. ex Planch. Microglossa pyrifolia (Lam.) Kuntze Mikania cordata (Burm. f.) B. L. Robins var. cordata syn. M. Chenopodiifolia Willd. Milicia exelsa (Welw.) Berg syn. Chlorophora excelsa (Welw.) benth. 202 Riparian forests plant species of Benin 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 Lmph mPh Lmph nph mPh mph mPh Th Lnph Lnph Lnph Lnph Lnph Ch Ch mph mph mph mph mph LmPh Lmph LmPh Lmph LTh He He Lmph Lmph Th mph Th LTh mph Ge Ge Ge Ge mPh mPh Ge mph mph mph mph mph Ch Lmph Th Th Th Th Th Lmph Lmph mph LmPh Th Th Ch GC G G Pan GC SZ SZ Paleo Pan G GC G G TA SZ G SZ SZ SG G GC TA Pan Pan Paleo TA Pan GC GC _ G Pan GC G Pan _ PRA PRA GC G Pan G GC SG SZ _ Paleo GC Pan Paleo AA _ GC G G SZ SZ SG SG TA Papilionaceae Papilionaceae Papilionaceae Mimosaceae Sapotaceae Sapotaceae Rubiaceae Molluginaceae Cucurbitaceae Cucurbitaceae Cucurbitaceae Annonaceae Annonaceae Acanthaceae Acanthaceae Annonaceae Dipterocarpaceae Rubiaceae Rubiaceae Rubiaceae Apocynaceae Papilionaceae Papilionaceae Papilionaceae Cucurbitaceae Commelinaceae Musaceae Rubiaceae Rubiaceae Najadaceae Lecythidaceae Acanthaceae Papilionaceae Annonaceae Davalliaceae Davalliaceae Orchidaceae Orchidaceae Boraginaceae Simaroubaceae Nymphaeaceae Ochnaceae Ochnaceae Ochnaceae Ochnaceae Ochnaceae Lamiaceae Olacaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Poaceae Apocynaceae Apocynaceae Flacourtiaceae Opiliaceae Poaceae Poaceae Papilionaceae Millettia barteri (Benth.) Dunn. Millettia thonningii (schum. & Thonn.) Baker Millettia warneckei Harms Mimosa pigra L. Mimusops andogensis Hiern Mimusops kummel Bruce ex A. DC. Mitragyna inermis (Willd.) O. Kuntze Mollugo nudicaulis Lam. Momordica balsamina L. Momordica charantia L. Momordica cissoides Benth. Monanthotaxis parviflora (Oliv.) Verdc. Monanthotaxis whytei (Stapf) Verdc. Monechma ciliatum (T. Anders.) C. B. Cl. Monechma depauperatum (Jacq.) Milne Monodora tenuifolia Benth. Monotes kerstingii Gilg Morelia senegalensis A. Rich. Morinda geminata DC. Morinda lucida Benth. Motandra guineensis (Thonning) A. DC. Mucuna poggei (L.) DC. var. poggei Mucuna pruriens (L.) DC. var. pruriens Mucuna sloanei Fawc. & Rendle Mukia maderaspatana (L.) M. J. Roem. syn. Melothria maderaspatana (L.) Cogn. Murdannia simplex (Vahl) Brenan Musa sapientum L. Mussaenda elegans Schum. & Thonn. Mussaenda isertiana DC. Naja sp. Napoleonaea vogelii Hook. & Planch. syn. N. leonensis Hutch. & Dalz. Nelsonia canescens (Lam.) Spreng. Neorautanenia mitis (A. Rich.) Verdc. Neostenanthera myristicifolia (Oliv.) Exell Nephrolepis biserrata (Sw.) Schott Nephrolepis sp. Nervilia petraea (Afzel. ex Pers.) Summerch. Nervilia umbrosa (Rchb.f.) Schltr. Newbouldia laevis (P. Beauv.) Seem. ex Bureau Nothospondias staudtii Engl. Nymphaea lotus L. Ochna afzelii R. Br. ex Oliv. Ochna membranacea Oliv. Ochna rhizomatosa (van Tiegh.) Keay Ochna schweinfurthiana F. Hoffm. Ochna sp. Ocimum gratissimum L. Olax subscorpioidea Oliv. var. subscorpioidea Oldenlandia corymbosa L. Oldenlandia herbacea (L.) Roxb. Oldenlandia lancifolia (Schumach.) DC. Oldenlandia sp. Olyra latifolia L. Oncinotis glabrata (Baill.) Stapf ex Hiern Oncinotis nitida Benth. Oncoba spinosa Forssk. Opilia amentacea Roxb. syn. O. celtidifolia (Guill. & Perr) Endl. ex Walp. Oplismenus burmannii (Retz.) P. Beauv. Oplismenus hirtellus (L.) P. Beauv. Ormocarpum sennoides (Willd.) DC. 203 Riparian forests plant species of Benin 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 He He Th mph nph nph nph mph mph Lnph He nph Lnph mph Th Ch Th He Th LGr MPh mph mPh LmPh He Lnph Lmph mph nph nph Th Th Th He mPh Th Lmph mph Th Ch Th Ch Th mph He He Ch Th mph Ch LmPh Th Lnph Th nph Th mPh mph Th Th Pan Pan PRA GC GC GC _ TA GC GC TA PRA Pan SG PRA SZ GC G _ GC GC SZ SZ SG Pan Pan AA SG SZ SG Paleo Pan _ TA GC TA TA SG Paleo PRA TA GC TA TA PRA Paleo G Pan Pan Pan TA Paleo TA _ _ Pan G TA Pan Paleo Poaceae Osmundaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Asclepiadaceae Poaceae Anacardiaceae Papilionaceae Pandanaceae Amaranthaceae Amaranthaceae Poaceae Poaceae Poaceae Aristolochiaceae Chrysobalanaceae Chrysobalanaceae Mimosaceae Asclepiadaceae Poaceae Passifloraceae Sapindaceae Rubiaceae Rubiaceae Rubiaceae Poaceae Poaceae Poaceae Poaceae Clusiaceae Rubiaceae Asclepiadaceae Papilionaceae Poaceae Acanthaceae Acanthaceae Acanthaceae Acanthaceae Arecaceae Poaceae Poaceae Euphorbiaceae Euphorbiaceae Euphorbiaceae Euphorbiaceae Euphorbiaceae Euphorbiaceae Euphorbiaceae Euphorbiaceae Euphorbiaceae Solanaceae Simaroubaceae Caesalpiniaceae Araceae Lamiaceae Oryza sativa L. Osmunda regalis L. Otomeria elatior (A. Rich. ex. DC.) Verdc. Oxyanthus formosus Hook. f. ex Planch. Oxyanthus pallidus Hiern Oxyanthus racemosus (Schum. & Thonn.) Keay Oxyanthus sp. Oxyanthus speciosus DC. Oxyanthus unilocularis Hiern Oxystelma bornouense R. Br. Oxytenanthera abyssinica (A. Rich.) Munro Ozoroa insignis Del. syn. Heeria insignis (del.) O. Ktze. Rev. Pachyrrhizus angulatus Rich. Pandanus candelabrum P. Beauv. Pandiaka angustifolia (Vahl) Hepper Pandiaka involucrata (Moq.) B. D. Jackson Panicum brevifolium L. Panicum maximum Jacq. Panicum sp. Pararistolochia goldieana (Hook. f.) Hutch. & Dalz. Parinari congensis F. Didr. Parinari curatellifolia Planch. ex Benth. Parkia biglobosa (Jacq.) R. Br. ex G. Don f. Parquetina nigrescens (Afzel.) Bullock Paspalum scrobiculatum L. syn. P. orbiculare G. Forest Passiflora foetida L. Paullinia pinnata L. Pavetta corymbosa (DC.) F. N. Williams Pavetta crassipes K. Schum. Pavetta oblongifolia (Hiern) Bremek. Pennisetum pedicellatum Trin. Pennisetum polystachion (L.) Schult. Pennisetum sp. Pennisetum unisetum (Nees) Benth. syn. Beckeropsis uniseta (Nees & Thonn.) Pentadesma butyracea Sab. Pentodon pentandrus (Schum. & Thonn.) Vatke Pergularia daemia (Forssk.) Chiov. Pericopsis laxiflora (Benth.) van. Meeuwen Perotis indica (L.) Kuntze Phaulopsis barteri (T. Anders.) Lindau Phaulopsis ciliata (Willd.) Hepper Phaulopsis imbricata (Forssk.) Sweet. subsp imbricata Phaulopsis sylvestre (Lindau) Lindau Phoenix reclinata Jacq. Phragmites australis subsp. altissimus (Benth.) W. D. Clayton Phragmites karka (Retz.) Trin. ex Steud. Phyllanthus alpestris Beille Phyllanthus amarus Schum. & Thonn. Phyllanthus kerstingii Brunel (ined.) syn. P. beillei auct. Phyllanthus maderaspatensis L. Phyllanthus muellerianus (O. Ktze.) Exell Phyllanthus niruri L. Phyllanthus reticulatus Poir. var. reticulatus Phyllanthus sp1. Phyllanthus sp2. (Yarpao stream) Physalis angulata L. Pierreodendron kerstingii (Engl.) Little Piliostigma thonningii (Schumach.) milne-Redh. Pistia stratrioides L. Platostoma africanum P. Beauv. 204 Riparian forests plant species of Benin 718 719 Ep Th TA TA Polypodiaceae Platycerium angolense Welw. Caryophyllaceae Polycarpaea eriantha Hochst. ex A. Rich. 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 Th He mph Ep Th nph mph Th Lmph Lmph Lmph Lnph mPh mph mPh Lnph mph Lnph mph nph nph nph nph nph nph nph mph Ch mPh mPh Lnph Th Ge He He Lmph mPh Lnph mph Lmph mph Th LTh LTh LTh LTh Lnph He Th nph mph Lmph Lmph Lnph mph nph Th Cosmo TA SZ _ Pan SG GC PRA GC G G _ SZ SZ PRA TA Pan GC SG G GC GC GC GC _ GC SG Pan SZ PRA Paleo Pan Cosmo Pan _ Paleo SZ G G Paleo TA Pan GC Paleo Pan _ GC Pan Paleo Pan G G G _ GC _ Pan Polygonaceae Polygonaceae Rubiaceae Orchidaceae Portulacaceae Rubiaceae Sapotaceae Urticaceae Verbenaceae Verbenaceae Verbenaceae Verbenaceae Mimosaceae Meliaceae Anacardiaceae Papilionaceae Myrtaceae Papilionaceae Hypericaceae Hypericaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Combretaceae Adianthaceae Papilionaceae Papilionaceae Papilionaceae Amaranthaceae Cyperaceae Cyperaceae Cyperaceae Combretaceae Arecaceae Icacinaceae Apocynaceae Celastraceae Anacardiaceae Poaceae Papilionaceae Papilionaceae Papilionaceae Papilionaceae Papilionaceae Cyperaceae Poaceae Euphorbiaceae Violaceae Capparidaceae Capparidaceae Capparidaceae Rubiaceae Rubiaceae Poaceae Polygonum salicifolium Brouss. ex Willd. syn. Persicaria salicifolia (Brouss. ex Willd) Assenov (1966) Polygonum senegalense Meisn. syn. Persicaria senegalensis (Meisn.) Sojak Polysphaeria arbuscula K. Schum. Polystachya sp. Portulaca oleracea Linn. Pouchetia africana A. Rich. ex DC. Pouteria alnifolia (Baker) Roberty syn. Malacantha alnifolia (Baker) Pierre Pouzolzia guineensis Benth Premna angolensis Gürke Premna hispida Benth. Premna lucens A. Chev. Premna sp. Prosopis africana (Guill. & Perr.) Taub. Pseudocedrela kotschyi (Schweinf.) Harms Pseudospondias microcarpa (A. Rich) Engl. Pseudovigna sp. Psidium guajava L. Psophocarpus tetragonolobus (L.) DC. Psorospermum glaberrimum Hochr. Psorospermum senegalense Spach Psychotria calva Hiern Psychotria latistipula Benth. Psychotria linderi Hepper Psychotria psychotrioides (DC.) Roberty Psychotria sp. Psychotria vogeliana Benth. Pteleopsis suberosa Engl. & Diels Pteris atrovirens Willd. Pterocarpus erinaceus Poir subsp. lucens Pterocarpus santalinoides DC. Pueraria phaseoloides (Roxb.) benth. var javanica Pupalia lappacea (L.) A. Juss. Pycreus lanceolatus (Poir.) C.B. Clarke Pycreus polystachyos (Rottb.) P.Beauv. Pycreus sp. Quisqualis indica L. Raphia sudanica A. Chev. Raphiostylis beninensis (Hook. f. ex Planch.) Planch. ex Benth. Rauvolfia vomitoria Afzel. Reissantia indica (Willd.) N. Hallé syn. Hippocratea indica Lam. Rhus natalensis Bernh. ex Krause Rhynchelytrum roseum (Nees) Stapf. & C.E. Hubbard ex Bews Rhynchosia congensis Baker Rhynchosia densiflora (Roth) DC. Rhynchosia minima (L.) DC. var. minima Rhynchosia sp. Rhynchosia viscosa (Roth) DC. Rhynchospora corymbosa (L.) Britt. Rhytachne triaristata (Steud.) Stapf Ricinus communis L. Rinorea dentata (P. Beauv.) O. Ktze Ritcheia capparioides (Andr.) Britten var. capparoides Ritcheia longipedicellata Gilg. Ritcheia sp. Rothmannia urcelliformis (Hiern) Robyns Rothmannia sp. Rottboellia cochinchinensis (Lour.) syn. R. exaltata L. f. 205 Riparian forests plant species of Benin 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 nph Ch Ch Ch Ch Lmph mph mph nph nph LmPh LmPh Lnph Lnph Lnph He Lmph Lmph nph nph nph Lmph Ge Ge Ge mph Lmph Th Ge Th Th He He Th Lnph MPh He He He He He He Th Lmph mph Th(fern) nph nph nph Th Ch Ch Th nph nph Ch Th Th He He TA GC G GC G _ GC SG _ GC AA GC GC SG _ SZ GC GC PRA G _ GC GC GC G TA TA AA TA Pan PRA PRA Pan _ Pan GC TA GC GC AA GC _ TA GC SZ _ Pan AA Pan Paleo G SG TA GC Pan SG Paleo SG TA GC Connaraceae Acanthaceae Acanthaceae Acanthaceae Acanthaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Apocynaceae Apocynaceae Rubiaceae Rubiaceae Rubiaceae Poaceae Celastraceae Celastraceae Celastraceae Celastraceae Celastraceae Celastraceae Agavaceae Agavaceae Agavaceae Euphorbiaceae Rubiaceae Ochnaceae Amaryllidaceae Poaceae Poaceae Poaceae Poaceae Poaceae Mimosaceae Oleaceae Cyperaceae Cyperaceae Cyperaceae Cyperaceae Cyperaceae Cyperaceae Scrophulariaceae Asclepiadaceae Polygalaceae Selaginaceae Caesalpiniaceae Caesalpiniaceae Caesalpiniaceae Caesalpiniaceae Caesalpiniaceae Rubiaceae Pedaliaceae Papilionaceae Papilionaceae Papilionaceae Poaceae Poaceae Poaceae Poaceae Rourea coccinea (Baker) Jongkind subsp. Coccinea syn. Byrsocarpus coccineus Thonn. Ruellia praetermissa Schweinf. ex Lindau Ruellia togoensis (Lindau) Heine Rungia guineensis Heine Ruspolia hypocrateriformis (Vahl) Milne-Redhead Rutidea sp. Rytigynia canthioides (Benth.) Robyns Rytigynia senegalensis Blume Rytigynia sp. Rytigynia umbellata (Hiern) Robyns Saba comorensis (Bojer) Pichon Saba thompsonii (A. Chev.) Pichon Sabicea brevipes Wernh. Sabicea calycina Benth. Sabicea sp. Sacciolepis africana C E. Hubbard & Snowden Salacia debilis (G. Don.) Walp. Salacia hispida Blakelock Salacia leptoclada Tul. Salacia pallescens Oliv. Salacia sp. Salacia staudtiana Loes. Sansevieria guineensis (L) Willd. syn. S. trifasciata prain Sansevieria liberica Gérôme & Labory Sansevieria longiflora Sims Sapium ellipticum (Hochst. ex Krauss) Pax Sarcocephalus latifolius (Smith) Bruce syn. Nauclea latifolia Smith Sauvagesia erecta L. Scadoxus multiflorus (Martyn) Raf. subsp multiflorus Schizachyrium brevifolium Nees var. brevifolium Schizachyrium exile (Hochst.) Pilger Schizachyrium platyphyllum (Franch.) Stapf Schizachyrium sanguineum (Retz.) Alston Schizachyrium sp. Schrankia leptocarpa DC. Schrebera arborea A. Chev. Scleria achtenii De Wild. Scleria bulbifera A. Rich. Scleria depressa (C. B. Cl.) Nelmes Scleria lacustris Wright ex Sauvelle Scleria naumanniana Boeck. Scleria sp. Scoparia dulcis Linn Secamone afzelii (Schultes) K. Schum. Securidaca longipedunculata Fres. Selaginela sp. Senna alata (L.) Roxb. syn. Cassia alata L. Senna hirsuta (L.) HS Irwin & Barneby syn. Cassia hirsuta (L.) Irwin & Barneby Senna obtusifolia (L.) HS Irwin & Barneby syn. Cassia tora L. var obtusifolia (L.) Haines Senna occidentalis (L.) Link syn. Cassia occidentalis L. Senna podocarpa (frill. & Perr.) Lock syn. Cassia podocarpa Guill. & Perr. Sericanthe chevalieri (K. Krause) Robbrecht var chevalieri Sesamum indicum Linn. Sesbania leptocarpa DC. Sesbania sesban (L.) Merrill Sesbania sudanica Gillett subsp. occidentalis Gillett Setaria barbata (Lam.) Kunth Setaria gracilipes C. E. Hubbard Setaria longisepa P. Beauv. Setaria megaphylla (Steud.) Th. Dur. & Schinz syn. S. chevaleri Stapf. 206 Riparian forests plant species of Benin 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 Th Th Th He Th Ch Ch Ch Th Ch Lnph Ch Lmph LGe Ch Ch Ch Th Th Th Th He He nph Th mPh Th Th Th Th Ch Th mPh He Th Lnph Th Ge mph mph mPh mph He Th Lmph Lmph LmPh Th Lmph LmPh LMPh LmPh mph LmPh LmPh LmPh Ge Ge Ge mph Pan TA _ TA Pan Pan AA GC _ Pan GC Pan GC TA Cosmo Pan Cosmo TA TA _ SZ SG AA G Pan TA SZ PRA _ SG SG AA TA SZ _ G AA GC TA SZ GC SG Paleo PRA GC GC GC _ GC G GC GC SZ GC G PRA G SZ SZ SG Poaceae Poaceae Poaceae Poaceae Malvaceae Malvaceae Malvaceae Malvaceae Malvaceae Malvaceae Malvaceae Malvaceae Hippocrateaceae Smilacaceae Solanaceae Solanaceae Solanaceae Lamiaceae Lamiaceae Lamiaceae Poaceae Poaceae Poaceae Anacardiaceae Asteraceae Bignoniaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Menispermaceae Loganiaceae Anacardiaceae Poaceae Poaceae Icacinaceae Verbenaceae Commelianaceae Apiaceae Sterculiaceae Sterculiaceae Bignoniaceae Poaceae Scrophlariaceae Apocynaceae Apocynaceae Apocynaceae Asteraceae Loganiaceae Loganiaceae Loganiaceae Loganiaceae Loganiaceae Loganiaceae Loganiaceae Loganiaceae Araceae Araceae Araceae Papilionaceae Setaria pallidae-fusca (Schumach.) Stapf & C.E.Hubb. ex M.B. Moss Setaria restioidea (Franch.) Stapf Setaria sp. Setaria sphacellata (Schumach.) Stapf & C.E.Hubb. ex M.B. Moss Sida acuta Burm. f. subsp. acuta Sida cordata (Burm. f.) Borss. Waalk. Sida linifolia Juss. ex Cav. Sida rhombifolia L. Sida sp. Sida spinosa L. syn. S. alba L. Sida urens L. var. urens Sida veronicifolia Lam. Simicratea welwitschii (Oliv.) N. Hallé syn. Simirestis welwitschii (Oliv.) N. Hallé Smilax anceps Willd. syn. S. kraussiana Meissner Solanum nigrum L. Solanum torvum Sw. Solanum verbascifolium L. Solenostemon latifolius (Hoechst. ex Benth.) J. K. Morton Solenostemon monostachyus (P. Beauv.) subsp. monostachyus Solenostemon sp. Sorghastrum bipennatum (Hack.) Pilg. Sorghum arundinaceum (Desv.) Stapf Sorghum bicolor (L.) Moench Sorindeia warneckei Engl. Sparganophorus sparganophara (L.) Jeffrey syn. Struchium sparganophora (L) O. Ktze Spathodea campanulata P. Beauv. Spermacoce filifolia Perr.& Lepr. ex DC. syn. Borreria filifolia (S.&T) K. Schum. Spermacoce ruelliae DC. syn. Borreria scabra (Schumach. & Thonn.) K. Schum. Spermacoce sp. Spermacoce stachydea DC. syn. Borreria stachydea (DC.) Hutch. & Dalz. Sphenostylis schweinfurthii Harms Spigelia anthelmia L. Spondias mombin L. Sporobolus pyramidalis P. Beauv. Sporobolus sp. Stachyanthus occidentalis (Keay & Miege) Boutique Stachytarpheta indica (L.) Vahl syn. S. angustifolia (Mill.) Stanfieldiella imperforata (C. B. Cl.) Breman Steganotaenia araliacea Hochst. var. araliacea Sterculia setigera Del. Sterculia tragacantha Lindl. Stereospermum kunthianum Cham. var. kunthianum Streptogyna crinita P. Beauv. Striga micrantha (Benth.) Benth. Strophanthus hispidus DC. Strophanthus preussii Engl. & Pax Strophanthus sarmentosus DC. Struchium sp. Strychnos afzelii Gilg. Strychnos barteri Solered. Strychnos congolana Gilg. Strychnos floribunda Gilg. Strychnos innocua Del. Strychnos nigritana Baker Strychnos splendens Gilg. Strychnos usambarensis Gilg. Stylochaeton hostifolius Engl. Stylochaeton hypogaeus Lepr. Stylochaeton lancifolius Kotschy & peyr. Swartzia madagascariensis Desv. 207 Riparian forests plant species of Benin 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 Th mPh mph mPh Lmph Ge Th mPh mPh mPh Th Ch mph mPh mph mph mPh LmPh LmPh mPh Ch(Fern) Ge(Par) Lnph Lnph Lnph Lnph Ge mph Th Lnph mph mph nph mph nph mph mph mPh mph mph Lnph Th mPh Ch Ch Ch Th Th Ch Lnph Lnph Lnph Lnph Lnph mPh mPh Ch Ge Lmph Lmph Pan GC TA TA TA Pan Pan Pan GC Paleo SG TA SG SG SZ _ GC GC GC G _ G Pan S GC GC Pan GC Pan G SZ SZ G _ GC SZ GC GC G _ G Pan GC Pan GC S Pan _ TA SG GC G _ GC GC TA Pan SG GC PRA Asteraceae Sapotaceae Sapotaceae Myrtaceae Asclepiadaceae Taccaceae Portulacaceae Caesalpiniaceae Rubiaceae Verbenaceae Papilionaceae Papilionaceae Combretaceae Combretaceae Combretaceae Combretaceae Combretaceae Dilleniaceae Dilleniaceae Mimosaceae Thelypteridaceae Balanophoraceae Acanthaceae Acanthaceae Acanthaceae Acanthaceae Cyperaceae Ulmaceae Aïzoaceae Malpighiaceae Rubiaceae Rubiaceae Rubiaceae Meliaceae Rubiaceae Meliaceae Meliaceae Meliaceae Anacardiaceae Anacardiaceae Menispermaceae Asteraceae Moraceae Tiliaceae Tiliaceae Tiliaceae Tiliaceae Tiliaceae Tiliaceae Cucurbitaceae Asclepiadaceae Asclepiadaceae Asclepiadaceae Asclepiadaceae Euphorbiaceae Euphorbiaceae Malvaceae Liliaceae Loganiaceae Annonaceae Synedrella nodiflora Gaertn. Synsepalum brevipes (Baker) Pennington syn. Pachystela brevipes (Baker) Engl. Synsepalum passargei (Engl.) Pennington syn. Vincentella passargei Engl. (Aubréville) Syzygium guineense (Willd.) DC. subsp. guineense Tacazzea apiculata Oliv. Tacca leontopetaloides (L.) O. Ktze syn. T. involucrata Schum Talinum triangulare (Jacq.) Willd. Tamarindus indica L. Tarenna eketensis Wernham. Tectona grandis L. f. Tephrosia platycarpa Guill. & Perr. syn. T. humilis. & Perr. Tephrosia vogelii Hook. f. Terminalia avicennioides Guill. & Perr. Terminalia glaucescens Planch. ex Benth. Terminalia macroptera Guill. & Perr. Terminalia sp. Terminalia superba Engl. & Diels Tetracera alnifolia Willd. Tetracera potatoria Afzel. ex G. Don Tetrapleura tetraptera (Schum. & Thonn.) Taub. Thelypteris sp. Thonningia sanguinea Vahl. Thunbergia alata Boj. ex Sims. Thunbergia atacoriensis Akoègninou & Lisowski Thunbergia cynanchifolia Benth. Thunbergia togoensis Lindau Torulinium odoratum (L.) Hopper Trema orientalis (L.) Blume syn. T. guineensis (Schum. & Thonn.) Ficalho Trianthema portulacastrum L. Triaspis odorata (Willd.) A. Juss. Tricalysia okelensis Hiern var. okelensis Tricalysia okelensis Hiern var. pubescens Tricalysia reticulata (Benth.) Hiern Tricalysia sp. Tricalysia subquadrata A. Rich. ex DC. Trichilia emetica Vahl subsp. emetica Trichilia prieuriana A. Juss. Trichilia retusa Oliv. Trichoscypha smythei Hutch. & Dalz. Trichoscypha sp. Triclisia subcordata Oliv. Tridax procumbens L. Trilepiseum madagascariense DC. syn. Bosqueia angolensis Ficalho Triumfetta cordifolia A. Rich Triumfetta dubia De Wild. Triumfetta lepidota K. Schum. Triumfetta rhomboidea Jacq. var. rhomboidea Triumfetta sp. Triumfetta tomentosa Jacq. Trochomeria macrocarpa (Sond.) syn. T. macroura Hook. F. Tylophora camerooniana N. E. Br. Tylophora dahomensis K. Schum. Tylophora sp. Tylophora sylvatica Decne. Uapaca heudelotii Baill. Uapaca togoensis Pax syn. U. somon Aubrév. & Léandri Urena lobata L. Urginea altissima (L.f.) Baker Usteria guineensis Willd. Uvaria chamae P. Beauv. 208 Riparian forests plant species of Benin 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 nph nph mph nph mph nph Ch Ch mph Ch Th Ch Ch Ch He Th LTh Lnph Lnph Lnph LTh LGe mPh mph mPh nph mph Th Ch Th mph nph mPh mPh mPh Ge mPh mph Th Th Lnph Lnph LTh Lnph mph mph mph G G GC _ SG _ SG G SZ SG TA G G _ SZ TA GC PRA PRA _ Pan AA SG SZ SZ SZ TA AA Pan SZ SZ Pan GC G GC PRA SG SG Cos Pan TA PRA _ PRA Paleo PRA Paleo Annonaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Asteraceae Asteraceae Asteraceae Asteraceae Asteraceae Asteraceae Asteraceae Asteraceae Poaceae Asteraceae Papilionaceae Papilionaceae Papilionaceae Papilionaceae Papilionaceae Papilionaceae Sapotaceae Verbenaceae Verbenaceae Verbenaceae Apocynaceae Campalunaceae Sterculiaceae Malvaceae Papilionaceae Olacaceae Annonaceae Annonaceae Annonaceae Xyridaceae Sapindaceae Rutaceae Mimosaceae Poaceae Cucurbitaceae Cucurbitaceae Cucurbitaceae Cucurbitaceae Rhamnaceae Rhamnaceae Rhamnaceae Uvaria ovata (Dunal) A. DC. Vangueria kerstingii Robyns Vangueria madagascariensis Gmelin syn. V. venosa Robyns Vangueria sp. Vangueriella spinosa (Schum & Thonn) syn.Vangueriopsis spinosa (Schumach & Thonn) Vangueriopsis sp. Vernonia bambilorensis Berhaut Vernonia camporum A. Chev. Vernonia colorata (Willd.) Drake Vernonia glaberrima Welw. ex O. Hoffm. Vernonia pauciflora (willd.) Less. Vernonia poskeana Vartke & Hildebrandt Vernonia purpurea Sch. Bip. ex Walp. Vernonia sp. Vetiveria nigritana (Benth.) Stapf Vicoa leptoclada (Webb) Dandy Vigna gracilis (Guill. & Perr.) Hook. f. syn. Dolichos gracilis Guill. & Perr. Vigna racemosa (G. Don) Hutch. & Dalz Vigna reticulata Hook. f. Vigna sp. Vigna unguiculata (L.) Walp. var. unguiculata Vigna vexillata (L.) A. Rich. var. vexillata Vitellaria paradoxa C. F. Gaertn. subsp. paradoxa Vitex chrysocarpa Planch. ex Benth. Vitex doniana Sweet Vitex madiensis Oliv. Voacanga africana Stapf Wahlenbergia perrottetii (A. DC.) Thulin syn. Cephalostigma perrottetii A. DC. Waltheria indica L. Wissadula amplissima (L.) R. E. Fries Xeroderris stuhlmannii (Taub) Mendonça & E.P. Sousa Ximenia americana L. Xylopia aethiopica (Dunal) A. Rich. Xylopia elliotii Engl. & Diels Xylopia parviflora (A. Rich.) Benth. Xyris capensis Thunb. Zanha golungensis Hiern Zanthoxylum zanthoxyloides (Lam.) Zepernick & Timber Zapoteca portoricensis (Jacq.) H. M. Hern. syn. Calliandra portoricensis (Jacq.) benth. Zea mays L. Zehneria capillacea (Schumach.) C. Jeffrey syn. Melothria capillacea (Schumach.) Cogn. Zehneria hallii C. Jeffrey syn. Melothria deltoidea Benth. Zehneria sp. Zehneria thwartesii (Schweinf.) C. Jeffrey syn. Melothria tridactyla Hook. f. Ziziphus mauritiana Lam. Ziziphus mucronata Willd. subsp. mucronata Ziziphus spina-cristi (L.) Desf. var. microphylla Hochst. A. ex A. Rich. Life forms (LF) followed Raunkaier (1934), Schnell (1971) and Keay & Hepper (1954-1972). They were: - Phanerophytes (Ph): mega-phanerophyte (MPh > 30 m), meso-phanerophyte (mPh: 8 to 29 m), microphanerophyte (mph: 2 to 7 m) and nano-phanerophyte (nph < 2m); - Therophytes (Th); - Hemicrytophytes (He); - Chamaephytes (Ch); - Lianas (L); - Geophytes (Ge); - Epiphytes (Ep) and - Parasites (Par). The phytogeographic types (PT) were named after White (1986) and Keay & Hepper (1954-1972). They were: - Species widely distributed in the tropics (Cosmopolitan-Cosmo, Pantropical-Pan, Afro-American-AA and Paleotropical-Paleo). - Species widely distributed in Africa (Tropical Africa-TA, Pluri Regional in Africa-PRA) - Regional species in Africa (Sudanian-S, Guinean-G, Sudano-Guinean-SG, Sudano-Zambesian-SZ, GuineoCongolian-GC). 209 Acknowledgements ACKNOWLEDGEMENTS This thesis could have not been completed without the help of many people and institutions. I have received a great deal of moral support from my family, friends and colleagues. Several people have read the various chapters and helped me very much in polishing the English. In the first place, my great appreciation and deep gratitude are due to my supervisor in Benin, Professor Dr. B. Sinsin, Faculty of Agronomic Sciences, Department of Environment Management, University of Abomey-Calavi. He has proposed and supported my candidature for the fellowship of the Project Flora of Benin. I wish to extend my sincere thanks for proposing the topic of the work and for his excellent guidance and constructive criticisms. I would like to thank my promotor Professor Dr. L.J.G. van der Maesen, for his open mind, gentleness, and excellent guidance. He is gratefully acknowledged for his skilful improvement of the English text. His critical reading and subtle comments were of great help in improving all the drafts of my chapters. I learned from him that patience is always rewarding. Thanks are due to Professor Dr. A. de Gier, Forest Science Division ITC, Enschede, for his valuable contribution, especially in methodological aspects of my research and chapter 10. Special gratitude goes to the Dutch Government for financing the Project Flora of Benin, which granted me financial support to carry out MSc and PhD studies. The staff of the Project Flora of Benin in Abomey-Calavi and Wageningen helped me during all the steps of this research. I take this opportunity to thank Dr. N. Sokpon, Dr. A. Akoègninou, Dr. J-P. Essou, Dr. V. Adjakidjè, and all lecturers of the Department of Environment Management of the Faculty of Agronomic Sciences, University of Abomey-Calavi, Benin. Special thanks are due to the colleagues of the Project Flora of Benin, Paul H. Yédomonhan, Pierre O. Agbani, Aristide C. Adomou, Ebenezer Ewedje, Raoul A. Adjobo and Nestor Gbéssi for their valuable help during data collection, plants determination and data analysis. The convivial atmosphere we had helped me very much overcome difficulties inherent to such studies. I would like to thank Dr. R. N’tia, Dr. Kaki, Dr. S. Seibou Toléba (University of Abomey-Calavi, Benin), G. Ossory, Dr. J.-P. Séro, who have constantly encouraged me. I would like to express my gratitude to the staff of the Direction des Forêts et Ressources Naturelles (DFRN), the Centre National de Gestion des Réserves de Faune (CENAGREF) and the Projet de Restauration des Ressources Forestières (PRRF – Bassila) for granting me special permission for working in the protected areas of Benin. Very kind hospitality and cooperation have been provided by all the foresters, fauna guards and all those who tirelessly worked with me all the days of my data collection at Samiondji, Toui-Kilibo, Idadjo, Bétérou, Pénéssoulou, Affon, Yarpao, Boukoumbé, Toucountouna, Kandi, Gbèssè and in the Pendjari Biosphere Reserve. It is practically impossible to name them all here. Without their help I could not have been able to complete my fieldwork. I also appreciated the hospitality of the people of the villages I visited throughout the country. I appreciated the very helpful reading and useful comments of Professor S. Porembski of Rostock University (Germany), especially for chapters 3 and 9. Also, I would like to express my sincere regards to Professor F. Bongers, of Wageningen University for his critical comments. My gratitude goes to all the staff and members of the Biosystematics group, Department of Plant Sciences, Wageningen University, The Netherlands, for the pleasant working atmosphere and their sincere help throughout my stay at Wageningen, and finalisation of this thesis. I also thank Kadry N.E. Abdel Khalik and other PhD students and project workers for the literature research they did for me during the time I was in Benin. They were crucial for the finalisation of my chapters. 211 Acknowledgements Ebenezer Ewedje should be thanked for the maps he draw, and Grace Nangendo for comments and helpful reading. I would like to acknowledge the companionship of Adi Mama, Marcel Houinato, Madjidou Oumorou, Théophile Sinadouwirou, Richard Bawa, Raoul Adjobo, Antonio Sinzogan, Roger Sodjinou, Florent Okry, Honoré Biaou, Brice Tenté, Yacoubou Boni, Etotépé Sogbohossou, Bienvenu Gangboché, and Barthélemy Kassa. I also appreciated the help of all the colleagues of the Laboratoire d’Ecologie Appliquée (FSA/DAGE), and all those whose names are not mentioned here. Finally my special thanks to my lovely mother for her never-ending support, and my gratitude to my sister Sylvia, to her family and to Espérance Obonté who have constantly encouraged me to achieve my aims. 212 Curriculum Vitae CURRICULUM VITAE The author of this dissertation, Armand Kuyéma NATTA, was born in Tanguiéta (North Benin), on 3 May 1972. He joined the Faculty of Agronomic Sciences, University of Abomey-Calavi in 1992. He obtained his BSc degree in General Agronomy in 1995. He continued his post-graduate studies in Agronomy (option Natural Resources Management) at the same University and obtained the diploma of Agronomic Engineer in 1997. He worked a short time as technical director of a Company specialised in the distribution of fertilisers and advice to cotton farmers. In 1998, he came back to the Faculty of Agronomic Sciences as research assistant. The same year he was granted a fellowship by the Project Flora of Benin to study Remote Sensing and Geographical Information Systems at the International Institute for Aerospace Survey and Earth Sciences (ITC), Enschede, The Netherlands. In February 2000, he was awarded an MSc degree in the Forest Science Division by ITC, with an overall qualification of very good. In March 2000, he began his PhD studies on riparian forests diversity in Benin, within the Project Flora of Benin, and in close cooperation with the Biosystematics group and National Herbarium of the Netherlands, Wageningen University Branch, Department of Plant Sciences, Wageningen University, The Netherlands. He has participated in several regional workshops in Benin, Burkina-Faso and Côte d’Ivoire, and in the XII World Forestry Congress held in Québec, Canada, in September 2003. His addresses in Benin are: Office: Faculty of Agronomic Sciences, Department of Environment Management (FSA/DAGE), University of Abomey-Calavi. 01 BP 526 Cotonou, Benin (West Africa). Tel/Fax: 00 229 303084 or Tel 00 229 360126 Home: 02 BP 906 Cotonou, Benin (West Africa) Tel: 00 229 055682 or 00 229 490087 or 00 229 822172 e-mail: natta@bj.refer.org aknatta@yahoo.com aknatta@hotmail.com 213 List of publications LIST OF PUBLICATIONS Natta A.K., Sinsin B. & van der Maesen L.J.G. 2002. Riparian forests, a unique but endangered ecosystem in Benin. Notulae Florae Beninensis 4. Botanische JahrbĦcher 124: 55-69. Natta A.K., Sinsin B. and van der Maesen L.J.G. 2003. Riparian forests and biodiversity conservation in Benin (West Africa). Paper presented at the XII World Forestry Congress, organised by FAO in Québec City. In: Proceedings of the XII World Forestry Congress Area B: Forest for the planet. B2a: Maintenance of Biodiversity. Voluntary paper of major interest to the deliberations (level 2, No. 0356). Canada, September 21 to 28, 2003. pp 126-127. Natta A.K. & Porembski S. Ouémé and Comoé: forest-savanna border relationships in two riparian ecosystems in West Africa. Notulae Florae Beninensis 8. Botanische JahrbĦcher. (in press). Natta A.K., Sinsin B. & van der Maesen L.J.G. A phytosociological study of Riparian forests in Benin (West Africa). Belgian Journal of Botany. (in press). 215