- Email: [email protected]
Late Quaternary Climate and Vegetation of the Sudanian Zone of Northeast Nigeria
Quaternary Research 58, 73–83 (2002) doi:10.1006/qres.2002.2356
Late Quaternary Climate and Vegetation of the Sudanian Zone of Northeast Nigeria Ulrich Salzmann1 Seminar f¨ur Vor- und Fr¨uhgeschichte, Arch¨aologie und Arch¨aobotanik Afrikas, J. W. Goethe-Universit¨at, Gr¨uneburgplatz 1, 60323 Frankfurt am Main, Germany E-mail: [email protected]
Philipp Hoelzmann Max-Planck-Institut f¨ur Biogeochemie, PO Box 100164, 07701 Jena, Germany
and Irena Morczinek Seminar f¨ur Vor- und Fr¨uhgeschichte, Arch¨aologie und Arch¨aobotanik Afrikas, J. W. Goethe-Universit¨at, Gr¨uneburgplatz 1, 60323 Frankfurt am Main, Germany Received March 8, 2001
INTRODUCTION ◦
The Lake Tilla crater lake in northeastern Nigeria (10 23 N, 12◦ 08 E) provides a ca. 17,000 14 C yr multiproxy record of the environmental history of a Sudanian savanna in West Africa. Evaluation of pollen, diatoms, and sedimentary geochemistry from cores suggests that dry climatic conditions prevailed throughout the late Pleistocene. Before the onset of the Holocene, the slow rise in lake levels was interrupted by a distinct dry event between ca. 10,900 and 10,500 14 C yr B.P., which may coincide with the Younger Dryas episode. The onset of the Holocene is marked by an abrupt increase in lake levels and a subsequent spread of Guinean and Sudanian tree taxa into the open grass savanna that predominated throughout the Late Pleistocene. The dominance of the mountain olive Olea hochstetteri suggests cool climatic conditions prior to ca. 8600 14 C yr B.P. The early to mid-Holocene humid period culminated between ca. 8500 and 7000 14 C yr B.P. with the establishment of a dense Guinean savanna during high lake levels. Frequent fires were important in promoting the open character of the vegetation. The palynological and palaeolimnological data demonstrate that the humid period terminated after ca. 7000 14 C yr B.P. in a gradual decline of the precipitation/evaporation ratio and was not interrupted by abrupt climatic events. The aridification trend intensified after ca. 3800 14 C yr B.P. and continued until the present. C
Palaeoecological studies presently available for West Africa outline a climatic history that is characterized by generally dry climatic conditions during the Last Glacial Maximum. Increasing humidity during the early to mid-Holocene led to a northward migration of numerous plants from more humid southern vegetation zones. An opposite trend can be recorded at the mid- to late Holocene transition, when climate became distinctly drier (e.g., L´ezine, 1989; Salzmann and Waller, 1998). It appears that Late Quaternary climatic changes were discontinuous with series of abrupt events. For the aridification trend of the middle to late Holocene, climatic fluctuations or cycles detected for several palaeoecological sites have been partly attributed to global climatic changes (e.g., Maley and Brenac, 1998; Gasse, 2000). The timing and degree of these climatic events often show strong regional variations, however, and the extent to which they reflect changes in global atmospheric circulation or local, site-specific conditions remains unclear. One major reason for the differences and uncertainties in reconstructing West African climates and vegetation is the paucity of palaeoecological sites. Whereas for the humid rain forest (e.g., Talbot and Delibrias, 1980; Maley and Brenac, 1998) and semiarid Sahelian zone (e.g., L´ezine, 1989; Salzmann and Waller, 1998) the environmental history is at least partially described by lake sediment profiles, palaeoecological data for the Sudanian and Guinean savanna zones are still scarce. It is widely assumed that these vegetation zones were formerly covered by closed forests and that human impact rather than climate caused degradation to modern savannas (e.g., Walter, 1979; Anhuf, 1997).
2002 University of Washington.
Key Words: palynology; palaeolimnology; diatoms; sedimentary geochemistry; stable isotopes; Holocene; Pleistocene; savanna; Nigeria; West Africa.
1
To whom correspondence should be addressed. 73
0033-5894/02 $35.00 C 2002 by the University of Washington. Copyright All rights of reproduction in any form reserved.
74
SALZMANN, HOELZMANN, AND MORCZINEK
However, this hypothesis is mainly based on botanical studies of modern vegetation and has not been tested by palynological investigations. The sediment core from Lake Tilla (northeast Nigeria) provides the first long-term palaeoecological record for the Sudanian zone. A detailed description and interpretation of the Holocene pollen diagram was published in Salzmann (2000). In order to detect long- and short-term climatic fluctuations as well as human impact, this study includes complementary analysis of pollen, charcoal particles, diatoms, sedimentary geochemistry, and stable isotopes. SITE DESCRIPTION
Lake Tilla (10◦ 32 N, 12◦ 08 E) is a crater lake (maar) situated about 400 km southwest of Lake Chad on the Biu Plateau (Fig. 1). The climate is marked by a strong seasonality with summer rains. Savannas are the characteristic vegetation of the Plateau whereas forests, which are defined by the absence of a continuous grass stratum (CSA, 1956), are rare. With a mean annual rainfall of about 1000 mm (Tuley, 1972), the Biu Plateau belongs to the latitudinal Sudanian vegetation zone, which is delimited to the north by the Sahel with annual rainfalls below 500 mm and to the south by the Guinean vegetation zone with an annual rainfall exceeding 1100 mm (White, 1983; see Fig. 1). With a basal diameter of 700 m, Lake Tilla is the largest crater of the Biu Plateau. Located at an altitude of 690 m above sea level (asl) the crater has the shape of an elongated bowl with a flat bottom (Fig. 1). The lake has no surface inlet or outlet and is fed mainly by groundwater. Since the early 1970s the lake has desiccated toward the end of each dry season. Acacia hockii savanna and degraded Terminalia–Combretum– Isoberlinia shrubland are the characteristic vegetation of the Biu Plateau (Tuley, 1972). The modern landscape is shaped by intensive farming, including the cultivation of Sorghum, maize, and cotton, resulting in an open parkland savanna with scattered trees and shrubs which are often of high economic use. During the dry season Lake Tilla is of substantial importance for those who raise livestock, who bring their animals to the crater for watering and grazing. METHODS
TABLE 1 Radiocarbon Ages from Lake Tilla
Profiles
Lab. no.
Age, 14 C yr B.P. (±1σ )
δ 13 C ‰
Age, cal yr B.P. (±1σ )
Depth of sample (cm)
LT97 LT97 LT97 LT97 LT97 LT97 LT97 LT97 LT97
UtC-7278 UtC-6461 UtC-7280 UtC-6460 UtC-6462 UtC-6464 UtC-7279 UtC-6463 UtC-6459
630 ± 40 1150 ± 30 1860 ± 40 2320 ± 50 4030 ± 50 5490 ± 50 8210 ± 60 9700 ± 60 11,120 ± 110
−25.2 −26.2 −24.4 −26.4 −31.8 −32.6 −32.4 −28.3 −19.6
650–560 1120–1000 1860–1740 2380–2210 4450–4600 6360–6220 9300–9080 11,170–10,890 13,270–13,010
23.5–24.5 51–52 199–200 313–314 451–452 551–552 679–680 797–798 889–890
LT95 LT95 LT95 LT95 LT95 LT95 LT95 LT95 LT95
UtC-5167 UtC-4316 UtC-4206 J-118 J-114 J-116 J-117 UtC-8511 J-115
1540 ± 30 3190 ± 40 10,300 ± 60 19610 ± 190 17,120 ± 170 17,880 ± 150 16,460 ± 130 25,300 ± 190 17,190 ± 140
−23.4 −28.2 −17.6 −20.9 −20.87 −14.64 −21.08 −20.4 −18.54
1490–1380 3450–3380 12,169–12,051 23,650–22,800 20,800–20,000 21,650–20,850 20,000–19,250 not available 20,850–20,100
82–83 289–290 777–778 1149–1151 1269–1270 1414–1416 1434–1436 1449–1451 1496–1498
Note. Radiocarbon dates (accelerator mass spectroscopy) on bulk organic carbon from Lake Tilla core LT97 and LT95. All radiocarbon dates were calibrated with the data set of OxCal 3.5 (Bronk Ramsey, 1995; Stuiver et al., 1998).
Pollen, Charcoal, and Diatom Analysis Subsamples for pollen analysis were taken every 10 cm from core LT97 using the standard preparation technique (Faegri and Iversen, 1989). Calculation of pollen percentages is based on a pollen sum that generally exceeds 800 pollen grains and includes arboreal and nonarboreal land pollen taxa (TLP). Among the 138 pollen taxa identified, trees and shrubs were grouped after their phytogeographical distribution. For the reconstruction of fire activities, microscopic charcoal particles were estimated for every pollen subsample of core LT97 by determining the surface area cover of all visible charcoal particles in volumetrically standardized pollen slides (Clark, 1982). Subsamples for diatom analysis were obtained from core LT95 approximately every 30 cm and prepared using HCl and H2 O2 . Diatoms were identified and grouped by their main habitat in planktonic and nonplanktonic taxa.
Material and Dating A 16-m-long sediment core (LT95) was taken in March 1995 on the western shore of Lake Tilla using a modified Livingstone piston corer (Fig. 1). In April 1997, when the lake had been completely desiccated, additional parallel cores (LT97) were taken from the center of the lake. However, we did not reach the base of the lacustrine sediments as an artesian groundwater horizon at ca. 9 m led to a continuous loss of sediment from the core chamber. Dating control is provided by 18 radiocarbon dates (accelerator mass spectrometry [AMS]) from organic matter (Table 1). All dates refer to uncalibrated 14 C yr with inferred ages based on linear interpolation.
Sedimentological and Geochemical Analyses TIC (total inorganic carbon), TOC (total organic carbon), TN (total nitrogen), and δ 13 C values of the organic fraction (δ 13 Corg ) were analyzed at 20-cm intervals. TIC was determined using a carbon analyzer and TOC was measured photometrically after oxidation with H2 SO4 and K2 Cr2 O7 . The C/N ratio was calculated directly from the percentages of TOC and TN. For the determination of δ 13 Corg the acid-pretreated, carbonate-free sample was combusted and CO2 was transferred to a Delta Plus XL isotope ratio mass spectrometer. For analysis of carbonates and X-ray diffractometry (minerogenic composition), subsamples
LATE QUARTERNARY CLIMATE AND VEGETATION, NIGERIA
75
FIG. 1. Location map of Lake Tilla on the Biu Plateau, modified after Directorate of Overseas Surveys for the Nigerian Government (1968), and vegetation zones of West Africa after White (1983).
76
SALZMANN, HOELZMANN, AND MORCZINEK
were obtained approximately every 70 cm. The Ca, Mg, and Sr contents of the bulk carbonates were analyzed by measuring a 3% HCl-dilution with an inductively coupled plasma-optical emission (ICP-OES). Qualitative and semiquantitative analyses of mineralogical compounds of the bulk samples were examined by X-ray powder diffraction analyses. The minerals contributing to TIC provide information about the brine saturation from which they precipitated. High Mg/Ca molar ratios and the existence of high Mgcalcites (>4–30 mole % of MgCO3 ) or Ca-dolomite (Ca0.5–0.7 Mg0.5–0.3 (CO3 )2 ) point to higher evaporation. Authigenic aragonite shows a high uptake-rate of Sr (F¨uchtbauer, 1988) that influences the Sr/Ca molar ratio and points to higher salinity of the palaeo-lake water. However, these interpretations were only be applied to samples with a TIC content of >0.5% and not for the upper part of the core (811–35 cm). For the same reason δ 13 Ccarb and δ 18 Ocarb values were determined only between 1500 to 811 cm where the TIC content exceeded 0.5%. CO2 was prepared from organic-free samples by treatment with 100% H3 PO4 and transferred to a MAT 252 isotope ratio mass spectrometer. As the Ca-dolomite content remains low in all samples (<100 counts/sec), no corrections for the fractionation coefficient were made. δ 13 Ccarb , δ 18 Ocarb , and δ 13 Corg values are given relative to the Vienna-Peedee Belemnite (V-PDB) standard. RESULTS AND INTERPRETATION
Radiocarbon Chronology The age–depth correlation, which is based for the Holocene on nine AMS dates from core LT97 (Fig. 2), shows sedimentation
FIG. 2. Age vs. depth plot. Sediment accumulation rates for the Holocene sequence are based on linear interpolation of 14 C dates from core LT97. The numbers are accumulation rates in cm/yr for each portion of the curve delimited by radiocarbon dates.
rates to be constant apart from a distinct increase between ca. 2500 and 1100 14 C yr B.P. For the Pleistocene sequence, the radiocarbon dates display a more complex pattern. Abrupt changes in lithology point to discontinuities in sedimentation (Fig. 3). However, the biostratigraphical and geochemical trends within the Pleistocene profile suggest that the sediments are in situ and not extensively reworked as might be suggested by the inverse radiocarbon dates. Given the very low organic carbon and high carbonate content of the Pleistocene sediments, we assume contamination has occurred. The detrital input of organic material from the slopes during low lake levels and/or the input of “old” aqueous carbon by groundwater seem most likely. Hence, a number of the radiocarbon ages are probably too old. Unfortunately the low carbon content and lack of pollen or macrofossils hinders the establishment of a more precise chronology for the Pleistocene. However, comparison with other West African palaeo-records, as discussed later in this paper, suggests that the Lake Tilla core goes back to the period of the Last Glacial Maximum. Therefore, the radiocarbon dates at 1435 cm (16,460 ± 130 14 C yr B.P.) and 1497 cm (17,190 ± 140 14 C yr B.P.) seem to be the most reliable. Late Pleistocene Sequence: ca. 17,000 14 C yr B.P.–ca. 12,000 14 C yr B.P. Sediment zone LTS-I to LTS-III, diatom zone LTD 1a and LTD1b. The late Pleistocene sediment sequence (1500–900 cm), which consists mainly of sandy silts with gravel, is characterized by very low TOC (<1%) and C/N ratios varying between 6.7 and 18.8, as well as strongly fluctuating carbonate contents (Fig. 3). Covariation in δ 18 Ocarb and δ 13 Ccarb values suggests that the stable isotope curves reflect hydrological changes within the lake. Pollen are not preserved in the Pleistocene section but constantly high δ 13 Corg values indicate C4dominated grass savannas on the Biu Plateau throughout the late Pleistocene. Between 1500 and 1405 cm (LTS-I) relatively low C/N ratios (7.7–11.7) suggest increasing proportions of aquatic plants. From the molar Sr/Ca ratio (0.0031–0.0037) of the carbonates oligosaline to mesosaline conditions are inferred. High numbers of nonplanktonic diatoms in particular Fragilaria and of the planktonic taxa Aulacoseira granulata var. angustissima (O. M¨uller) Simonsen at the corresponding diatom zone LTD1a (1465–1405 cm) indicate fluctuating water levels (Fig. 4). However, the dominance of Aulacoseira granulata (Ehrenberg) Simonsen with a maximum at 1450 cm and a low δ 18 Ocarb value suggest that during this period Lake Tilla might have reached significantly higher levels than throughout the following late Pleistocene section. The basal silts and clays are covered by a 1-m-thick layer (1405–1300 cm) of almost pure carbonates (high-Mg calcite, Ca dolomite, aragonite). Positive values of δ 18 Ocarb and δ 13 Ccarb point to a reduced precipitation/evaporation ratio with high ionic concentration of the palaeo-lake water (oligo- to mesosaline) as inferred from the molar Sr/Ca ratio (0.0034–0.0046). The C/N
LATE QUARTERNARY CLIMATE AND VEGETATION, NIGERIA
77
FIG. 3. Stratigraphy, sedimentary geochemistry, stable isotopes, and radiocarbon dates for the Holocene (LT97) and Pleistocene (LT95) sediments of Lake Tilla. (See Fig. 6 for a description of sedimentary units). TOC = total organic carbon; TIC = total inorganic carbon; V-PDB = Vienna-Peedee Belemnite.
ratios range from 12.1 to 18.8 and reach highest values of the entire sequence indicating a high input of terrestrial plants at low water depths. A thin clayey layer rich in diatoms between 1285 and 1265 cm (LTS-III) suggests a return to slightly higher lake levels. However, the dominance of nonplanktonic taxa at diatom zone LTD-1b (Fig. 4) implies that the water table remained permanently low. Between 1300 and 1047 cm, fluctuating curves of C/N, δ 13 Corg , 13 δ Ccarb , δ 18 Ocarb , and TIC suggest changing water depths and varying ionic concentrations of the palaeo-lake water. High δ 13 Ccarb and positive δ 18 Ocarb values represent low precipitation/evaporation ratios and coincide with the formation of Cadolomite, whereas lower δ 13 Ccarb and δ 18 Ocarb values are associated with low-Ca dolomite contents and precipitation of low-Mg calcite, pointing to elevated P/E ratios. High quartz grain content indicates increased input from the surrounding slopes and infers lower water depths. Between 1047 and 937 cm (LTS-III), the lower and more constant C/N ratio (average 7.8), lower isotopic composition of the carbonates (−2.64 to −2.16 ‰ δ 18 Ocarb ), and Sr/Ca molar ratios between 0.0026 and 0.0042 point to a slightly deeper water depth, although the high amount of quartz and basaltic gravel invoke an increased input of detrital material.
Late Pleistocene–Holocene Transition: ca. 12,000 14 C yr B.P.–10,000 14 C yr B.P. Sediment zone LTS-IV, pollen zone LT-1, diatom zone LTD-2. Alternating clay and sandy silt layers between 900 and 835 cm (LTS-IV, Fig. 3) and an increase in planktonic diatoms in zone LTD-2 (Fig. 4) characterize the transition of Lake Tilla to a permanent freshwater lake. However, increasing C/N ratios, which point to a higher input of terrestrial plant material, and an alternating abundance of planktonic (e.g., Aulacoseira granulata var. angustissima) and nonplanktonic taxa (e.g., Fragilaria ulna (Nitzsch) Lange-Bertalot) suggest relatively low and fluctuating water levels. Large quantities of grass pollen and only few arboreal taxa indicate an open grassland on the Biu Plateau at the end of the late Pleistocene (pollen zone LT-1, Fig. 5). Increasing percentages of herbs and Typha suggest a rise in littoral marginal zones inside the crater caused by a significant lowering of lake levels between ca. 10,900 and 10,500 14 C yr B.P. Evidence for a short and more arid period shortly before the onset of the Holocene is also given by the absence of tree pollen, the high Sr/Ca (0.0052) and Mg/Ca (1.27) molar ratios as well as a relatively high δ 18 Ocarb value of +0.1 ‰ at 840 cm
78
SALZMANN, HOELZMANN, AND MORCZINEK
FIG. 4.
Summary diatom diagram for the Holocene and Late Pleistocene sequences of core LT95. Plus sign (+) indicates values below 5%.
which can all be attributed to a lower precipitation/evaporation ratio. Holocene (<10,000 14 C yr B.P.) Sediment zone LTS-V, pollen zone LT-1 to LT-3, diatom zone LTD-3 to LTD5. Humid conditions throughout the Holocene
are indicated by the deposition of black organic lake mud with high TOC and low TIC values. The Holocene diatom spectra, dominated by the circumneutral to alkaliphilous species A. granulata (pH >7), indicate stable hydrological conditions (Servant and Servant-Vildary, 1980). High and relatively undeviating quantities of charred grass epidermis suggest regular savanna fires throughout the entire Holocene.
FIG. 5. Summary Holocene pollen diagram (LT97) showing the percentage of main taxa with depth bars exaggerated 5× and charcoal particles content in cm2 /cm3 plotted against depth.
LATE QUARTERNARY CLIMATE AND VEGETATION, NIGERIA
79
80
SALZMANN, HOELZMANN, AND MORCZINEK
Early to Mid-Holocene (ca. 10,000 14 C yr B.P.–3800 14 C yr B.P.) Sediment zone LTS-Va and LTS-Vb, pollen zone LT-2, diatom zone LTD-3 and LTD-4. The wetter condition at the onset of the Holocene led to an increase in several Sudanian and Guinean tree taxa, notably Combretaceae, Lannea/Sclerocarya, Uapaca, Alchornea, and Cussonia, as well as a decline in herbs (pollen zone LT-1 and LT-2a, Fig. 5). Whereas humidity increased, temperatures might have been significantly lower than today, as suggested by high numbers of Olea hochstetterii (max. 15%) between ca. 10,000 and 8600 14 C yr B.P. This species is absent on the Biu Plateau today and the nearest population can be found on the Jos Plateau and Mandara mountains above 1300 m asl (White, 1983). A highly diverse, densely wooded Guinean savanna became established on the Biu Plateau between ca. 8600 and 7000 14 C yr B.P. The corresponding pollen zone LT-2b shows highest tree pollen percentages and the greatest number of taxa (Fig. 5). Although the tree cover was presumably denser than today, high percentages of grass pollen (>52%), charred particles of grass epidermis, and the presence of characteristic savanna trees such as Lophira, Cussonia, and Burkea africana clearly indicate that closed forest never became dominant. In floristic composition the mid-Holocene vegetation closely resembles the modern “Southern Guinean Savannas” (Keay, 1959) at the forest– savanna boundary about 350 km south of Lake Tilla with annual rainfall exceeding 1200 mm. The Guinean savanna tree Uapaca and the shrub Alchornea might have been a substantial part of the swamp forest fringing the mid-Holocene lake. From ca. 7000 14 C yr B.P. onward a decline in woody climbers, Pteropsida, and several Guinean tree taxa, notably Cussonia, Burkea africana, and Alchornea, can be recorded at pollen zone LT-2c indicating the onset of drier environmental conditions. After ca. 5200 14 C yr B.P., the Guinean swamp forest fringing Lake Tilla might have become more open as suggested by a distinct decrease in Uapaca percentages followed by increasing Typha and Poaceae (LT-2d). The decline in tree taxa from ca. 7000 14 C yr B.P. onward was accompanied by a change in floristic composition from a wet Southern-Guinean to a drier SudanoGuinean savanna. This is in particular shown by the successive occurrence of pollen from the characteristic Northern Guinean savanna tree Monotes kerstingii (Keay, 1959) at ca. 6000 14 C yr B.P. (LT-2c, Fig. 5) and the Sudanian tree taxa Isoberlinia at 4800 14 C yr B.P. (LT-2d). Fluctuating percentages of A. granulata and A. distans var. africana (LTD-4, Fig. 4) implies contemporaneous changes in water chemistry which are accompanied by a decline in TOC and a trend toward heavier δ 13 Corg values, suggesting an increased input of terrestrial C4 organic material (LTS-Vb, Fig. 3). Mid- to Late Holocene (<3800 14 C yr B.P.) Sediment zone LTS-Vc, pollen zone LT-3, diatom zone LTD-5. A shift toward drier environmental conditions is recorded for the transition to the late Holocene after ca. 3800 14 C
yr B.P. The sharp decrease in TOC values and a rise in δ 13 Corg values (LTS-Vc, Fig. 3) indicate increasing importance of C4savanna grasses. Consistently lower C/N ratios at ca. 10.2 are recorded below 475 cm, suggesting that algae became the predominant source of organic matter from about 4300 14 C yr B.P. onward. After ca. 3800 14 C yr B.P. an open Sudanian savanna with woody taxa from the drier Sahelian zone such as Balanites, Commiphora, and Salvadora became established (pollen zone LT-3; Fig. 5). Like today, numerous herbs (e.g., Typha, Amaranthaceae/Chenopodiaceae, Asteroideae, Mitracarpus) colonized the seasonally flooded shorelines of the shrinking lake. The former Guinean swamp forest with Uapaca and Alchornea was replaced by a fringing Sudanian woodland with Celtis cf. integrifolia, and Ficus (pollen zone LT-3a). An increase in quartz reflects a higher detrital input. The reoccurrence of the diatom Aulacoseira granulata var. angustissima (LTD-5a, Fig. 4), the regular presence of the halophilous indicator species Cyclostephanos dubius (Fricke) Round (Krammer and LangeBertalot, 1991), and a Mg-calcite formation relatively rich in Sr and Mg after ca. 3000 14 C yr B.P. confirm a further aridification that was accompanied by a rise in salinity of the lake water. Dry, grayish, calcareous silts with high Sr/Ca (0.0042) and Mg/Ca (1.58) molar ratios at the top of the sequence and rising pollen percentages of fringing herbs, in particular Cyperaceae and Typha, at pollen zone LT-3c document an advanced silting up of Lake Tilla after 1100 14 C yr B.P. The distinct increase in sedimentation rates between ca. 2500 and 1100 14 C yr B.P. (Fig. 2) is difficult to interpret. Gradually increasing δ 13 Corg and TIC values with precipitation of high-Mg calcite as well as rising Mg/Ca and Sr/Ca ratios point to a continuation of the aridification trend which might have enforced the soil erosion on the crater slopes (sediment zone LTS-Vc, Fig. 3). However, although the pollen data show a slight decline in tree pollen, the undeviating percentages of fringing herbs and Sahelian tree taxa (Fig. 5, pollen zone LTD-3b) contradict the interpretation of a rapidly increasing aridity after 2500 14 C yr B.P. Apart from climatic change, human impact by early shepards and farmers, for whom the lake became an important water reservoir during the dry season, may have been responsible. As it is today, trampling and grazing by livestock may have enforced slope erosion while the reduction in tree pollen percentages reflects anthropogenic clearing. DISCUSSION AND CONCLUSIONS
Late Quaternary Climatic Changes With a dry late Pleistocene, a humid early to mid-Holocene, and a subsequent aridification, Lake Tilla confirms the general climatic trends previously described by other palaeoecological records from northern tropical and subtropical Africa (e.g., L´ezine, 1989; Dupont and Agwu, 1992; Gasse, 2000; Salzmann and Waller, 1998). The few late Pleistocene records presently available for West Africa indicate that the period of maximum aridity associated with low temperatures took place between
LATE QUARTERNARY CLIMATE AND VEGETATION, NIGERIA
ca. 20,000 and 15,000 14 C yr B.P. (e.g., Servant and ServantVildary, 1980; Maley and Brenac, 1998). At all sites a slow return to wetter climatic conditions, indicated by rising lake levels and a spread of rain forest in West Cameroon and Ghana, can be detected from ca. 15,000 to 14,000 14 C yr B.P. onward (Maley, 1996). The correlation with other west African sites suggests that the Pleistocene sequence of Lake Tilla, for which precise radiocarbon datings are not available, goes back to at least 17,000 14 C yr B.P. As is the case with the Pleistocene records from the West African rain-forest zone (e.g., Talbot and Johannessen, 1992), the Lake Tilla profile indicates that the Last Glacial Maximum and the transition to the Holocene humid period was evidently interrupted by several short arid and humid intervals. Of particular note is a distinct dry event at Lake Tilla between ca. 10,900 and 10,500 14 C yr B.P. (Fig. 6). Comparable aridity events, generally placed in the interval 11,000–10,000 14 C yr B.P., have been recorded at numerous other sites in tropical
81
Africa and were attributed to the climatic change of the Younger Dryas period (e.g., Talbot and Johannessen, 1992; Maley and Brenac, 1998; Salzmann and Waller, 1998). The onset of the Holocene humid period is marked at nearly all West African sites by an abrupt increase in lake levels at ca. 10,000 14 C yr B.P. At Lake Tilla, wetter climatic conditions led to a rapid spread of Guinean and Sudanian tree taxa into the formerly open grass savanna. Temperatures might have been lower than today as indicated by the dominance of the mountain olive Olea hochstetterii until ca. 8600 14 C yr B.P. This coincides with palynological results from Lake Bosumtwi in Ghana, where the appearance of montane elements at 400 m asl was interpreted by Maley (1996) as an indicator of low temperatures. However, it has to be considered that the early Holocene vegetation on the Biu Plateau which consisted of a mixture of Guinean lowland savanna with a few mountainous taxa apparently has no modern analogue. The rapid spread of the mountain olive on the sparsely vegetated Biu Plateau and its dominance during the
FIG. 6. Summary diagram showing the woody pollen taxa grouped after their main phytogeographical distribution, the percentages of planktonic vs. nonplanktonic diatoms and the main geochemical data for Lake Tilla plotted against uncalibrated and calibrated radiocarbon dates. The palaeoclimate curve shows trends in precipitation/evaporation since approximately the last 17,000 14 C yr B.P. as deduced from reconstructed lake level and vegetation changes. Dashed line indicate sequences were the interpretation is obscured by insufficient radiocarbon datings or possible hiati.
82
SALZMANN, HOELZMANN, AND MORCZINEK
early Holocene might have been also caused by a preemption of sites. The early to mid-Holocene humid period culminated between ca. 8500 and 7000 14 C yr B.P. Dense Guinean savannas became established on the Biu Plateau, supporting the assumption of a latitudinal shift of vegetation zones of West Africa of about 400 km northward in response to climatic changes (e.g., Lezine, 1989; Dupont and Agwu, 1992). The maximum northward shift of vegetation was accompanied by high water levels of Lake Tilla. Palynological and palaeolimnological data suggest that the optimum of the Holocene humid period had been passed in northeastern Nigeria after ca. 7000 14 C yr B.P. The course of the termination of the African Holocene humid period and transition to modern more arid conditions is subject of controversy. Abrupt dry climatic events have been detected at several sites in tropical and subtropical West Africa, in particular around 7500, 4500, 3700, and 2800 14 C yr B.P., each followed by a return to wetter climates (e.g., Servant and Servant-Vildary, 1980; Talbot and Delibrias, 1980; Maley and Brenac, 1998; Gasse, 2000). However, as these events show a high variability in timing, it remains unclear in how far they represent short term episodes with different site related amplitudes or rather changes in global atmospheric circulation. At Lake Tilla the humid period terminated in a gradual decline in precipitation from ca. 7000 14 C yr B.P. onward, resulting in a floristic change from Guinean to drier Sudanian savannas and a parallel lowering of lake levels (Fig. 6). At the Manga Grasslands situated about 300 km north of Lake Tilla in the Sahelian zone of Nigeria, a gradual climatic deterioration set in at ca. 5000 14 C yr B.P. (Salzmann and Waller, 1998; Street-Perrott et al., 2000). The multiple pollen profiles taken from several dune depressions of the Manga Grasslands revealed differences in timing of vegetational changes of up to 1000 yr which might have been caused by specific local conditions such as variations in the relative height of the water table (Salzmann and Waller, 1998). Similar reasons might have also been responsible for the disparity between Lake Tilla and the Manga Grasslands where the end of the mid-Holocene humid period is recorded in the Sahelian pollen diagrams about 2000 yr later. The profiles from both regions give no evidence for a short dry spell around 7500 14 C yr B.P. which has been detected in numerous lake records from the northern monsoon domain of Africa (e.g., Servant and Servant-Vildary, 1980; Talbot and Delibrias 1980; Gasse 2000) and was partly connected with a short but distinct cooling event in the North Atlantic and Greenland (Alley et al., 1997; Gasse, 2000). However, the inference that the period between ca. 7000 and 7500 14 C yr B.P. was not a short event but rather marks the end of most humid conditions in the northern monsoon domain of Africa followed by a gradual decline in precipitation and evaporation was recently supported by a highresolution eolian sediment record taken from a marine core off Cap Blanc, Mauritania (DeMenocal et al., 2000). At Cap Blanc the gradual mid-Holocene aridification culminated in a very abrupt onset of much more arid condi-
tions at ca. 4800 14 C yr B.P. (DeMenocal et al., 2000). At Lake Tilla the gradual trend toward aridity seems to have intensified about 1000 yr later. At ca. 3800 14 C yr B.P. a distinct drop in water levels of Lake Tilla indicates a shift toward drier climatic conditions. A distinct biostratigraphical and sedimentological change can be recorded at nearly all West African sites for the beginning of the Late Holocene between ca. 4000–2000 14 C yr B.P. (e.g., Lezine, 1989; Ballouche and Neumann, 1995; Maley and Brenac, 1998; Salzmann and Waller, 1998). The strong temporal variations between sites suggest that this change was caused by an intensification of a general trend with different regional amplitudes rather than by one abrupt climatic event. In the Sahel of Nigeria (Salzmann and Waller, 1998) and Burkina Faso (Ballouche and Neumann, 1995) the establishment of the modern vegetation coupled with a distinct drop in lake levels occurred between ca. 3300 and 3000 14 C yr B.P. In western Senegal a similar change occurred about 1000 yr later (Lezine, 1989). In the rain-forest zone of western Cameroon an abrupt opening of the vegetation took place at ca. 2500 14 C yr B.P., which was followed by a return to wetter climates after 2000 14 C yr B.P. (Maley and Brenac, 1998). In northeastern Nigeria the aridification continued throughout the late Holocene resulting in the recent desiccation of Lake Tilla. The Role of Climate Change and Human Impact on the Distribution of Savannas and Forests on the Biu Plateau Competition for limited water resources is often cited as being a major factor which controls the distribution of savanna and forest in tropical West Africa (e.g., Walter, 1979; Cole, 1986). Walter (1979) postulated that natural savannas as a stable “climatic climax” are confined to regions with summer rains and a mean annual rainfall of 300–500 mm. This hypothesis was supported by studies on modern vegetation in protected savanna areas which suggest that the Sudanian zone with a mean annual rainfall of 500–1100 mm receives enough precipitation to promote the establishment and spread of dense forests (e.g., Swaine, 1992). Human impact has been emphasized as being responsible for the progressive degradation of these closed forests to modern savannas by clearing, grazing, agriculture, and iron smelting—a process which might have started during the midto late Holocene (e.g., Anhuf, 1997). On the basis of the new palynological results from Lake Tilla, the hypothesis of the anthropogenic origin of Sudanian savannas during the Holocene must be rejected. High percentages of grass pollen, the regular occurrence of savanna fires, and the presence of characteristic savanna trees clearly indicate that savannas were always present on the plateau and never replaced closed forests. The charcoal particle curve of Lake Tilla supports the assumption that frequent fires constitute a major agent which maintains the open character of the savanna vegetation (Swaine, 1992). The constancy of the charcoal curve provides no evidence whether these fires were mainly of natural origin or induced by early hunters, who may have started them long before the onset of the Holocene.
LATE QUARTERNARY CLIMATE AND VEGETATION, NIGERIA
Today most of the West African savannas are cultural landscapes which have been strongly shaped by humans. Archaeological and archaeobotanical investigations suggest that in northeastern Nigeria human impact increased from at least 3000 14 C yr B.P. onward (Breunig et al., 1996). Nevertheless, although different palaeoecological methods were applied, the Lake Tilla record neither supports the archaeological findings nor does it provide evidence at which point the anthropogenic transformation to the modern “farmed parklands” have started. The lack of unambiguous indicators even for the late Holocene demonstrates the difficulties of detecting human activity within the open vegetation of a savanna.
ACKNOWLEDGMENTS The research is part of the West African Savanna project (SFB 268) at the University of Frankfurt, which is funded by the German Research Foundation (DFG). Special thanks go to Alhaji Yamta Thetur, Mohammed Bata, R. Byer, J. Maley, K. Neumann, G. Schenk, and E. Schulz . We also thank H. F. Lamb and an anonymous reviewer for their constructive comments.
REFERENCES Alley, L. J., Mayewski, P. A., Sowers, T., Stuiver, M., Taylor, K. C., and Clark, P. U. (1997). Holocene climatic instability: A prominent, widespread event 8200 years ago. Geology 25, 483–486. Anhuf, D. (1997). Paleovegetation in West Africa for 18.000 B.P. and 8.500 B.P. Eiszeitalter und Gegenwart 47, 112–119. Ballouche, A., and Neumann, K. (1995). A new contribution to the Holocene vegetation history of the West African Sahel: Pollen from Oursi, Burkina Faso and charcoal from three sites in northeast Nigeria. Vegetation History and Archaeobotany 4, 31–39. Breunig, P., Neumann, K., and Van Neer, W. (1996). New research on the Holocene settlement and environment of the Chad basin in Nigeria. African Archaeological Review 13, 111–145. Bronk Ramsey, C. (1995). Radiocarbon calibration and analysis of stratigraphy: The OxCal program. Radiocarbon 37, 425–430. Clark, R. L. (1982). Point count estimation of charcoal in pollen preparation and thin sections of sediments. Pollen et Spores 24, 523–535. Cole, M. M. (1986). “The Savannas. Biogeography and Geobotany.” Academic Press, London. CSA (Conseil Scientifique pour l’Afrique au Sud du Sahara). (1956). Phytog´eographie Phytogeography. Yangambi Publication 53. Yangambi, DR of Congo. DeMenocal, P., Ortiz, J., Guilderson, T., Adkins, J., Sarnthein, M., Baker, L., and Yarusinsky, M. (2000). Abrupt onset and termination of the African Humid Period. Quaternary Science Reviews 19, 347–361. Directorate of Overseas Surveys for the Nigerian Government. (1968). Nigeria 1:50.000. Biu Sheet 133 S.W. Lagos, London. Dupont, L. M., and Agwu, C. O. C. (1992). Latitudinal shifts of forest and savanna in NW Africa during the Brunhes chron: Further marine palyno-
83
logical results from site M 16415 (9◦ N, 19◦ W). Vegetation History and Archaeobotany 1, 163–175. Faegri, K., and Iversen, J. (1989). “Textbook of Pollen Analysis.” Wiley, New York. F¨uchtbauer, H. (Ed.) (1988). “Sedimente und Sedimentgesteine II. Teil.” Schweizerbart, Stuttgart. [In German] Gasse, F. (2000). Hydrological changes in the African tropics since the Last Glacial Maximum. Quaternary Science Reviews 19, 189–211. Keay, R. W. J. (1959). “An outline of Nigerian vegetation.” Government Printer, Lagos. Krammer, K., and Lange-Bertalot, H. (1991). “S¨usswasserflora von Mitteleuropa, Bacillariophyceae, 3. Teil: Centrales, Fragilariaceae, Eunotiaceae.” Fischer, Stuttgart. [In German] L´ezine, A.-M. (1989). Late Quaternary vegetation and climate of the Sahel. Quaternary Research 2, 317–334. Maley, J. (1996). The African rain forest—Main characteristics of changes in vegetation and climate from the Upper Cretacous to the Quaternary. Proceedings of the Royal Society of Edinburgh 104B, 31–74. Maley, J., and Brenac, P. (1998). Vegetation dynamics, palaeoenvironments and climatic change in the forests of western Cameroon during the last 28000 years B.P. Review of Palaeobotany and Palynology 99, 157–187. Salzmann, U. (2000). Are savannas degraded forests?—A Holocene pollen record from the Sudanian zone of NE-Nigeria. Vegetation History and Archaeobotany 9, 1–15. Salzmann, U., and Waller, M. (1998). The Holocene vegetational history of the Nigerian Sahel based on multiple pollen profiles. Review of Palaeobotany and Palynology 100, 39–72. Servant, M., and Servant-Vildary, S. (1980). L’environnement Quaternaire du bassin du Tchad. In “The Sahara and the Nile” (M. A. J. Williams and H. Faure, Eds.), pp. 133–162. Balkema, Rotterdam. [In French] Street-Perrott, F. A., Holmes, J. A., Waller, M. P., Allen, M. J., Barber, N. G. H., Fothergill, P. A., Harkness, D. D., Ivanovich, M., Kroon, D., and Perrot, R. A. (2000). Drought and dust deposition in the West African Sahel: A 5500-year record from Kajemarum Oasis, northeastern Nigeria. The Holocene 10, 293– 302. Stuiver, M., Reimer, P. J., Bard, E., Beck, J. W., Burr, G. S., Hughen, K. A., Kromer, B., McCormac, G., van der Plicht, J., and Spurk, M. (1998). INTCAL98 radiocarbon age calibration, 24000–0 cal B.P. Radiocarbon 40, 1041–1083. Swaine, M. D. (1992). Characteristics of dry forest in West Africa and the influence of fire. Journal of Vegetation Science 3, 365–374. Talbot, M. R., and Delibrias, G. (1980). A new Pleistocene-Holocene water-level curve for Lake Bosumtwi, Ghana. Earth and Planetary Science Letters 47, 336–344. Talbot, M. R., and Johannessen, T. (1992). A high resolution palaeoclimatic record for the last 27,500 years in tropical West Africa from the carbon and nitrogen isotopic composition of organic matter. Earth and Planetary Science Letter 110, 23–37. Tuley, P. (Ed.). (1972). “The Land Resources of North East Nigeria Volume 1: The Environment.” Foreign and Commonwealth Office Overseas Development Administration, Land Resources Division, Surbiton. Walter, H. (1979). “Vegetation of the Earth and Ecological Systems of the GeoBiosphere.” Springer, Berlin. White, F. (1983). “The Vegetation of Africa. A Descriptive Memoir to Accompany the Unesco/AETFAT/UNSO Vegetation Map of Africa.” Unesco, Paris.
Late Quaternary Climate and Vegetation of the Sudanian Zone of Northeast Nigeria Ulrich Salzmann1 Seminar f¨ur Vor- und Fr¨uhgeschichte, Arch¨aologie und Arch¨aobotanik Afrikas, J. W. Goethe-Universit¨at, Gr¨uneburgplatz 1, 60323 Frankfurt am Main, Germany E-mail: [email protected]
Philipp Hoelzmann Max-Planck-Institut f¨ur Biogeochemie, PO Box 100164, 07701 Jena, Germany
and Irena Morczinek Seminar f¨ur Vor- und Fr¨uhgeschichte, Arch¨aologie und Arch¨aobotanik Afrikas, J. W. Goethe-Universit¨at, Gr¨uneburgplatz 1, 60323 Frankfurt am Main, Germany Received March 8, 2001
INTRODUCTION ◦
The Lake Tilla crater lake in northeastern Nigeria (10 23 N, 12◦ 08 E) provides a ca. 17,000 14 C yr multiproxy record of the environmental history of a Sudanian savanna in West Africa. Evaluation of pollen, diatoms, and sedimentary geochemistry from cores suggests that dry climatic conditions prevailed throughout the late Pleistocene. Before the onset of the Holocene, the slow rise in lake levels was interrupted by a distinct dry event between ca. 10,900 and 10,500 14 C yr B.P., which may coincide with the Younger Dryas episode. The onset of the Holocene is marked by an abrupt increase in lake levels and a subsequent spread of Guinean and Sudanian tree taxa into the open grass savanna that predominated throughout the Late Pleistocene. The dominance of the mountain olive Olea hochstetteri suggests cool climatic conditions prior to ca. 8600 14 C yr B.P. The early to mid-Holocene humid period culminated between ca. 8500 and 7000 14 C yr B.P. with the establishment of a dense Guinean savanna during high lake levels. Frequent fires were important in promoting the open character of the vegetation. The palynological and palaeolimnological data demonstrate that the humid period terminated after ca. 7000 14 C yr B.P. in a gradual decline of the precipitation/evaporation ratio and was not interrupted by abrupt climatic events. The aridification trend intensified after ca. 3800 14 C yr B.P. and continued until the present. C
Palaeoecological studies presently available for West Africa outline a climatic history that is characterized by generally dry climatic conditions during the Last Glacial Maximum. Increasing humidity during the early to mid-Holocene led to a northward migration of numerous plants from more humid southern vegetation zones. An opposite trend can be recorded at the mid- to late Holocene transition, when climate became distinctly drier (e.g., L´ezine, 1989; Salzmann and Waller, 1998). It appears that Late Quaternary climatic changes were discontinuous with series of abrupt events. For the aridification trend of the middle to late Holocene, climatic fluctuations or cycles detected for several palaeoecological sites have been partly attributed to global climatic changes (e.g., Maley and Brenac, 1998; Gasse, 2000). The timing and degree of these climatic events often show strong regional variations, however, and the extent to which they reflect changes in global atmospheric circulation or local, site-specific conditions remains unclear. One major reason for the differences and uncertainties in reconstructing West African climates and vegetation is the paucity of palaeoecological sites. Whereas for the humid rain forest (e.g., Talbot and Delibrias, 1980; Maley and Brenac, 1998) and semiarid Sahelian zone (e.g., L´ezine, 1989; Salzmann and Waller, 1998) the environmental history is at least partially described by lake sediment profiles, palaeoecological data for the Sudanian and Guinean savanna zones are still scarce. It is widely assumed that these vegetation zones were formerly covered by closed forests and that human impact rather than climate caused degradation to modern savannas (e.g., Walter, 1979; Anhuf, 1997).
2002 University of Washington.
Key Words: palynology; palaeolimnology; diatoms; sedimentary geochemistry; stable isotopes; Holocene; Pleistocene; savanna; Nigeria; West Africa.
1
To whom correspondence should be addressed. 73
0033-5894/02 $35.00 C 2002 by the University of Washington. Copyright All rights of reproduction in any form reserved.
74
SALZMANN, HOELZMANN, AND MORCZINEK
However, this hypothesis is mainly based on botanical studies of modern vegetation and has not been tested by palynological investigations. The sediment core from Lake Tilla (northeast Nigeria) provides the first long-term palaeoecological record for the Sudanian zone. A detailed description and interpretation of the Holocene pollen diagram was published in Salzmann (2000). In order to detect long- and short-term climatic fluctuations as well as human impact, this study includes complementary analysis of pollen, charcoal particles, diatoms, sedimentary geochemistry, and stable isotopes. SITE DESCRIPTION
Lake Tilla (10◦ 32 N, 12◦ 08 E) is a crater lake (maar) situated about 400 km southwest of Lake Chad on the Biu Plateau (Fig. 1). The climate is marked by a strong seasonality with summer rains. Savannas are the characteristic vegetation of the Plateau whereas forests, which are defined by the absence of a continuous grass stratum (CSA, 1956), are rare. With a mean annual rainfall of about 1000 mm (Tuley, 1972), the Biu Plateau belongs to the latitudinal Sudanian vegetation zone, which is delimited to the north by the Sahel with annual rainfalls below 500 mm and to the south by the Guinean vegetation zone with an annual rainfall exceeding 1100 mm (White, 1983; see Fig. 1). With a basal diameter of 700 m, Lake Tilla is the largest crater of the Biu Plateau. Located at an altitude of 690 m above sea level (asl) the crater has the shape of an elongated bowl with a flat bottom (Fig. 1). The lake has no surface inlet or outlet and is fed mainly by groundwater. Since the early 1970s the lake has desiccated toward the end of each dry season. Acacia hockii savanna and degraded Terminalia–Combretum– Isoberlinia shrubland are the characteristic vegetation of the Biu Plateau (Tuley, 1972). The modern landscape is shaped by intensive farming, including the cultivation of Sorghum, maize, and cotton, resulting in an open parkland savanna with scattered trees and shrubs which are often of high economic use. During the dry season Lake Tilla is of substantial importance for those who raise livestock, who bring their animals to the crater for watering and grazing. METHODS
TABLE 1 Radiocarbon Ages from Lake Tilla
Profiles
Lab. no.
Age, 14 C yr B.P. (±1σ )
δ 13 C ‰
Age, cal yr B.P. (±1σ )
Depth of sample (cm)
LT97 LT97 LT97 LT97 LT97 LT97 LT97 LT97 LT97
UtC-7278 UtC-6461 UtC-7280 UtC-6460 UtC-6462 UtC-6464 UtC-7279 UtC-6463 UtC-6459
630 ± 40 1150 ± 30 1860 ± 40 2320 ± 50 4030 ± 50 5490 ± 50 8210 ± 60 9700 ± 60 11,120 ± 110
−25.2 −26.2 −24.4 −26.4 −31.8 −32.6 −32.4 −28.3 −19.6
650–560 1120–1000 1860–1740 2380–2210 4450–4600 6360–6220 9300–9080 11,170–10,890 13,270–13,010
23.5–24.5 51–52 199–200 313–314 451–452 551–552 679–680 797–798 889–890
LT95 LT95 LT95 LT95 LT95 LT95 LT95 LT95 LT95
UtC-5167 UtC-4316 UtC-4206 J-118 J-114 J-116 J-117 UtC-8511 J-115
1540 ± 30 3190 ± 40 10,300 ± 60 19610 ± 190 17,120 ± 170 17,880 ± 150 16,460 ± 130 25,300 ± 190 17,190 ± 140
−23.4 −28.2 −17.6 −20.9 −20.87 −14.64 −21.08 −20.4 −18.54
1490–1380 3450–3380 12,169–12,051 23,650–22,800 20,800–20,000 21,650–20,850 20,000–19,250 not available 20,850–20,100
82–83 289–290 777–778 1149–1151 1269–1270 1414–1416 1434–1436 1449–1451 1496–1498
Note. Radiocarbon dates (accelerator mass spectroscopy) on bulk organic carbon from Lake Tilla core LT97 and LT95. All radiocarbon dates were calibrated with the data set of OxCal 3.5 (Bronk Ramsey, 1995; Stuiver et al., 1998).
Pollen, Charcoal, and Diatom Analysis Subsamples for pollen analysis were taken every 10 cm from core LT97 using the standard preparation technique (Faegri and Iversen, 1989). Calculation of pollen percentages is based on a pollen sum that generally exceeds 800 pollen grains and includes arboreal and nonarboreal land pollen taxa (TLP). Among the 138 pollen taxa identified, trees and shrubs were grouped after their phytogeographical distribution. For the reconstruction of fire activities, microscopic charcoal particles were estimated for every pollen subsample of core LT97 by determining the surface area cover of all visible charcoal particles in volumetrically standardized pollen slides (Clark, 1982). Subsamples for diatom analysis were obtained from core LT95 approximately every 30 cm and prepared using HCl and H2 O2 . Diatoms were identified and grouped by their main habitat in planktonic and nonplanktonic taxa.
Material and Dating A 16-m-long sediment core (LT95) was taken in March 1995 on the western shore of Lake Tilla using a modified Livingstone piston corer (Fig. 1). In April 1997, when the lake had been completely desiccated, additional parallel cores (LT97) were taken from the center of the lake. However, we did not reach the base of the lacustrine sediments as an artesian groundwater horizon at ca. 9 m led to a continuous loss of sediment from the core chamber. Dating control is provided by 18 radiocarbon dates (accelerator mass spectrometry [AMS]) from organic matter (Table 1). All dates refer to uncalibrated 14 C yr with inferred ages based on linear interpolation.
Sedimentological and Geochemical Analyses TIC (total inorganic carbon), TOC (total organic carbon), TN (total nitrogen), and δ 13 C values of the organic fraction (δ 13 Corg ) were analyzed at 20-cm intervals. TIC was determined using a carbon analyzer and TOC was measured photometrically after oxidation with H2 SO4 and K2 Cr2 O7 . The C/N ratio was calculated directly from the percentages of TOC and TN. For the determination of δ 13 Corg the acid-pretreated, carbonate-free sample was combusted and CO2 was transferred to a Delta Plus XL isotope ratio mass spectrometer. For analysis of carbonates and X-ray diffractometry (minerogenic composition), subsamples
LATE QUARTERNARY CLIMATE AND VEGETATION, NIGERIA
75
FIG. 1. Location map of Lake Tilla on the Biu Plateau, modified after Directorate of Overseas Surveys for the Nigerian Government (1968), and vegetation zones of West Africa after White (1983).
76
SALZMANN, HOELZMANN, AND MORCZINEK
were obtained approximately every 70 cm. The Ca, Mg, and Sr contents of the bulk carbonates were analyzed by measuring a 3% HCl-dilution with an inductively coupled plasma-optical emission (ICP-OES). Qualitative and semiquantitative analyses of mineralogical compounds of the bulk samples were examined by X-ray powder diffraction analyses. The minerals contributing to TIC provide information about the brine saturation from which they precipitated. High Mg/Ca molar ratios and the existence of high Mgcalcites (>4–30 mole % of MgCO3 ) or Ca-dolomite (Ca0.5–0.7 Mg0.5–0.3 (CO3 )2 ) point to higher evaporation. Authigenic aragonite shows a high uptake-rate of Sr (F¨uchtbauer, 1988) that influences the Sr/Ca molar ratio and points to higher salinity of the palaeo-lake water. However, these interpretations were only be applied to samples with a TIC content of >0.5% and not for the upper part of the core (811–35 cm). For the same reason δ 13 Ccarb and δ 18 Ocarb values were determined only between 1500 to 811 cm where the TIC content exceeded 0.5%. CO2 was prepared from organic-free samples by treatment with 100% H3 PO4 and transferred to a MAT 252 isotope ratio mass spectrometer. As the Ca-dolomite content remains low in all samples (<100 counts/sec), no corrections for the fractionation coefficient were made. δ 13 Ccarb , δ 18 Ocarb , and δ 13 Corg values are given relative to the Vienna-Peedee Belemnite (V-PDB) standard. RESULTS AND INTERPRETATION
Radiocarbon Chronology The age–depth correlation, which is based for the Holocene on nine AMS dates from core LT97 (Fig. 2), shows sedimentation
FIG. 2. Age vs. depth plot. Sediment accumulation rates for the Holocene sequence are based on linear interpolation of 14 C dates from core LT97. The numbers are accumulation rates in cm/yr for each portion of the curve delimited by radiocarbon dates.
rates to be constant apart from a distinct increase between ca. 2500 and 1100 14 C yr B.P. For the Pleistocene sequence, the radiocarbon dates display a more complex pattern. Abrupt changes in lithology point to discontinuities in sedimentation (Fig. 3). However, the biostratigraphical and geochemical trends within the Pleistocene profile suggest that the sediments are in situ and not extensively reworked as might be suggested by the inverse radiocarbon dates. Given the very low organic carbon and high carbonate content of the Pleistocene sediments, we assume contamination has occurred. The detrital input of organic material from the slopes during low lake levels and/or the input of “old” aqueous carbon by groundwater seem most likely. Hence, a number of the radiocarbon ages are probably too old. Unfortunately the low carbon content and lack of pollen or macrofossils hinders the establishment of a more precise chronology for the Pleistocene. However, comparison with other West African palaeo-records, as discussed later in this paper, suggests that the Lake Tilla core goes back to the period of the Last Glacial Maximum. Therefore, the radiocarbon dates at 1435 cm (16,460 ± 130 14 C yr B.P.) and 1497 cm (17,190 ± 140 14 C yr B.P.) seem to be the most reliable. Late Pleistocene Sequence: ca. 17,000 14 C yr B.P.–ca. 12,000 14 C yr B.P. Sediment zone LTS-I to LTS-III, diatom zone LTD 1a and LTD1b. The late Pleistocene sediment sequence (1500–900 cm), which consists mainly of sandy silts with gravel, is characterized by very low TOC (<1%) and C/N ratios varying between 6.7 and 18.8, as well as strongly fluctuating carbonate contents (Fig. 3). Covariation in δ 18 Ocarb and δ 13 Ccarb values suggests that the stable isotope curves reflect hydrological changes within the lake. Pollen are not preserved in the Pleistocene section but constantly high δ 13 Corg values indicate C4dominated grass savannas on the Biu Plateau throughout the late Pleistocene. Between 1500 and 1405 cm (LTS-I) relatively low C/N ratios (7.7–11.7) suggest increasing proportions of aquatic plants. From the molar Sr/Ca ratio (0.0031–0.0037) of the carbonates oligosaline to mesosaline conditions are inferred. High numbers of nonplanktonic diatoms in particular Fragilaria and of the planktonic taxa Aulacoseira granulata var. angustissima (O. M¨uller) Simonsen at the corresponding diatom zone LTD1a (1465–1405 cm) indicate fluctuating water levels (Fig. 4). However, the dominance of Aulacoseira granulata (Ehrenberg) Simonsen with a maximum at 1450 cm and a low δ 18 Ocarb value suggest that during this period Lake Tilla might have reached significantly higher levels than throughout the following late Pleistocene section. The basal silts and clays are covered by a 1-m-thick layer (1405–1300 cm) of almost pure carbonates (high-Mg calcite, Ca dolomite, aragonite). Positive values of δ 18 Ocarb and δ 13 Ccarb point to a reduced precipitation/evaporation ratio with high ionic concentration of the palaeo-lake water (oligo- to mesosaline) as inferred from the molar Sr/Ca ratio (0.0034–0.0046). The C/N
LATE QUARTERNARY CLIMATE AND VEGETATION, NIGERIA
77
FIG. 3. Stratigraphy, sedimentary geochemistry, stable isotopes, and radiocarbon dates for the Holocene (LT97) and Pleistocene (LT95) sediments of Lake Tilla. (See Fig. 6 for a description of sedimentary units). TOC = total organic carbon; TIC = total inorganic carbon; V-PDB = Vienna-Peedee Belemnite.
ratios range from 12.1 to 18.8 and reach highest values of the entire sequence indicating a high input of terrestrial plants at low water depths. A thin clayey layer rich in diatoms between 1285 and 1265 cm (LTS-III) suggests a return to slightly higher lake levels. However, the dominance of nonplanktonic taxa at diatom zone LTD-1b (Fig. 4) implies that the water table remained permanently low. Between 1300 and 1047 cm, fluctuating curves of C/N, δ 13 Corg , 13 δ Ccarb , δ 18 Ocarb , and TIC suggest changing water depths and varying ionic concentrations of the palaeo-lake water. High δ 13 Ccarb and positive δ 18 Ocarb values represent low precipitation/evaporation ratios and coincide with the formation of Cadolomite, whereas lower δ 13 Ccarb and δ 18 Ocarb values are associated with low-Ca dolomite contents and precipitation of low-Mg calcite, pointing to elevated P/E ratios. High quartz grain content indicates increased input from the surrounding slopes and infers lower water depths. Between 1047 and 937 cm (LTS-III), the lower and more constant C/N ratio (average 7.8), lower isotopic composition of the carbonates (−2.64 to −2.16 ‰ δ 18 Ocarb ), and Sr/Ca molar ratios between 0.0026 and 0.0042 point to a slightly deeper water depth, although the high amount of quartz and basaltic gravel invoke an increased input of detrital material.
Late Pleistocene–Holocene Transition: ca. 12,000 14 C yr B.P.–10,000 14 C yr B.P. Sediment zone LTS-IV, pollen zone LT-1, diatom zone LTD-2. Alternating clay and sandy silt layers between 900 and 835 cm (LTS-IV, Fig. 3) and an increase in planktonic diatoms in zone LTD-2 (Fig. 4) characterize the transition of Lake Tilla to a permanent freshwater lake. However, increasing C/N ratios, which point to a higher input of terrestrial plant material, and an alternating abundance of planktonic (e.g., Aulacoseira granulata var. angustissima) and nonplanktonic taxa (e.g., Fragilaria ulna (Nitzsch) Lange-Bertalot) suggest relatively low and fluctuating water levels. Large quantities of grass pollen and only few arboreal taxa indicate an open grassland on the Biu Plateau at the end of the late Pleistocene (pollen zone LT-1, Fig. 5). Increasing percentages of herbs and Typha suggest a rise in littoral marginal zones inside the crater caused by a significant lowering of lake levels between ca. 10,900 and 10,500 14 C yr B.P. Evidence for a short and more arid period shortly before the onset of the Holocene is also given by the absence of tree pollen, the high Sr/Ca (0.0052) and Mg/Ca (1.27) molar ratios as well as a relatively high δ 18 Ocarb value of +0.1 ‰ at 840 cm
78
SALZMANN, HOELZMANN, AND MORCZINEK
FIG. 4.
Summary diatom diagram for the Holocene and Late Pleistocene sequences of core LT95. Plus sign (+) indicates values below 5%.
which can all be attributed to a lower precipitation/evaporation ratio. Holocene (<10,000 14 C yr B.P.) Sediment zone LTS-V, pollen zone LT-1 to LT-3, diatom zone LTD-3 to LTD5. Humid conditions throughout the Holocene
are indicated by the deposition of black organic lake mud with high TOC and low TIC values. The Holocene diatom spectra, dominated by the circumneutral to alkaliphilous species A. granulata (pH >7), indicate stable hydrological conditions (Servant and Servant-Vildary, 1980). High and relatively undeviating quantities of charred grass epidermis suggest regular savanna fires throughout the entire Holocene.
FIG. 5. Summary Holocene pollen diagram (LT97) showing the percentage of main taxa with depth bars exaggerated 5× and charcoal particles content in cm2 /cm3 plotted against depth.
LATE QUARTERNARY CLIMATE AND VEGETATION, NIGERIA
79
80
SALZMANN, HOELZMANN, AND MORCZINEK
Early to Mid-Holocene (ca. 10,000 14 C yr B.P.–3800 14 C yr B.P.) Sediment zone LTS-Va and LTS-Vb, pollen zone LT-2, diatom zone LTD-3 and LTD-4. The wetter condition at the onset of the Holocene led to an increase in several Sudanian and Guinean tree taxa, notably Combretaceae, Lannea/Sclerocarya, Uapaca, Alchornea, and Cussonia, as well as a decline in herbs (pollen zone LT-1 and LT-2a, Fig. 5). Whereas humidity increased, temperatures might have been significantly lower than today, as suggested by high numbers of Olea hochstetterii (max. 15%) between ca. 10,000 and 8600 14 C yr B.P. This species is absent on the Biu Plateau today and the nearest population can be found on the Jos Plateau and Mandara mountains above 1300 m asl (White, 1983). A highly diverse, densely wooded Guinean savanna became established on the Biu Plateau between ca. 8600 and 7000 14 C yr B.P. The corresponding pollen zone LT-2b shows highest tree pollen percentages and the greatest number of taxa (Fig. 5). Although the tree cover was presumably denser than today, high percentages of grass pollen (>52%), charred particles of grass epidermis, and the presence of characteristic savanna trees such as Lophira, Cussonia, and Burkea africana clearly indicate that closed forest never became dominant. In floristic composition the mid-Holocene vegetation closely resembles the modern “Southern Guinean Savannas” (Keay, 1959) at the forest– savanna boundary about 350 km south of Lake Tilla with annual rainfall exceeding 1200 mm. The Guinean savanna tree Uapaca and the shrub Alchornea might have been a substantial part of the swamp forest fringing the mid-Holocene lake. From ca. 7000 14 C yr B.P. onward a decline in woody climbers, Pteropsida, and several Guinean tree taxa, notably Cussonia, Burkea africana, and Alchornea, can be recorded at pollen zone LT-2c indicating the onset of drier environmental conditions. After ca. 5200 14 C yr B.P., the Guinean swamp forest fringing Lake Tilla might have become more open as suggested by a distinct decrease in Uapaca percentages followed by increasing Typha and Poaceae (LT-2d). The decline in tree taxa from ca. 7000 14 C yr B.P. onward was accompanied by a change in floristic composition from a wet Southern-Guinean to a drier SudanoGuinean savanna. This is in particular shown by the successive occurrence of pollen from the characteristic Northern Guinean savanna tree Monotes kerstingii (Keay, 1959) at ca. 6000 14 C yr B.P. (LT-2c, Fig. 5) and the Sudanian tree taxa Isoberlinia at 4800 14 C yr B.P. (LT-2d). Fluctuating percentages of A. granulata and A. distans var. africana (LTD-4, Fig. 4) implies contemporaneous changes in water chemistry which are accompanied by a decline in TOC and a trend toward heavier δ 13 Corg values, suggesting an increased input of terrestrial C4 organic material (LTS-Vb, Fig. 3). Mid- to Late Holocene (<3800 14 C yr B.P.) Sediment zone LTS-Vc, pollen zone LT-3, diatom zone LTD-5. A shift toward drier environmental conditions is recorded for the transition to the late Holocene after ca. 3800 14 C
yr B.P. The sharp decrease in TOC values and a rise in δ 13 Corg values (LTS-Vc, Fig. 3) indicate increasing importance of C4savanna grasses. Consistently lower C/N ratios at ca. 10.2 are recorded below 475 cm, suggesting that algae became the predominant source of organic matter from about 4300 14 C yr B.P. onward. After ca. 3800 14 C yr B.P. an open Sudanian savanna with woody taxa from the drier Sahelian zone such as Balanites, Commiphora, and Salvadora became established (pollen zone LT-3; Fig. 5). Like today, numerous herbs (e.g., Typha, Amaranthaceae/Chenopodiaceae, Asteroideae, Mitracarpus) colonized the seasonally flooded shorelines of the shrinking lake. The former Guinean swamp forest with Uapaca and Alchornea was replaced by a fringing Sudanian woodland with Celtis cf. integrifolia, and Ficus (pollen zone LT-3a). An increase in quartz reflects a higher detrital input. The reoccurrence of the diatom Aulacoseira granulata var. angustissima (LTD-5a, Fig. 4), the regular presence of the halophilous indicator species Cyclostephanos dubius (Fricke) Round (Krammer and LangeBertalot, 1991), and a Mg-calcite formation relatively rich in Sr and Mg after ca. 3000 14 C yr B.P. confirm a further aridification that was accompanied by a rise in salinity of the lake water. Dry, grayish, calcareous silts with high Sr/Ca (0.0042) and Mg/Ca (1.58) molar ratios at the top of the sequence and rising pollen percentages of fringing herbs, in particular Cyperaceae and Typha, at pollen zone LT-3c document an advanced silting up of Lake Tilla after 1100 14 C yr B.P. The distinct increase in sedimentation rates between ca. 2500 and 1100 14 C yr B.P. (Fig. 2) is difficult to interpret. Gradually increasing δ 13 Corg and TIC values with precipitation of high-Mg calcite as well as rising Mg/Ca and Sr/Ca ratios point to a continuation of the aridification trend which might have enforced the soil erosion on the crater slopes (sediment zone LTS-Vc, Fig. 3). However, although the pollen data show a slight decline in tree pollen, the undeviating percentages of fringing herbs and Sahelian tree taxa (Fig. 5, pollen zone LTD-3b) contradict the interpretation of a rapidly increasing aridity after 2500 14 C yr B.P. Apart from climatic change, human impact by early shepards and farmers, for whom the lake became an important water reservoir during the dry season, may have been responsible. As it is today, trampling and grazing by livestock may have enforced slope erosion while the reduction in tree pollen percentages reflects anthropogenic clearing. DISCUSSION AND CONCLUSIONS
Late Quaternary Climatic Changes With a dry late Pleistocene, a humid early to mid-Holocene, and a subsequent aridification, Lake Tilla confirms the general climatic trends previously described by other palaeoecological records from northern tropical and subtropical Africa (e.g., L´ezine, 1989; Dupont and Agwu, 1992; Gasse, 2000; Salzmann and Waller, 1998). The few late Pleistocene records presently available for West Africa indicate that the period of maximum aridity associated with low temperatures took place between
LATE QUARTERNARY CLIMATE AND VEGETATION, NIGERIA
ca. 20,000 and 15,000 14 C yr B.P. (e.g., Servant and ServantVildary, 1980; Maley and Brenac, 1998). At all sites a slow return to wetter climatic conditions, indicated by rising lake levels and a spread of rain forest in West Cameroon and Ghana, can be detected from ca. 15,000 to 14,000 14 C yr B.P. onward (Maley, 1996). The correlation with other west African sites suggests that the Pleistocene sequence of Lake Tilla, for which precise radiocarbon datings are not available, goes back to at least 17,000 14 C yr B.P. As is the case with the Pleistocene records from the West African rain-forest zone (e.g., Talbot and Johannessen, 1992), the Lake Tilla profile indicates that the Last Glacial Maximum and the transition to the Holocene humid period was evidently interrupted by several short arid and humid intervals. Of particular note is a distinct dry event at Lake Tilla between ca. 10,900 and 10,500 14 C yr B.P. (Fig. 6). Comparable aridity events, generally placed in the interval 11,000–10,000 14 C yr B.P., have been recorded at numerous other sites in tropical
81
Africa and were attributed to the climatic change of the Younger Dryas period (e.g., Talbot and Johannessen, 1992; Maley and Brenac, 1998; Salzmann and Waller, 1998). The onset of the Holocene humid period is marked at nearly all West African sites by an abrupt increase in lake levels at ca. 10,000 14 C yr B.P. At Lake Tilla, wetter climatic conditions led to a rapid spread of Guinean and Sudanian tree taxa into the formerly open grass savanna. Temperatures might have been lower than today as indicated by the dominance of the mountain olive Olea hochstetterii until ca. 8600 14 C yr B.P. This coincides with palynological results from Lake Bosumtwi in Ghana, where the appearance of montane elements at 400 m asl was interpreted by Maley (1996) as an indicator of low temperatures. However, it has to be considered that the early Holocene vegetation on the Biu Plateau which consisted of a mixture of Guinean lowland savanna with a few mountainous taxa apparently has no modern analogue. The rapid spread of the mountain olive on the sparsely vegetated Biu Plateau and its dominance during the
FIG. 6. Summary diagram showing the woody pollen taxa grouped after their main phytogeographical distribution, the percentages of planktonic vs. nonplanktonic diatoms and the main geochemical data for Lake Tilla plotted against uncalibrated and calibrated radiocarbon dates. The palaeoclimate curve shows trends in precipitation/evaporation since approximately the last 17,000 14 C yr B.P. as deduced from reconstructed lake level and vegetation changes. Dashed line indicate sequences were the interpretation is obscured by insufficient radiocarbon datings or possible hiati.
82
SALZMANN, HOELZMANN, AND MORCZINEK
early Holocene might have been also caused by a preemption of sites. The early to mid-Holocene humid period culminated between ca. 8500 and 7000 14 C yr B.P. Dense Guinean savannas became established on the Biu Plateau, supporting the assumption of a latitudinal shift of vegetation zones of West Africa of about 400 km northward in response to climatic changes (e.g., Lezine, 1989; Dupont and Agwu, 1992). The maximum northward shift of vegetation was accompanied by high water levels of Lake Tilla. Palynological and palaeolimnological data suggest that the optimum of the Holocene humid period had been passed in northeastern Nigeria after ca. 7000 14 C yr B.P. The course of the termination of the African Holocene humid period and transition to modern more arid conditions is subject of controversy. Abrupt dry climatic events have been detected at several sites in tropical and subtropical West Africa, in particular around 7500, 4500, 3700, and 2800 14 C yr B.P., each followed by a return to wetter climates (e.g., Servant and Servant-Vildary, 1980; Talbot and Delibrias, 1980; Maley and Brenac, 1998; Gasse, 2000). However, as these events show a high variability in timing, it remains unclear in how far they represent short term episodes with different site related amplitudes or rather changes in global atmospheric circulation. At Lake Tilla the humid period terminated in a gradual decline in precipitation from ca. 7000 14 C yr B.P. onward, resulting in a floristic change from Guinean to drier Sudanian savannas and a parallel lowering of lake levels (Fig. 6). At the Manga Grasslands situated about 300 km north of Lake Tilla in the Sahelian zone of Nigeria, a gradual climatic deterioration set in at ca. 5000 14 C yr B.P. (Salzmann and Waller, 1998; Street-Perrott et al., 2000). The multiple pollen profiles taken from several dune depressions of the Manga Grasslands revealed differences in timing of vegetational changes of up to 1000 yr which might have been caused by specific local conditions such as variations in the relative height of the water table (Salzmann and Waller, 1998). Similar reasons might have also been responsible for the disparity between Lake Tilla and the Manga Grasslands where the end of the mid-Holocene humid period is recorded in the Sahelian pollen diagrams about 2000 yr later. The profiles from both regions give no evidence for a short dry spell around 7500 14 C yr B.P. which has been detected in numerous lake records from the northern monsoon domain of Africa (e.g., Servant and Servant-Vildary, 1980; Talbot and Delibrias 1980; Gasse 2000) and was partly connected with a short but distinct cooling event in the North Atlantic and Greenland (Alley et al., 1997; Gasse, 2000). However, the inference that the period between ca. 7000 and 7500 14 C yr B.P. was not a short event but rather marks the end of most humid conditions in the northern monsoon domain of Africa followed by a gradual decline in precipitation and evaporation was recently supported by a highresolution eolian sediment record taken from a marine core off Cap Blanc, Mauritania (DeMenocal et al., 2000). At Cap Blanc the gradual mid-Holocene aridification culminated in a very abrupt onset of much more arid condi-
tions at ca. 4800 14 C yr B.P. (DeMenocal et al., 2000). At Lake Tilla the gradual trend toward aridity seems to have intensified about 1000 yr later. At ca. 3800 14 C yr B.P. a distinct drop in water levels of Lake Tilla indicates a shift toward drier climatic conditions. A distinct biostratigraphical and sedimentological change can be recorded at nearly all West African sites for the beginning of the Late Holocene between ca. 4000–2000 14 C yr B.P. (e.g., Lezine, 1989; Ballouche and Neumann, 1995; Maley and Brenac, 1998; Salzmann and Waller, 1998). The strong temporal variations between sites suggest that this change was caused by an intensification of a general trend with different regional amplitudes rather than by one abrupt climatic event. In the Sahel of Nigeria (Salzmann and Waller, 1998) and Burkina Faso (Ballouche and Neumann, 1995) the establishment of the modern vegetation coupled with a distinct drop in lake levels occurred between ca. 3300 and 3000 14 C yr B.P. In western Senegal a similar change occurred about 1000 yr later (Lezine, 1989). In the rain-forest zone of western Cameroon an abrupt opening of the vegetation took place at ca. 2500 14 C yr B.P., which was followed by a return to wetter climates after 2000 14 C yr B.P. (Maley and Brenac, 1998). In northeastern Nigeria the aridification continued throughout the late Holocene resulting in the recent desiccation of Lake Tilla. The Role of Climate Change and Human Impact on the Distribution of Savannas and Forests on the Biu Plateau Competition for limited water resources is often cited as being a major factor which controls the distribution of savanna and forest in tropical West Africa (e.g., Walter, 1979; Cole, 1986). Walter (1979) postulated that natural savannas as a stable “climatic climax” are confined to regions with summer rains and a mean annual rainfall of 300–500 mm. This hypothesis was supported by studies on modern vegetation in protected savanna areas which suggest that the Sudanian zone with a mean annual rainfall of 500–1100 mm receives enough precipitation to promote the establishment and spread of dense forests (e.g., Swaine, 1992). Human impact has been emphasized as being responsible for the progressive degradation of these closed forests to modern savannas by clearing, grazing, agriculture, and iron smelting—a process which might have started during the midto late Holocene (e.g., Anhuf, 1997). On the basis of the new palynological results from Lake Tilla, the hypothesis of the anthropogenic origin of Sudanian savannas during the Holocene must be rejected. High percentages of grass pollen, the regular occurrence of savanna fires, and the presence of characteristic savanna trees clearly indicate that savannas were always present on the plateau and never replaced closed forests. The charcoal particle curve of Lake Tilla supports the assumption that frequent fires constitute a major agent which maintains the open character of the savanna vegetation (Swaine, 1992). The constancy of the charcoal curve provides no evidence whether these fires were mainly of natural origin or induced by early hunters, who may have started them long before the onset of the Holocene.
LATE QUARTERNARY CLIMATE AND VEGETATION, NIGERIA
Today most of the West African savannas are cultural landscapes which have been strongly shaped by humans. Archaeological and archaeobotanical investigations suggest that in northeastern Nigeria human impact increased from at least 3000 14 C yr B.P. onward (Breunig et al., 1996). Nevertheless, although different palaeoecological methods were applied, the Lake Tilla record neither supports the archaeological findings nor does it provide evidence at which point the anthropogenic transformation to the modern “farmed parklands” have started. The lack of unambiguous indicators even for the late Holocene demonstrates the difficulties of detecting human activity within the open vegetation of a savanna.
ACKNOWLEDGMENTS The research is part of the West African Savanna project (SFB 268) at the University of Frankfurt, which is funded by the German Research Foundation (DFG). Special thanks go to Alhaji Yamta Thetur, Mohammed Bata, R. Byer, J. Maley, K. Neumann, G. Schenk, and E. Schulz . We also thank H. F. Lamb and an anonymous reviewer for their constructive comments.
REFERENCES Alley, L. J., Mayewski, P. A., Sowers, T., Stuiver, M., Taylor, K. C., and Clark, P. U. (1997). Holocene climatic instability: A prominent, widespread event 8200 years ago. Geology 25, 483–486. Anhuf, D. (1997). Paleovegetation in West Africa for 18.000 B.P. and 8.500 B.P. Eiszeitalter und Gegenwart 47, 112–119. Ballouche, A., and Neumann, K. (1995). A new contribution to the Holocene vegetation history of the West African Sahel: Pollen from Oursi, Burkina Faso and charcoal from three sites in northeast Nigeria. Vegetation History and Archaeobotany 4, 31–39. Breunig, P., Neumann, K., and Van Neer, W. (1996). New research on the Holocene settlement and environment of the Chad basin in Nigeria. African Archaeological Review 13, 111–145. Bronk Ramsey, C. (1995). Radiocarbon calibration and analysis of stratigraphy: The OxCal program. Radiocarbon 37, 425–430. Clark, R. L. (1982). Point count estimation of charcoal in pollen preparation and thin sections of sediments. Pollen et Spores 24, 523–535. Cole, M. M. (1986). “The Savannas. Biogeography and Geobotany.” Academic Press, London. CSA (Conseil Scientifique pour l’Afrique au Sud du Sahara). (1956). Phytog´eographie Phytogeography. Yangambi Publication 53. Yangambi, DR of Congo. DeMenocal, P., Ortiz, J., Guilderson, T., Adkins, J., Sarnthein, M., Baker, L., and Yarusinsky, M. (2000). Abrupt onset and termination of the African Humid Period. Quaternary Science Reviews 19, 347–361. Directorate of Overseas Surveys for the Nigerian Government. (1968). Nigeria 1:50.000. Biu Sheet 133 S.W. Lagos, London. Dupont, L. M., and Agwu, C. O. C. (1992). Latitudinal shifts of forest and savanna in NW Africa during the Brunhes chron: Further marine palyno-
83
logical results from site M 16415 (9◦ N, 19◦ W). Vegetation History and Archaeobotany 1, 163–175. Faegri, K., and Iversen, J. (1989). “Textbook of Pollen Analysis.” Wiley, New York. F¨uchtbauer, H. (Ed.) (1988). “Sedimente und Sedimentgesteine II. Teil.” Schweizerbart, Stuttgart. [In German] Gasse, F. (2000). Hydrological changes in the African tropics since the Last Glacial Maximum. Quaternary Science Reviews 19, 189–211. Keay, R. W. J. (1959). “An outline of Nigerian vegetation.” Government Printer, Lagos. Krammer, K., and Lange-Bertalot, H. (1991). “S¨usswasserflora von Mitteleuropa, Bacillariophyceae, 3. Teil: Centrales, Fragilariaceae, Eunotiaceae.” Fischer, Stuttgart. [In German] L´ezine, A.-M. (1989). Late Quaternary vegetation and climate of the Sahel. Quaternary Research 2, 317–334. Maley, J. (1996). The African rain forest—Main characteristics of changes in vegetation and climate from the Upper Cretacous to the Quaternary. Proceedings of the Royal Society of Edinburgh 104B, 31–74. Maley, J., and Brenac, P. (1998). Vegetation dynamics, palaeoenvironments and climatic change in the forests of western Cameroon during the last 28000 years B.P. Review of Palaeobotany and Palynology 99, 157–187. Salzmann, U. (2000). Are savannas degraded forests?—A Holocene pollen record from the Sudanian zone of NE-Nigeria. Vegetation History and Archaeobotany 9, 1–15. Salzmann, U., and Waller, M. (1998). The Holocene vegetational history of the Nigerian Sahel based on multiple pollen profiles. Review of Palaeobotany and Palynology 100, 39–72. Servant, M., and Servant-Vildary, S. (1980). L’environnement Quaternaire du bassin du Tchad. In “The Sahara and the Nile” (M. A. J. Williams and H. Faure, Eds.), pp. 133–162. Balkema, Rotterdam. [In French] Street-Perrott, F. A., Holmes, J. A., Waller, M. P., Allen, M. J., Barber, N. G. H., Fothergill, P. A., Harkness, D. D., Ivanovich, M., Kroon, D., and Perrot, R. A. (2000). Drought and dust deposition in the West African Sahel: A 5500-year record from Kajemarum Oasis, northeastern Nigeria. The Holocene 10, 293– 302. Stuiver, M., Reimer, P. J., Bard, E., Beck, J. W., Burr, G. S., Hughen, K. A., Kromer, B., McCormac, G., van der Plicht, J., and Spurk, M. (1998). INTCAL98 radiocarbon age calibration, 24000–0 cal B.P. Radiocarbon 40, 1041–1083. Swaine, M. D. (1992). Characteristics of dry forest in West Africa and the influence of fire. Journal of Vegetation Science 3, 365–374. Talbot, M. R., and Delibrias, G. (1980). A new Pleistocene-Holocene water-level curve for Lake Bosumtwi, Ghana. Earth and Planetary Science Letters 47, 336–344. Talbot, M. R., and Johannessen, T. (1992). A high resolution palaeoclimatic record for the last 27,500 years in tropical West Africa from the carbon and nitrogen isotopic composition of organic matter. Earth and Planetary Science Letter 110, 23–37. Tuley, P. (Ed.). (1972). “The Land Resources of North East Nigeria Volume 1: The Environment.” Foreign and Commonwealth Office Overseas Development Administration, Land Resources Division, Surbiton. Walter, H. (1979). “Vegetation of the Earth and Ecological Systems of the GeoBiosphere.” Springer, Berlin. White, F. (1983). “The Vegetation of Africa. A Descriptive Memoir to Accompany the Unesco/AETFAT/UNSO Vegetation Map of Africa.” Unesco, Paris.