Open Access
How to translate text using browser tools
1 July 1995 Forest Regeneration in the Tana River Primate National Reserve, Kenya
Kimberly E. Medley
Author Affiliations +

This paper examines forest regeneration in the Tana River Primate National Reserve, using data on the size-class abundances and site conditions of major canopy tree species. Forests vary from those dominated by Pachystela msolo to mixed forests of greater species richness with Sorindeia madagascariensis and Diospyros mespiliformis or with Garcinia livingstonei and Mimusops obtusifolia. The low occurrence of saplings and narrow range of soil conditions supporting mature individuals project a relative decline for P. msolo in the study area. Size-class associations between six major canopy trees show a successional tendency toward greater species diversity but a decline in the regeneration of primate resources. Significant differences in the environmental conditions of Ficus sycomorus, P. msolo and S. madagascariensis document possible causal changes in the physical environment. The effects of riverine changes on the regeneration of primate habitat and its persistence potentially jeopardises long-term conservation in this-small reserve.

FOREST REGENERATION IN THE TANA RIVER PRIMATE NATIONAL RESERVE, KENYAKimberly E. MedleyInstitute of Primate ResearchNational Museums of Kenya andDepartment of GeographyMiami University, Oxford, OH 45056, USA;ABSTRACTThis paper examines forest regeneration in the Tana River Primate National Reserve, using data on the size-class abundances and site conditions of major canopy tree species. Forests vary from those dominated by Pachystela msolo to mixed forests of greater species richness with Sorindeia madagascariensis and Diospyros mespiliformis or with Garcinia livingstonei and Mimusops obtusifolia. The low occurrence of saplings and narrow range of soil conditions supporting mature individuals project a relative decline for P. msolo in the study area. Size-class associations between six major canopy trees show a successional tendency toward greater species diversity but a decline in the regeneration of primate resources. Significant differences in the environmental conditions of Ficus sycomorus, P. msolo and S. madagascariensis document possible causal changes in the physical environment. The effects of riverine changes on the regeneration of primate habitat and its persistence potentially jeopardises long-term conservation in this small reserve.INTRODUCTIONThe lower floodplain of the Tana River, Kenya, provides a groundwater regime suitable for a narrow corridor of tropical forest, up to 30 m in stature, in an otherwise thorn-scrub environment. Riverine forest patches persist in isolation from the rain forests of Central Africa and the Indian Ocean coast (Lind & Morrison, 1974). Along the Tana River, as in other riverine systems, the distribution off forests and their composition are determined by the groundwater-hydrologic regime (Hughes, 1988) and the dynamics of a meandering river system (cf. Brinson, 1990; Salo and Räsänen, 1989). Floodplain disturbances are primary factors influencing forest dynamics in this region.The Tana River Primate National Reserve was established in 1976 to preserve the best remaining riverine forest along the Tana River and important habitat for the endangered Tana River red colobus (Colobus badius rufomitratus) and crested mangabey (Cercocebus galeritus galeritus) (Marsh, 1976). The riverine forest mosaic provides an isolated refuge for plant and animal species adapted to a moist climatic regime (Andrews et al., 1975). A regional examination of the flora and structural characteristics of floodplain forests along the Tana River suggests that river meanders, erosion, and deposition result in a high diversity of land forms and forest types in the TRPNR vicinity (Medley, 1992).Irrigation projects, upstream and downstream from the Reserve, jeopardise regional conservation of riverine forest along the Tana River (Hughes, 1987, Ledec, 1987, Medley et al., 1989), and argue for stewardship activities in the protected area. The Reserve, however, is one of the smallest in East Africa (171 km⟨sup⟩2⟨/sup⟩) and the forest area is fragmented and much smaller (9.5 km⟨sup⟩2⟨/sup⟩ in approximately 26 areas). A primate survey conducted in 1985, ten years after the establishment of the Reserve, showed population declines equal to 80% for the red colobus and 25% for the crested mangabey (Marsh, 1986). Marsh (1986) reported that the sharp decline in their populations may be attributable to a corresponding decline in forest habitat, primarily through the loss of important canopy-tree food resources. Furthermore, Hughes (1988) concluded from a study of forests between Bura and the Reserve that pioneer forest areas were absent, and that tree regeneration levels were low at all sampling plots. These findings, while providing an overview of forest condition at locations along the Tana River, support the need for a detailed examination of regeneration in the protected area. Preservation of primate habitat depends on understanding, and potentially managing for, vegetation change in a small land area (e.g., White and Bratton, 1980).Based on field research conducted between 1987-1988 in the Reserve (Medley, 1990), I address three questions concerning forest regeneration: (1) what is the current status of tree-species establishment; (2) do significant associations exist among the size classes of major canopy trees that may reflect species preferences or successional tendencies; and (3) do sites near three tree species, important as primate food resources (sensu Decker, 1989 and Kinnaird, 1990; see Medley, 1993a), differ significantly in their environmental conditions and potential for forest regeneration? The overall objective is to quantify the current status of forest regeneration through an examination of the ecology and successional tendencies of important riverine trees. Although an absence of long-term data on the vegetation and floodplain land forms limits successional projections, the results should show several key patterns of forest regeneration within this region, especially as it may be related to establishment and change in primate habitat. The study thereby provides a basis for further research in this region.STUDYAREAThe Tana River flows from the humid highlands near Mt Kenya and the Aberdare Mountain Range, through an arid-semiarid floodplain between Garissa and Garsen, and to the Indian Ocean north of Malindi (fig. 1). At Hola (1° 30' S, 40° 2' E; 100 m), approximately 40 km upstream from the Reserve, annual minimum and maximum temperatures average 21.4 °C and 33 °C, respectively (Muchena, 1987). The climate is anomalously dry with an annual mean precipitation of 470 mm, no months with greater than 100 mm precipitation, and a growing season of 40 days occurring coincident with the short rains (October-December) (FAO 1984). The life zone is thorn woodland (sensu Holdridge 1967).Evergreen-semi-evergreen forest vegetation along the floodplain is groundwater-dependent and its lateral extent is determined sharply by access to the groundwater with distance from the river (Hughes, 1988). Furthermore, forest community types correspond with particular flooding frequencies (Hughes, 1990) and land forms (Hughes, 1988, Medley, 1992), emphasising the close relationship between river dynamics, floodplain development, and the mechanisms of forest succession (Hughes, 1994). Given the dynamics of a meandering stream (Medley and Hughes, in Press), human and use activities (Medley, 1993b), and large mammal populations (Allaway, 1979; Marsh, 1976), riverine forest occurs as a patch mosaic (fig. 2).Figure 1. The Tana River basin, Kenya (modified from Medley 1992).METHODSStatus of tree-species establishmentI selected twelve forest areas in the Reserve that were representative of the forest-community types in this region. They varied in tree-species composition and primate abundances (Fig. 2; Medley, 1993a). In each forest, sample points were located at 50 m distances along transects placed approximately 100 m apart, and quadrats or plots were randomly orientated from each point centre. The number of transects and points per study area varied, depending on the size of the forest. In total, 363 points were used to document intraforest variation in the Reserve, as well as differences among the twelve forest study areas in tree-species composition and levels of regeneration.Figure 2. The study area located in the south-central sector of the Reserve. The twelve forest areas include: Guru North (gn), Guru South (gs), Mchelelo West (mw), Congolani Central (cc), Congolani West (cw), Sifa West (sw), Baomo North (bn), Baomo South A (bsa), Baomo South B (bsb), Kitere West (kw), Mnazini North (mn), and Mnazini South (ms). The top inset shows the reserve boundary and riverine forest (shaded) (reprinted from Medley, 1992).Vegetation data were collected from four structural layers at each sample point, using point-centred quarter sampling for large trees (⟩20 cm dbh) and nested rectangular plots for the sub-canopy (10-20 cm dbh in 8 x 14m plots), sapling (⟩1 m in height and ⟨ 10 cm dbh in 4 x 6 m plots), and seedling (.03-1 m in height in 2 x 2 m plots) layers. Species that obtain heights greater than 10m were considered canopy trees, and their size-class abundances were derived from point frequencies, total densities, and relative densities. Data analyses focused on major canopy species that were most abundant as large trees in the study areas. A geographic information system (GIS) was used to show spatial patterns in the occurrences of the major tree species in the study forest areas by extrapolating from the sample points.Associations among major forest canopy tree speciesSuccessional tendencies between important riverine trees were investigated using the inter-associations among species size classes (Pielou, 1977, Zedler and Goff, 1973). Spearman correlations were computed to determine the direction and significance of associations between abundances in the large-tree size class of each major canopy tree species and abundances in their respective size classes (i.e., seedlings, saplings, sub-canopy trees, and large trees).Three associations with the large-tree size class of a major canopy species by another species were possible: (1) Invader: positive correlation for the sub-canopy, sapling, and/or seedling size-class; (2) Associate: positive correlations for the large-tree size class or all size classes; and (3) Nonassociate: negative correlations with all size classes. Negative correlations between the abundances of a canopy species and its sapling size class, but positive correlations with the sapling size class of other canopy species, suggest a change in forest composition. If a successional pathway is evident, one would expect a canopy tree species to enter (invade), associate, and decline (nonassociate) according to a particular sequence. Negative correlations between the large-tree size class of one canopy species and the sapling size classes of all canopy species indicate low forest regeneration.Site and plant community characteristics near three canopy tree speciesFicus sycomorus, Sorindeia madagascariensis, and Pachystela msolo are highly used food resources for the endangered primates (Decker, 1989, Homewood, 1978, Kinnaird, 1990, Marsh, 1981), were identified as common components of suitable primate habitat (Medley 1993a), and appeared representative of different forest communities and/or stages of development (Hughes, 1988, Medley, 1992). Working in the same 12 forest areas, I selected 16 mature F. sycomorus, 15 P. msolo, and 10 S. madagascariensis trees for a comparative study of their site conditions and evidence of forest regeneration. Three individuals of P. msolo were selected from a large open area of the Baomo South forest that had experienced stand-level death of this species after a major flood in 1969.Circular plots (radius = 15 m; 707 m⟨sup⟩2⟨/sup⟩) were established around each tree and divided into four quadrats (A, B, C, D) along a random orientation. Soil texture, subsurface soil moisture, and flood heights were compared as three attributes directly related to the flooding regime and moisture characteristics of the floodplain environment (see Hughes, 1990, 1994). Composite soil samples were collected in each circular plot at 30 cm in depth, and soil textures were determined by the Soil Testing Laboratory at the National Agricultural Laboratories, Kenya Ministry of Agriculture. Soil moisture percentages were determined by taking initial and dry weight measurements from samples collected at 1.5 m. Samples were collected from all plots during a 26.5 hour period in order to limit moisture differences attributable to a fluctuating water table. After a flood event in May 1988 (fig. 3), I measured the maximum water height as evidenced by the high-water markings on each selected tree and the depth of newly deposited sediments. The species richness and abundances of tree saplings, greater than 1 m in height and less than 10 cm dbh, were measured in quadrat A (177 m⟨sup⟩2⟨/sup⟩).Figure 3. Tana River discharge measured daily at the Garissa station gauge. Data obtained from the Hydrology Department, Kenya Ministry of Water Development. Discharge is measured in cubic meters per second (cumecs) and the recurrence interval for the flood in May 1988 is derived from long-term data for the basin (see Hughes, 1990).Figure 4. Spatial distribution of the six major canopy-tree species in the forest study areas. Forest names are provided in figure 2. Note that Mimusops obtusifolia was formerly identified as Mimusops fruticosa and that Pachystela msolo is now called Synsepalum msolo by Pennington (1992) in the Genera of Sapotaceae (see Turrill et al., 1952)Figure 5. Densities of the six major canopy-tree species at four forest layers.Figure 6. Relative densities of the six major canopy-tree species in the large-tree and sapling size classes.RESULTSStatus of tree-species establishmentCompositional differences are evident among the twelve forest areas, based on the densities of canopy trees (table 1). Pachystela msolo, identified in earlier Tana River research as P. brevipes (see Homewood, 1978; Hughes, 1988; Marsh, 1976) and now listed as Synsepalum msolo by Pennington (1992) in the Genera of Sapotaceae (see Turrill et al., 1952), is the most abundant tree in four forest areas, accounting for approximately 23-54% of the total large-tree density in each area. The rest of the forests have a mixed composition with Sorindeia madagascariensis, often in association with Diospyros mespiliformis, or with Mimusops obtusifolia, identified in earlier research as M fruticosa, and Garcinia livingstonei. The density of Ficus sycomorus is high in several forests, mostly in association with P. msolo. Based on their densities in the twelve forest areas, these six trees (F. sycomorus, P. msolo, S. madagascariensis, D. mespiliformis, G. livingstonei, and M. obtusifolia) are identified as the most important canopy-tree species. Their current distribution in the study area appears non-random, especially evident by the clustered distribution of P. msolo in the southern forests (except Mnazini South) and the concentration of F. sycomorus along the river (e.g. Guru North, Baomo South A, and Mnazini North; fig.4).Figure 7. Spearman correlations (Rs) between the size classes of five canopy trees and the large-tree size class of Pachystela msolo (top graph) and Mimusops obtusifolia (bottom graph).Figure 7. Spearman correlations (Rs) between the size classes of five canopy trees and the large-tree size class of Pachystela msolo (top graph) and Mimusops obtusifolia (bottom graph).Levels of establishment by these six trees vary from relatively low densities in the seedling and sapling classes (Ficus sycomorus) to a nearly bimodal distribution pattern with lowest densities in the sub-canopy size class (fig. 5). In the large-tree size class, Pachystela msolo is the most abundant tree, representing nearly 35% of all the canopy-tree individuals (fig. 6). The tree occurs as the single dominant species (relative density ⟨0.4) in three forest patches (table 1). At the sapling layer, however, P. msolo is among the least abundant of the canopy species (fig. 6). Saplings were recorded at only seven points out of 363, and the low number of individuals at those points resulted in a low overall density (fig. 5).Table 1. Density (No/ha) of canopy trees recorded in the large-tree size class (⟩20 cm dbh). The twelve forest areas include: Mnazini North (mn), Baomo South b (bsb), Kitere West (kw), Baomo South a (bsa), Baomo North (bn), Mchelelo West (mw), Guru South (gs), Guru North (gn), Congolani Central (cc), Congolani West (cw), Sifa West (sw), and Mnazini South (ms) (see Fig. 2). Forests are arranged according to their similarity to Mnazini North, as measured by the relative abundances of all co-occurring canopy species and the six most important canopy trees are identified in bold print.Tree speciesmnbsbkwbsabnmwgsgncccwswmsPachystela msolo31.739.727. salviifolium7. sycomorus6. paniculata5. venosum2. unijugata1. mespiliformis1. goetzii0. minor0.50.415.9Ficus bussei0.50.5Ficus natalensis0.50.4Garcinia livingstonei0. africana0. madagascariensis21.819. sessiliflora1.3Hunteria africana1.30.80.5Lepisanthus senagalensis1.3Mimusops obtusifolia1. reclinata2.90.40.6Tree speciesmnbsbkwbsabnmwgsgncccwswmsAlbizia glaberrima0. narcissodora0.7Oxystigma msoo0.71.3Mangifera indica2.71.00.4Acacia robusta0. micrantha0.7Lannea schweinfurthii0. rovumae0. aethiopum0.41.3Polysphaeria multiflora0.4Spirostachys venenifera0. appendiculata0. orientalis0.40.3Albizia gummifera0. kabuyeana0.4Thespesia danis0.4Hyphaene compressa11.73.70.610.04.69.4Majidea zanguebarica0. ferrea0.30.20.6Markhamia zanzibarica0.30.60.4Oncoba spinosa0.30.2Pavetta sphaerobotrys0.30.4Tamarindus indica0.3Ziziphus pubescens2. speciesmnbsbkwbsabnmwgsgncccwswmsCeltis occidentalis1.01.9Populus ilicifolia0.41.00.5Afzelia quanzensis0.21.20.5Dobera glabra0.21.0Ficus scasselatii0.2Cynometra lukei0. inoploeum0.6Lecaniodiscus fraxinifolius0.60.7Terminalia brevipes1.5Cassia abbreviata0.41.0Lawsonia inermis0.2Cordia sinensis1.00.4Salvadora persica1.0Lamprothamnas zanguebaricus04Total Density (# trees/ha)58.984.666.544.236.742.137.71933.836.937.269.4Number of Tree Species131212211726311525201122Sample Points31172337234152104721952Table 2. Size-class association analyses. Data columns show significant (+) and negative (-) Spearman correlations (alpha ⟨0.05) between the size classes of species listed by rows (sd = seedling, sp = sapling, sc = sub-canopy, and c = large tree) and the large-tree size class of six major canopy tree species.Large-Tree Size Class of Six Major Canopy Tree SpeciesPlant species (rows) and their size classes (columns)Ficus sycomorusPachystela msoloSorindeia madagascariensisDiospyros mespiliformisGarcinia livingstoneiMimusops obtusifoliasdspsccsdspsccsdspsccsdspsccsdspsccsdspsccFicus sycomorus++++--Pachystela msolo++++++---+-Sorindeia madagascariensis+++++++----Diospyros mespiliformis-+++++--Garcinia livingstonei-++-++++-Mimusops obtusifolia--+++Associations between major forest canopy tree speciesSpearman correlations between the size classes of five canopy species and the large-tree size class of Pachystela msolo and Mimusops obtusifolia show two contrasting patterns (fig. 7). Seedlings and saplings of most species are positively associated with large trees of P. msolo (except M obtusifolia) and negatively associated with M obtusifolia (except M obtusifolia). Although self-replacement (i.e., positive association with its saplings) is non-significant, other canopy species are established under the P. msolo canopy. In contrast, M obtusifolia shows a negative relationship with those size classes, or an absence of invading species. P. msolo and M obtusifolia are nonassociates as large trees. Significant positive and negative associations between the respective size classes and the large-tree size class of the six major canopy species (table 2), coupled with field observations and summaries for the twelve forest areas, illustrate some relative preferences and successional tendencies. P. msolo and S. madagascariensis invade beneath F. sycomorus and occur as positive associates at the canopy layer. The co-occurrences between these two trees and Diospyros mespiliformis at the canopy (e.g., Kitere West and Baomo South; table 1 and fig. 4), and the significant invasion of S. madagascariensis and D. mespiliformis under P. msolo indicate a successional tendency toward a mixed-forest composition. Garcinia livingstonei and Mimusops obtusifolia are significant invaders under Sorindeia madagascariensis and Diospyros mespiliformis, respectively, but show a negative or non-significant association with these species as large trees. Community differences are greatest between Pachystela-dominated forests (e.g., Mnazini North) and Mimusops-Garcinia forests (e.g., Mnazini South; see table 1). Ficus sycomorus, Pachystela msolo, and S. madagascariensis show negative correlations with large trees of one or both of these species. These trees either decline upon this change in canopy-species dominance, or segregate to different sites (sensu Hughes, 1988) irrespective of their preferences as saplings.Site conditions and forest regeneration near three canopy tree speciesIndividuals of Ficus sycomorus, Pachystela msolo, and Sorindeia madagascariensis show significant differences in their site conditions and evidence of sapling; establishment (fig. 8). Ficus sycomorus shows an adaptation to a wide range of sand percentages (mean = 35.9; sd = 21.1; range = 14-86%), in contrast to the greater sand content of soils associated with Sorindeia madagascariensis (mean = 53.6; sd = 19.2; range 24-86%) and the low percentages of sand associated with Pachystela msolo (mean = 26.5; sd = 10.9; range = 2-48%). These differences correspond with those observed for soil moisture percentages at 1.5 m in depth. Soil moisture is highest and least variable under P. msolo (mean = 22.3; sd= 1.72; range = 14.3-31.4%). Whereas highest flooding heights (92 cm) and sediment deposits (8 cm) are measured for F. sycomorus plots, 11 of the 15 P. msolo plots flooded (up to 56 cm) and the mean sediment depth (1.4 cm) exceeds that recorded for F. sycomorus.Highest species richness at the sapling layer occurs in the Sorindeia madagascariensis plots (mean = 14; sd = 2 spp.). The mean value closely approximates that measured for Pachystela msolo (mean = 14; sd= 5 spp.), which is especially high for individuals in the area of stand-level death (mean = 23 spp.). Primate food resources show an opposite relationship. They are most abundant in the Ficus sycomorus plots (mean =210; sd = 120/ha), decline in the P. msolo plots (mean = 128; sd = 162), and are significantly lower in the S. madagascariensis plots (mean = 78; sd= 58/ha).DISCUSSIONThe twelve forest areas studied in the Reserve vary in composition from those dominated byFigure 8. Site and regeneration characteristics of the plots centered on Ficus sycomorus (n=16), Pachystela msolo (n=15), and Sorindeia madagascariensis (n=10). The graphs show the means and standard deviations for the following attributes: a) percent sand in soils collected at 30 cm in depth; b) percent soil moisture measured at 1.5 m in depth; c) maximum water height during the May 1988 flood; d) depth of sediments deposited during the May 1988 flood; e) number of species measured as saplings (⟩1m ht and ⟨ 10 cm dbh) in 177 m⟨sup⟩2⟨/sup⟩ plots; and, f) density of primate food resource trees (Ficus spp., Diospyros mespiliformis, S. madagascariensis, P. msolo, and Acacia robusta subsp. usambarensis).Pachystela msolo, with Ficus sycomorus, to mixed forests of greater species richness dominated by Sorindeia madagascariensis and Diospyros mespiliformis or by Garcinia livingstonei and Mimusops obtusifolia. Studies of community diversity in this region by Marsh (1976), Homewood (1978), Hughes (1988), Medley (1992), and Njue (1992) concur that the forest types represented by these trees vary with major floodplain land forms: flooded low-levees (e.g., Mnazini North), well-drained high levees (e.g., Mchelelo West), and clay-backwater swamps (e.g., Mnazini South; fig. 8). Some of the study areas contain a mosaic of these forest types as represented by the distribution of canopy trees in the sample quadrats, documenting a heterogeneity of site conditions over small (⟨200 m) distances (fig. 4; see Guru South, and Baomo South a).Forest regeneration, as it may influence change in these community types, directly relates to the levels of establishment and potential for recruitment by the major canopy tree species. Ficus sycomorus shows pioneer growth characteristics, with low abundances at the seedling and sapling layers in the forest study areas (fig. 5) and a large number of individuals near the river edge (fig. 4). Other studies show a record of invasion on point bars (with Populus illicifolia) and cleared openings and a record of high mean growth rates at geater thanl cm dbh/year (Hughes, 1988; Medley, 1994; Kahumbu, 1993). In contrast, Sorindeia madagascariensis appears more shade tolerant, with high levels of regeneration in the forest study areas.Low abundances of major canopy species in the sub-canopy layer suggest a delay will occur in the closure of the canopy following the death of large trees. This size structure may be explained by episodic establishment (sensu Hughes, 1994; Wissmar and Swanson, 1990), continuous regeneration with a low number of species reaching the sub-canopy size-class (Webb et al., 1972), or recruitment dependent on fast growth through the sub-canopy size class in canopy gaps (Hartshorn, 1980; Salo and Kalliola, 1991). Large-magnitude floods that occur infrequently may change the status for regeneration and initiate a new successional cycle. Earlier studies by Marsh (1976) and Allaway (1979) document a more open riverine forest, attributable mostly to disturbances imposed by large mammal populations. The recent population crash in elephants along the Tana River and local extinction of the black rhino have allowed for greater establishment and growth in the forest understory (Marsh, 1986). Both regional (e.g., river flows) and local (e.g., animal populations) environmental changes may partially explain the profound differences in the abundances of trees at the sapling and sub-canopy layers, and the overall size-class structure of the forests.Pachystela msolo, which is presently most abundant in the large-tree size class and occurs in nearly mono-dominant stands (e.g., Mnazini North), shows a low abundance at the subcanopy size class (Fig. 5) and the lowest relative abundance in the sapling size class (Fig. 6). Areas of significant regeneration by P. msolo were not measured or observed. Furthermore, stand-level death of mature individuals is documented in the Congolani and Baomo South forests (Marsh, 1986; Medley, 1990). These observations question the persistence of this species, at least in the short-term, as the most abundant canopy tree in the Reserve. The flooding and/or geomorphological conditions that promote the establishment of a new cohort of trees are not clearly documented.Significant associations between the size-classes of the six major canopy species show successional tendencies among the forest communities in which they dominate (Fig. 9). Ficus sycomorus is a riparian pioneer tree; it is not found invading significantly under mature trees of any other major canopy species. In contrast, Pachystela msolo and Sorindeia madagascariensis both invade under F. sycomorus canopies. While the occurrence of mono-dominant stands of P. msolo is an obvious feature in some forest patches, its low level of regeneration complicates any understanding of conditions necessary for establishment. From the association analyses, it appears that P. msolo establishes and dominates at the transition toward a high-levee forest with S. madagascariensis and Diospyros mespiliformis. In view of the low regeneration recorded for P. msolo, I hypothesise a trend toward greater dominance by these other two species.Figure 9, Model showing the site preferences and successional tendencies of the six major canopy tree species in the riverine forest communities. A primary successional pathway is shown along the dotted line in association with vertical development of the floodplain through high-frequency, low magnitude floods. The solid line shows disruption of the pathway, attributable to low-frequency, high magnitude floods and lateral movement of the river channel.Along this section of the Tana River, the upland vegetation is thorn-scrub. Forest is confined to a narrow band with environmental factors determining its expansion on either edge. With the exception of F. sycomorus, none of the major canopy species has a significant association with its saplings, identifying a low level of self-replacement in the studied forests. Forest community establishment and persistence appears dependent on the destabilising influence of river meanders, floods, and the continual development of new sites for colonisation. Riverine forest is necessarily sensitive to changes in the hydrologic regime, such as those associated with upstream dam construction (see Hughes, 1987; Njue, 1992). At present, five dams are constructed in the upper basin of the Tana River (Masinga, Kamburu, Gitaru, Kindaruma, and Kiambere). The much larger Three Forks Dam proposed for construction at Mutonga and Grand Falls is projected to drop groundwater levels, decrease the sediment load, and reduce seasonal flooding patterns. Unless mitigation measures, such as controlled water releases from the reservoir, are a part of the final plan, the impacts on riverine forest could be devastating (Butynski, 1995).A comparative study of Ficus sycomorus, Pachystela msolo, and Sorindeia madagascariensis further documents the ecological differences among the forest communities where these species dominate. Together, they represent the range in conditions characteristic of high- quality primate habitat (Decker, 1989; Kinnaird, 1990; Medley, 1993a). Soils vary from sandy loam under S. madagascariensis to predominantly clay soils under P. msolo, with Corresponding differences from low to relatively higher moisture availability at depth. Flooding conditions also differ significantly between the three trees. After the flood in May 1988, with a predicted two-year recurrence interval, floodwater heights varied from mean levels above 30 cm for F. sycomorus to low mean heights (6.2 cm) or an absence of flooding for six of the ten S. madagascariensis plots. Greater site variability in F. sycomorus may be attributable to relatively longer life spans and changing environmental conditions. In contrast, P. msolo shows the narrowest range of site conditions.Greatest species richness at the sapling layer occurs in association with S. madagascariensis, but the regeneration of primate food resources is highest under F. sycomorus. These results support a trend toward greater species richness from low-levee to high-levee forest communities. The regeneration of primate habitat corresponds with riverbank to low-levee positions, consequently dependent on active deposition and flooding from the river channel. Primates, through their high use of fruits from these tree species, may be important in the dispersal of seeds between the community types. Again, regional changes in the deposition of new sites for early colonisation, a predicted outcome of dam construction, and local human needs for agricultural lands will reduce the potential for early-forest development and the establishment of high-quality primate habitat.A mosaic of forest communities is represented in the Tana River Primate National Reserve. Other studies have shown a correspondence between the distribution of forest types and the characteristics of the floodplain disturbance regime (Hughes, 1988, 1990). This study quantifies regeneration patterns for the major canopy species, which vary in their relative abundances in different forests, and identifies some possible successional tendencies. The overall mechanisms of forest change and/or persistence are complicated by the coupling of allogenic (riverine), and autogenic (species establishment and recruitment) factors (Medley and Hughes, in press). A projection of community change based on current patterns of regeneration necessarily assumes no major shifts in the river position or hydrologic regime (Fig. 9). Forest community change progresses along the projected pathway in accordance with vertical floodplain development through high-frequency and low magnitude floods, as represented by the flood that occurred in May 1988. Community change would be notably disrupted by a major shift in the river position associated with low-frequency and high magnitude events (cf Hughes, 1994). The relative abundances of tree species and their size-class structures documented in this study should provide a basis for comparative studies of change through time in the Reserve. Turner (1989) states that a dynamic landscape may exhibit a stable mosaic of community types at one spatial scale, but not at another. Along the Tana, the spatial scale at which a stable mosaic of all forest communities is maintained, or at best a sustainable distribution of primate habitat, is not yet predictable.ACKNOWLEDGEMENTSWildlife Conservation International (now The Wildlife Conservation Society) sponsored the field research with additional financial assistance and language training provided by the Michigan State University African Studies Centre through a Department of Education National Resource Fellowship. The study was affiliated with the National Museums of Kenya, Institute of Primate Research, under permission granted by the Office of the President (permit # OP. 13/001/17 C 24/9). I thank the staff at the East African Herbarium, and especially Stephen Rucina, for assistance with my plant identifications, Dr. Jim Else for administrative guidance, and Bakari Mohammed Garise for his assistance during the field research. Helpful reviews of the manuscript were provided by several colleagues at the Institute of Ecosystem Studies, New York and Miami University. I also thank Dr. Edward Vanden Berghe and two anonymous reviewers with the East African Natural History Society.REFERENCESAllaway, J.D. (1979). Elephants and Their Interactions with People in the Tana River Region of Kenya. Ph. D. Dissertation, Cornell University.Andrews, P., C.P. Groves. & J.F.M. Horne (1975). Ecology of the lower Tana River floodplain (Kenya) J. E. Afr. Nat. Hist. Soc. and Nat. Mus. 151: 1-31.Brinson, M.M. (1990). Riverine forests. In: A. Lugo, S. Brown, S. & M. Brinson (eds.), Forested Wetlands. Elsevier, New York.Butynski. T. (1995). Report says dam could threaten Kenya's endangered primates. African Primates (IUCN) 1 (1): 4-8.Decker, B.S. (1989). Effects of Habitat Disturbance on the Behavioral Ecology and Demographics of the Tana River Red Colobus (Colobus badius rufomitratus). Ph. D. Dissertation, Emory University.FAO (1984). Agriclimatologica1 data for Africa. Volume 2. Countries south of the equator. FAO Plant Production and Protection Series. No. 22. FAO, Rome.Hartshorn, G.S. (1980). Neotropical forest dynamics. Biotropica 12 (Supplement): 23-30.Holdridge, L. R (1967). Life Zone Ecology. Revised edition. Tropical Science Center, San Jose, Costa Rica.Homewood, K. (1978). Feeding strategy of the Tana mangabey (Cercocebus galeritus). (Mammalia: Primates). Journal of Zoology (London) 186:375-391.Hughes, F.M.R. (1987). Conflicting uses for forest resources in the lower Tana River basin of Kenya. In: Anderson, D. & Grove, R. (eds.), Conservation in Africa. Cambridge University Press, Cambridge.Hughes, F.M R (1988). The ecology of African floodplain forests in semiarid and arid zones: a review. Journal of Biogeography 15: 127-140.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.Hughes. F.M R (1994). Environmental change, disturbance, and regeneration in semiarid floodplain forests. In: A. C. Millington and K Pye (eds.), Environmental Change in Drylands: Biogeographical and Geomorphological Perspectives. John Wiley & Sons, New York.Kahumbu, P. (1992). The Sustainability of Fig Tree (Ficus sycomorus) Harvesting in a Kenyan Reserve. MSc Thesis, University of Florida.Kinnaird, M. F. (1990). Behavioral and Demographic Responses to Habitat Change by the Tana River Crested Mangabey (Cercocebus galeritus galeritus). Ph. D. Dissertation, University of Florida.Ledec, G. (1987). Effects of Kenya's Bura Settlement Project on biological diversity and other conservation concerns. Conservation Biology 3(1): 247-258.Lind, E.M. & Morrison, M.E.S. (1974) East African Vegetation. Longman Group Limited, London.Marsh, C.W. (1976). A Management Plan for the Tana River Game Reserve. Report to the Kenya Department of Wildlife Conservation and Management, Nairobi. New York Zoological Society, Bronx, NYMarsh, C.W. (1981). Diet choice among rec colobus (Colobus badius rufomitratus) on the Tana River, Kenya. Folia Primatology 35: 147-178.Marsh, C.W. (1986). A resurvey of Tana River primates and their forest habitat. Primate Conservation 7:72-81.Medley, KE. (1990). Forest Ecology and Conservation in the Tana River National Primate Reserve, Kenya. Ph. D. Dissertation. Michigan State University.Medley, KE. (1992). Patterns of forest diversity along the Tana River, Kenya. Journal of Tropical Ecology 8: 353-371.Medley, K E. (1993a). Primate conservation along the Tana River, Kenya. An examination of the forest habitat. Conservation Biology 7: 109-121.Medley, K E. (1993b). Extractive forest resources of the Tana River National Primate Reserve, Kenya. Economic Botany 42: 171-183.Medley, K E. (1994). Identifying a strategy for forest restoration in the Tana River National Primate Reserve, Kenya. In: D. Baldwin, 1. DeLuce, and C. PIetsch (eds.), Beyond Preservation: Restoring and Inventing Landscapes. University of Minnesota Press, Minneapolis.Medley, K E., M. F. Kinnaird & B. S. Decker (1989). A survey of the riverine forests in the Wema/Hewani vicinity with reference to development and the preservation of human resources. Utafiti 2(1):1-6Medley, K.E. and F.M.R. Hughes (in press). Riverine Forests. In: T. McClanahan and T. Young (eds.), Ecosystems of East Africa and Their Conservation. Oxford University Press, Oxford.Muchena, F.N. (1987). Soils and Irrigation of Three Areas in the Lower Tana Region, Kenya. University of Wageningen.Njue, A. (1992). The Tana River Floodplain Forest, Kenya: Hydrologic and Edaphic Factors as Determinants of Vegetation Structure and Function. Ph.D. Dissertation, University of California-Davis.Pielou, E.C. (1977). Mathematical ecology. John Wiley and Sons, New York.Salo, J. S. & R.J. Kalliola (1991). River dynamics and natural forest regeneration in Peruvian Amazonia. In: A. Gómez-Pompa, T.C. Whitmore & M. Hadley (eds.), Rain Forest Regeneration and Management. UNESCO and the Parthenon Publishing Group, Paris.Salo, J. and M.E. Räsäsen (1989). Hierarchy of landscape patterns in western Amazon. In: L.B. Holm-Nielson, I.C. Nielsen & H. Balslev (eds.), Tropical Forests. Botanical Dynamics. Speciation and Diversity. Academic Press, New York.Turner, M.G. (1989). Landscape ecology; effects of pattern on process. Annual Review of Ecology and Systematics 20: 171-197.Turrill, W.B. et al. (eds.) (1952). Flora of Tropical East Africa. In fascicles. Crown Agents, London (from 1980, Balkema, Rotterdam).Webb, L.J., J.G. Tracey & W.T. Williams (1972). Regeneration and patterns in subtropical rainforest. Journal of Ecology 60:675-696.White, P.S. & S.P. Bratton (1980). After preservation: philosophical and practical problems of change. Biological Conservation 18:242-255.Wissmar, R.C. & F.J. Swanson (1990). Landscape disturbances and lotic ecotomes. In: R.J. Naiman & H. Décamps (eds.), The Ecology and Management of Aquatic-Terrestrial Ecotones. UNESCO and The Parthenon Publishing Group, Paris.Zedler, P.H. & F.G. Goff (1973). Size-association analysis of forest successional trends in Wisconsin. Ecological Monographs 45(1):79-94.

Kimberly E. Medley "Forest Regeneration in the Tana River Primate National Reserve, Kenya," Journal of East African Natural History 84(2), 77-96, (1 July 1995).[77:FRITTR]2.0.CO;2
Published: 1 July 1995
Back to Top