A

i Samples Distance (m)

"O

i Samples Distance (m)

Raphia swamp

Rinorea forest

Albizia forest

Ecotone Savanna Vegetation zones

Figure 5.3. Modern pollen rain from the drier semi-evergreen forest in Cameroon, Kandara site: (a) location map of the studied transect (samples 1 to 26); (b) comparison between arboreal pollen and leaf area index (Cournac et al., 2002). After Bremond et al., 2005 and Vincens et al., 2000.

Figure 5.3. (c) Simplified pollen diagram (after Yincens et al., 2000) (% calculated versus pollen sum including all identified taxa, Cyperaceae, and spores). * Denotes names representing the greatest probability for identification of pollen type, but for some pollen types other genera or species having the same pollen morphology cannot be excluded (pollen types <1%). Dots = pollen % < 1.

(Mimosaceae), Margaritaria discoidea (Apocynaceae) and Combretaceae. These taxa can be considered as good indicators of semi-evergreen forest under a seasonal climate, with a short dry season. However, and as expected, their pollen abundances do not exactly correspond to plant abundances in the vegetation cover. This is unavoidable due to differential pollen production. The soil samples from the Albizia forest contain a few grains of Albizia but are dominated by 50 to 70% Myrianthus arboreus (Moraceae). Near the contact with the savanna, Chaetachme aristata (Ulmaceae) dispersed more pollen (20%) than Albizia (5 to 10%) despite the latter being more strongly represented in the vegetation cover. Tetrorchidium, Ficus, Trilepisium (all Moraceae), Antidesma (Euphorbiaceae), Rubiaceae, and Sapindaceae are normal components of the semi-evergreen forest pollen rain. At contact with the Rinorea forest, Piptadeniastrum pollen was relatively abundant whereas none of the different species of Rinorea were recorded as pollen. Raphia pollen dominates near the swamp, which is also marked by an abundance of fern spores (Pteridophytes monoletes). At the northern end of the transect, close to the savanna, the abundance of Pteridium aquilinum spores indicates frequent burning. The pollen diagram illustrates the spatial colonization of a burnt savanna by a dry semi-evergreen forest. The succession starts with Moraceae, followed by Celtis, and then by Sterculiaceae/ Raphia/Pteridophyta, following dry to wetter local conditions under climatic conditions favorable to the establishment of dry semi-evergreen forest. Although this succession might be valid only for the northern semi-evergreen forest region, the results improve our understanding of tropical rainforest dynamics in the past and should stimulate similar studies for other regions.

While there is some agreement among botanists that the forest dominated by shade-intolerant tree species established from relatively open conditions several hundred years ago, the origin of savanna "islands" in the semi-evergreen forest of southeast Cameroon still remains controversial. Analyses of opal phytolith assemblages from topsoil samples of the forests and the included savanna patches at Lobeke—2°17'N, 15°42'E, 300 to 700 m in elevation, south of Kandara (Figure 5.1) in a region receiving 1,600-1,700-mm/yr precipitation—suggest stable conditions with no evidence of recent disturbances, such as fire or logging (Runge and Fimbel, 1999). Elsewhere, evidence for recent invasion of forest into the savanna is also provided from soil organic carbon isotopic studies (Guillet et al., 2001) whereas 813 C of organic matter in soil profiles provides information on previous forest development in areas now occupied by savanna (Schwartz et al., 1996). Constant re-organization of the distribution of forest and savanna patches at the limit of the drier semi- evergreen forest may be forced by the seasonal distribution of the rain through the year and variability in dry-season length. But, recent human deforestation has modified the pattern. Sorting out the respective effect of climate from that of human impact on modern vegetation deserves further thorough investigation.

5.3.3.2 South of the equator

The coastal semi-evergreen forest does not extend to southern Congo beyond 4°S latitude (White, 1983). Instead, a mosaic of forest-grassland occupies the Atlantic

Congo Rainforest Location
Figure 5.4. Location map of sites for floristic inventories and modern pollen surface soil samples collected within the semi-evergreen forests from Congo (after Elenga et al., 2000a).

coast, receiving less than 1,100 mm/yr of rainfall and with temperatures averaging 18° C during the dry season. Several types of semi-evergreen forests are present further inland, in the Mayombe massif, which reaches 730 m maximum elevation and receives 1,400 to 1,600 mm/yr of precipitation. Botanical investigation was made at 12 geographical sites (A to L), distributed between 4°05' to 4°50'S latitude and 11°45' to 12°35'E longitude (Figure 5.1). Among the investigated sites, 8 are distributed in Mayombe and 4 correspond to the coastal plain (including the mangrove) (Figure 5.4). At each site, trees with a diameter at breast height >5 cm were counted within 20 x 100-m plots (Elenga et al., 2000a). All the floristic inventories highlight the great number of tree species and the great spatial heterogeneity of their distribution. Out of a total of 620 individual trees, 352 species were identified, distributed among 47 botanical families. Of the represented tree species, the most important ones are Annonaceae (15 species), Euphorbiaceae (8 species), Caesalpiniaceae (7 species), Rubiaceae (6 species), followed by Burseraceae, Olacaceae, Sapotaceae, and Moraceae. Within all the studied plots, 110 species have tree frequencies >1%, but only 41 species were present in at least two sites where the average density of trees reached 120 individuals per 100 m2 (Elenga et al., 2000a). The forests of the Mayombe are remarkable for high floristic diversity and little overlap between tree composition at different localities. Correlation between floristic richness and climatic variables— such as insolation, water availability, or primary production—are now being explained by a new hypothesis. Palynological studies explore how such floristic diversity is reproduced in pollen assemblages extracted from 50 surface soil samples collected at the 12 sites where tree counts had been made (Figure 5.5).

5.3.4 Forests of the Mayombe

Within all the samples from the semi-evergreen forests of the Mayombe massif, c. 19 to 23 plant species make up more than 75% of the total tree counts. But, the dominant tree varies greatly within the 8 studied plots and the results of pollen analysis show that this dominance is not necessarily reproduced in the corresponding pollen assemblages. At La Tour (site A), Treculia obovoidea (Moraceae) represents 8.5% of the trees, but its pollen frequencies reach much higher percentages (30 to 60%) in all of the 10 pollen assemblages (Figure 5.5). Other trees—such as Plagiostyles africana (Euphorbiaceae), Maranthes (Chrysobalanaceae), Irvingia, Uapaca (Euphorbiaceae), common to the floristic list (Elenga et al., 2000a, table 2) and the pollen diagram (Figure 5.5)—show similar percentages both as plants and pollen. Strombosia (Olaca-ceae) are under-represented by their pollen, whereas Annonaceae and Myristicaceae have not been found as pollen. At Dimonika (site B), 11 of the 22 most abundant trees have been recognized as pollen taxa. Among them, Treculia obovoidea (Moraceae), Anisophyllea myriostricta (Anisophyllaceae), Trichoscypha (Anacardiaceae), Dacryodes (Burseraceae), Caesalpiniaceae, and Allanblachia (Guttifereae) have significant percentages, both as plant and pollen. However, other genera among Burseraceae, Apocynaceae, Clusiaceae, and Annonaceae were found weakly or not represented by their pollen. At Mindou (site C), the most abundant trees—Antho-stema (Euphorbiaceae, 13%), Dialium and Guibourtia (all Caesalpiniaceae totaling 22% trees)—are represented by significant pollen percentage values. Syzygium, not in the tree list, provided the greatest amount of pollen (20 to 50%) in samples 16 to 22 (Figure 5.5). The Mindou site, located on the humid western slopes of Mayombe, is close to the coastal forest. Proximity of Syzygium clumps in local swamps perhaps explains such high pollen percentages. At Mandzi (site D), of the most abundant trees—such as Microdesmis sp. (Pandanaceae 10%), Grewia (Tiliaceae 8%), and Tessmania sp. (Caesalpiniaceae 6%)—only Grewia is recorded by its pollen (2 to 5%). Other taxa—such as Irvingia, Pancovia (Sapindaceae), and Aidia mi-crantha (Rubiaceae)—have about the same representation both in trees and in pollen. However, associated pollen from Piptadeniastrum, Calpocalyx (Mimosaceae), Dacryodes/Santiria, Ganophyllum (Sapindaceae), Fagara (Rutaceae), Macaranga and Elaeis do not correspond to the listed trees at this site (Elenga et al., 2000a, table 5.2), a discrepancy possibly explained by the fact that pollen origin may be found outside the sampled plots used for tree counts. At les Saras (site E), Treculia—an abundant tree

Figure 5.5. Frequencies of the main pollen taxa identified within modern soil surface samples from Southern Congo (after Elenga et al., 2000a) (% calculated versus pollen sum including all identified taxa; Cyperaceae and Pteridophyta spores being unimportant, except at site B).

Figure 5.5. Frequencies of the main pollen taxa identified within modern soil surface samples from Southern Congo (after Elenga et al., 2000a) (% calculated versus pollen sum including all identified taxa; Cyperaceae and Pteridophyta spores being unimportant, except at site B).

(12.6%)—is represented by similar pollen frequencies, whereas three species of Annonaceae (13.5% distributed between Anonidium, Polyalthia, and Enantia) are totally absent as pollen. At Kitina (site F, samples 39, 40), the dominant (12.4%) Anthostema trees (Euphorbiaceae) are recorded by pollen, but at much lower percentages (<5%). Other abundant trees—such as Scytopetalum klaineanum (8% Scy-topetalaceae), Ctenolophon englerianus (6% Ctenolophonaceae), Spathandra blackeoides (6% Melastomataceae)—not identified during pollen analysis might correspond to unidentified pollen taxa, their pollen morphology not currently well-known (Figure 5.5). Dialium has about the same abundance (5%) both as plant and pollen. At Mpassi Mpassi (site G, sample 41), the two most abundant trees—Hua gabonensis (16% Huaceae) and Pancovia (7% Sapindaceae)—were not represented as pollen. Pentaclethra (Mimosoideae, 2 species, 5% trees) has the same representation as plant and as pollen, whereas the abundance of Berlinia pollen (20%) may correspond to another unknown Caesalpiniaceae pollen, since Berlinia was not in the list of counted trees. At les Bandas (site H), the dominant tree (56%) Parkia (Mimosaceae) was not found as pollen, whereas the diagram includes Pentachlethra pollen (10 %), which is clearly distinguishable from the polyad of Parkia. Except for Dacryodes and Aidia (Rubiaceae), both present in the plant record and in the pollen, there is not much overlapping of other common trees. Joint occurrences of pollen from Allophyllus, Celtis, Hymenocardia, and Combretaceae together with more Poaceae indicate much drier climatic conditions for sites located close to the drier Niari valley located in the rain shadow slope of the Mayombe massif.

5.3.5 Coastal forests

At Koubotchi (site I) the dominant (30% trees) forest component Celocaryon preussii (Myristicaceae) was not found as pollen, nor were Xylopia aethiopica (Annonaceae), Carapa (Meliaceae), Staudia (Myristicaceae), Vitex sp. (Verbenaceae). But the pollen representation of Symphonia, Uapaca, Maranthes (Chrysobalanaceae), Pycnanthus (Myristicaceae), and Sapotaceae correspond fairly well to the number of trees counted in the plot. Remarkably, Macaranga and Alchornea (<1.4% in the tree counts) are over-represented by significant pollen percentages (>20%), and Tetracera (Dillenia-ceae) is also over-represented by its pollen (>10% pollen). At Tchissanga (site J), two pollen samples (47, 48) indicate pollen values for Symphonia and Sapotaceae, Fegimanra (Anacardiaceae), and Syzygium (Myrtaceae), that are also characteristic trees of the Symphonia globulifera forest in valleys of the coastal plain (Elenga et al., 2000b). However, they lack the record of Memecylon (Melastomataceae) which accounted for 34% of the total number of trees in the same plot. At Ntombo (site K), Anthostema pollen (Euphorbiaceae) was found less abundant (5%) than in the tree cover dominated by Anthostema aubryanum (51%). But, there is good correspondence between the plant and pollen representation of Syzygium guineensis, Hallea ciliata (Rubiaceae), Elaeis guineensis (Palmae), and Alstonia congensis (Apocynaceae). Tetracera pollen (5%) was found, whereas this climber represents less than 1% in botanical inventories. This may indicate an over-representation of pollen from climbers, or an under-estimation of plant specimens, those with diameter lower than 5 cm not being counted. The mangrove (site L) is dominated by the abundance of Rhizophora, associated with Phoenix and Pandanus, also abundant plants in the plant cover.

In order to clarify the distribution of plants and pollen versus environmental factors, correspondence analysis done on all the samples compared the composition of the floristic inventories from lowland coastal forests with those of Mayombe mid-elevation forests. Indeed, there is a west/east increasing rainfall gradient from 1,100 to 1,600 mm/yr and an increased elevation between the coastal plain (sea level) and the Mayombe (700 m). But, the authors favoured an explanation involving different soil composition. They distinguished the pollen association of Syzygium, Symphonia globulifera, Phoenix, Tetracera, Sclerosperma (Arecaceae) as characterizing swamp forests (Elenga et al., 2000b). However, differences in elevation and also strong variations in precipitation could partly explain the differences in floristic and pollen composition of the coastal and Mayombe forests. Variations in the amount of rainfall are not negligible. Moreover, during the dry season, the effect of clouds on evapo-transpiration (Maley and Elenga, 1993) and that of the Benguela Current induce cooler temperature in southern Congo (Maley, 1997).

In conclusion, the study on modern pollen rain from Congo shows that arboreal pollen percentages from 70 to 90% characterize samples collected under closed forest, lower values being found within disturbed forest. These high values are obtained despite the fact that important families—such as Annonaceae (all species), most Myristicaceae (except Pycnanthus), Chrysobalanaceae, Olacaceae (except Strombosia), Clusiaceae (except Symphonia), Apocynaceae, Meliaceae, Melastoma-taceae, etc.—were poorly documented in modern pollen rain. Well-diversified pollen assemblages from southern Congo document the floristic diversity of semi-evergreen forests south of the equator. Although there is no direct overlap between pollen and tree composition, pollen assemblages clearly distinguish the different types of forest that have produced them. Associated Symphonia globulifera, Uapaca, Hallea, Dacryodes, Anthostema, Dialium, Plagiostyles, and Sapotaceae, both in the vegetation and in the corresponding modern pollen rain, characterize the dry semi-evergreen forest of Mayombe. This is enough to appreciate that the semi-evergreen forests south of the equator appear palynologically distinct from the same vegetation unit mapped north of the equator in Cameroon (sites Q and E, Figure 5.2), including Kandara. Possible explanation for such differences may be searched for in the long-term geological history and (or) in the differential ecological requirements and threshold climatic limits of the various forest trees. More investigation on this line is needed.

5.3.6 Swamp forest

Preliminary information about modern pollen rain from the inundated evergreen swamp forests of Central Congo was provided by three samples collected within the Guibourtia demeusii (Caesalpiniaceae) dominated association. Located in the central Congo Basin below 400 m in elevation (1°34'N, 17°30'E), the area receives more than 1,600 mm/yr of precipitation with a very short dry season. The results show that arboreal pollen (AP) again ranges from 75 to 90%. Pollen assemblages are dominated by Lophira (up to 60% in one sample), followed by Guibourtia, Alchornea, Macaranga, Uapaca, Combretaceae, and Myrianthus, and a few pteridophyte spores. These pollen assemblages are different from those collected from the hygrophilous evergreen forest of Cameroon and from any types of the semi-evergreen drier peripheral forest (Elenga, 1992).

The pioneer studies summarized in this chapter represent significant progress in tropical modern pollen rain of Africa. First, they clearly demonstrate that a high proportion of tree pollen can identify forest cover. Second, they indicate that the main mapping vegetation units (evergreen, semi-evergreen, and mixed) as well as secondary subdivisions in the vegetation (from Cameroon and Congo) are characterized by different pollen assemblages. Third, there is partial overlap between pollen and plant representation. In conclusion, differences in pollen (taxa composition and abundance) can be used to recognize the vegetation units and sub-units within the Guineo-Congolian rainforest, despite the lack of pollen representation of some dominant trees. Although not covering continuous climatic gradients, the results discussed bear critical information for interpreting fossil pollen data from the region. Extracting individual or associated pollen markers for all the vegetation units within the rainforest, however, requires additional and more homogenously distributed samples before being statistically valid. A complete inventory of forest types is essential because of the high diversity of rainforests. Collecting along two distinct transects—one from south to north crossing the equator to address the climatic influence of the ITCZ (Haug et al., 2001), the other from west to east, to address inland monsoon penetration—would be most valuable. This section has shown that modern pollen from the rainforest can be studied in the same way as other forests in the world. It is a long, but feasible task.

Was this article helpful?

0 0
Waste Management And Control

Waste Management And Control

Get All The Support And Guidance You Need To Be A Success At Understanding Waste Management. This Book Is One Of The Most Valuable Resources In The World When It Comes To The Truth about Environment, Waste and Landfills.

Get My Free Ebook


Post a comment