triar (1135)

Upload: wellington-fava

Post on 07-Apr-2018

228 views

Category:

Documents


0 download

TRANSCRIPT

  • 8/6/2019 triar (1135)

    1/16

    461 Journal of Paleolimnology 21: 461476, 1999. 1999Kluwer Academic Publishers. Printed in the Netherlands.

    Environmental history of the Colombian savannas of theLlanos Orientales since the Last Glacial Maximum from

    lake records El Pinal and Carimagua*

    Introduction

    The Late Quaternary environmental history of Neo-

    tropical lowland ecosystems is still poorly known

    despite their importance in the verification of global

    climate models and a better understanding of global

    biodiversity (Clapperton, 1993; Markgraf, 1993). The

    major Neotropical lowland ecosystems are the two

    main rain forest areas (Amazonia, and the Pacific rain

    forest area of Choc) and several savanna areas. In

    northern South America, the Llanos Orientales in

    Colombia and the Orinoco Llanos in Venezuela

    represent, after the cerrado in Brazil, the second

    major Neotropical savanna ecosystem. Other large

    savanna areas in northern South America (Figure 1)

    are the Gran Sabana in Venezuela, the Rupununi-Rio

    Branco savanna in Surinam, the Rio Jari savanna, and*This paper was presented at the 7th International Symposium

    on Palaeolimnology (1997), held at Heiligkreuztal, Germany

    Hermann Behling & Henry Hooghiemstra

    University of Amsterdam, Hugo de Vries-Laboratory, Department of Palynology and Paleo/Actuo-ecology,

    Kruislaan 318, 1098 SM Amsterdam (The Netherlands Centre for Geo-ecological Research, ICG), The

    Netherlands (E-mail: [email protected])

    Received 9 December 1997; accepted 6 June 1998

    Key words: Colombia, Llanos Orientales, pollen analysis, savanna, gallery forest, vegetation and climate

    history, Late Quaternary

    Abstract

    Late Quaternary environments have been studied by pollen analysis of lake sediments from the savannas of the

    Colombian Llanos Orientales at 180 m elevation. The pollen record form Laguna El Pinal (408N, 7023W),

    dated by 6 AMS radiocarbon dates, starts at 18,290 14C yr B.P. The record from Laguna Carimagua (404N,

    7014W), also dated by 6 AMS dates, starts at 8270 14C yr B.P. Both records show a landscape dominated by

    grassland savanna with only few woody savanna taxa, such as Curatella andByrsonima, frequent fires, and little

    occurrence of forest and/or gallery forest along the rivers. The savanna ecosystem at the studied sites was relatively

    stable during the last 18,000 yrs, but minor changes in floral composition, and in the proportion of savanna/forest,

    have been recorded. Very little gallery forest and the non permanent lake conditions of Laguna El Pinal reflect

    the driest period, interpreted to reflect low rainfall rates and long dry seasons during the Last Glacial Maximum

    until 10,690 14C yr B.P. During the Late Glacial, Laguna El Pinal was a permanent shallow lake, and changed into

    a lake with higher water levels during the Holocene, indicating wetter conditions. Expansion of regional gallery

    forest also started at around 10,690 14C yr B.P. Little vegetational change observed in Laguna Carimagua at

    5570 14C yr B.P., in combination with a simultaneous decrease of savanna observed in previously studied lakes,

    suggest a change to regional wetter conditions. Thus, the Holocene before 5500 14C yr B.P. was somewhat drier

    than the following period until about 3850 14C yr B.P. In both records, Late Holocene lake deposits are incomplete.

    Shore vegetation of Laguna Carimagua always included a minor contribution of the palmsMauritia andMauritiella.

    The marked increase of palms during the last c. 3800 yrs points to increased human impact on the vegetation

    under the wettest Holocene climate regime.

  • 8/6/2019 triar (1135)

    2/16

    462

    Figure 1. Map showing the major savanna areas of northern South America: (1) Llanos Orientales in Colombia and Orinoco Llanos in

    Venezuela, (2) Gran Sabana, (3) Rupununi-Rio Branco savanna, (4) Rio Jari savanna and (5) coastal Guianan savanna. The inset map shows the

    geographical location of the savanna sites mentioned in the text: Laguna El Pinal, Laguna Carimagua, Laguna Angel, Laguna Sardinas and

    Laguna de Agua Sucia.

    the belt of coastal Guianan savannas (Hueck & Seibert,

    1972).

    The first palynological data on the Holocene

    vegetation history and climatic change of the Llanos

    Orientales were published by Wijmstra & Van der

    Hammen (1966). Other palynological data from

    savanna ecosystems of northern South America are

    available from the Gran Sabana in Venezuela (Rull,

    1992), the Rupununi savanna in Surinam (Wijmstra

    & Van der Hammen, 1966) and the belt of coastal

    Guianan savannas in Surinam (Wijmstra, 1971),

    Guyana (Van der Hammen, 1963) and French Guiana

    (Tissot & Marius, 1992).

    We collected lacustrine sediment cores for palyno-

    logical studies to explore the regional history of

    vegetation, climate, and human influence on the

    tropical savanna ecosystem of the Llanos Orientales.

    Starting from the city of Villavicencio, which is located

    at the foothills of the Eastern Cordillera, we cored 13

    lakes and swamps along a 500 km long transect from W

    to E (Figure 1). The sites are located between the

    latitudes of 330 and 430 N and the longitudes of 69

    and 74 W, at elevations between 80 and 450 m in the

    provinces of Meta and Vichada. In this paper we present

    new pollen records of Laguna El Pinal and Laguna

    Carimagua to compliment the information of the

    previously studied Laguna Angel and Laguna Sardinas

    (Behling & Hooghiemstra, 1998).

    Study sites

    Both lakes, Laguna El Pinal (408N, 7023W) and

    Laguna Carimagua (404N, 7014 W) are about 20 km

    distant from each other and lie at the same elevation of

  • 8/6/2019 triar (1135)

    3/16

    463

    about 180 m in a relatively flat area in the central part

    of the Llanos Orientales in the province of Meta

    (Figure 1). The distance to Villavicencio, located atthe foothills of the Andes, is about 270 km.

    Laguna El Pinal (Figure 2) is a shallow lake and had

    a water depth of 70 cm during the coring operation, at

    the end of February 1996 in the dry season. The lake

    is c. 1200 m long and c. 300 m wide and looks like an

    old meander. However, it is unlikely that it was a part

    of a river, as the distance to the next river system of

    the Rio Meta is several kilometers. Laguna Carimagua

    (Figure 3) is also an isolated and shallow lake, c. 3 km

    long and c. 1.5 km wide. The water depth was 80 cm

    during the period of the coring, at the end of February.

    The lake is located in a shallow depression in a flat area,

    located above the large river system of the Rio Metain the north, and several much smaller river systems

    at the south. Within several kilometers distance to the

    lake, several small rivers, bordered by gallery forest,

    originate.

    The present-day vegetation of the studied region of

    the two lakes is a grass savanna, which is being used

    as pasture land. Extended areas of gallery forest grow

    along the small river bodies. Laguna El Pinal is

    surrounded by stands of the palmMauritia (local name

    morichi) and by small areas of shrubby vegetation,

    probably secondary. Laguna Carimagua, which is

    located within the protected area of the agriculturalexperimental station Centro de Investigacion para la

    Atillanura, Carimagua, also called Carimagua, is

    surrounded mainly by stands ofMauritiella and some

    Mauritia. Several small islands of pure Mauritiella

    palms are found in the lake area (Figure 4). At the

    northern edge, the lake is surrounded by grassland

    (Figure 3), whereas the other parts of the lake arebordered by small patches of forest.

    The vegetation of the Llanos Orientales is primarily

    characterized by different herbaceous savannas with

    shrubs and shrubby trees, and gallery forest along the

    rivers. At some places between the rivers, patches of

    forest occur in which the plant composition is

    comparable to Amazonian forest. General data on the

    flora and ecology of the savannas of the Llanos were

    published by Cuatrecasas (1989), Gentry (1993),

    Huber (1987), Hueck (1966), Hueck and Seibert

    (1972), Pinto-Escobar (1993) and Vareschi (1980).

    Blydenstein (1967) recognized 10 different plant

    associations in the savannas of the Llanos Orientales.

    According to Cuatrecasas (1989), characteristic taxa

    for grass savannas are Poaceae, belonging to the genera

    Andropogon , Eragrostis , Axonopus, Paspalum,

    Aristida, Ctenium andPanicum. Characteristic shrubs

    belong to the family Melastomataceae (Miconia,

    Tibouchina), Papilionaceae (Clitoria), Ceasalpiniaceae

    (Cassia), Lamiaceae (Hyptis), Sterculiaceae (Wal-

    theria), Malvaceae (Sida,Pavonia) and other families.

    Characteristic larger shrubs or trees, which can be also

    very small in size (less than 50 cm high in areas with

    frequent burning), are Curatella americana (Dill-

    eniaceae), Byrsonima crassifolia (Malpighiaceae),Bowdichia virgilioides (Papilionaceae) and Pal-

    icourea rigida (Rubiaceae). The composition and

    distribution of different savanna plant communities

    can be seen as the result of several environmental

    Figure 2. View on Laguna El Pinal with stands ofMauritia palm around the lake.

  • 8/6/2019 triar (1135)

    4/16

    464

    factors, including pedological, physical, biotic and

    climatic factors (Cole, 1982, 1986; Medina & Silva, 1990).

    Along rivers, lake shores and swampy areas, dense

    stands of the palmMauritia are often found.Mauritia

    swamp forest is called morichal. The less tall palm

    Mauritiella grows at lake shores in shallow water and

    surrounds several lakes in the Llanos Orientales.

    The climate of the Llanos Orientales is strongly

    seasonal. The marked annual dry period lasts about 4

    5 months, from mid November to mid March. In the

    northern part of the Llanos Orientales (province of

    Arauca) the dry season is longer than in the southern

    part (provinces of Meta and Vichada). In the Llanos

    Orientales there is a precipitation gradient from

    12002000 mm/yr in the northern region near the

    Venezuelan border, to about 20002500 mm/yr in the

    southern and southwestern parts of the savanna area.

    The transition from savanna to the Amazon rain forest

    occurs in the south, where annual precipitation is over

    24002500 mm/yr. Mean annual temperature in the

    study area is 2627 C with less than 3 C variation

    between monthly means. The small annual temperature

    oscillation contrasts to the 1015 C daily variation

    (Blydenstein, 1967; Snow, 1976).

    Figure 3. View on Laguna Carimagua, bordered by mainlyMauritiella palm vegetation, but at the northern border by open grass savanna.

    Figure 4. Mauritiella-dominated vegetation on the islands of Laguna Carimagua.

  • 8/6/2019 triar (1135)

    5/16

    465

    Methods

    We used an echo sounder to determine the bathymetryof each lake. Sediment cores were collected in the lake

    center where water was deepest. We used a modified

    Livingstone piston corer from a wooden platform,

    which was fixed on two inflatable rubber boats. The

    studied core from Laguna El Pinal is 72 cm long and

    from Laguna Carimagua is 95 cm long. Within 3 weeks

    sediment cores were transported by air freight to

    Amsterdam and stored at 4C in the laboratory prior

    to analysis.

    Accelerator Mass Spectrometry (AMS) radio-

    carbon dates were obtained from 6 bulk samples of

    1 cm core interval from core Laguna El Pinal, and 6

    samples from core Laguna Carimagua were collectedand dated at the Van der Graaff Laboratory of the

    University of Utrecht (Van der Borg et al., 1987).

    For pollen analysis, samples of 1 cm3 were taken

    in 2 cm intervals along the two cores. Prior to pro-

    cessing, one tablet of exoticLycopodium spores was

    added to each sample for calculation of pollen

    concentration and pollen influx. All samples were

    prepared using standard pretreatment techniques

    including sodium pyrophosphate, acetolysis, and heavy

    liquid separation by bromoform (Faegri & Iversen,

    1989). Pollen residues were mounted in glycerin

    gelatin and analyzed with a Zeiss microsope at 500 magnification. Identification of pollen grains and

    spores was aided by pollen morphological descriptions

    published by Behling (1993), Herrera & Urrego

    (1996), Hooghiemstra (1984), Roubik & Moreno

    (1991) and the reference collection of the Hugo de

    Vries-Laboratory. About 300 pollen grains from non-

    aquatic taxa were counted for the pollen sum.

    The sample from Laguna El Pinal at 60 cm depth

    contains too little pollen to obtain reliable statistics.

    Pollen grains below 60 cm core depth were less well

    preserved but a meaningful analysis was still possible.

    Samples below 81 cm from Laguna Carimagua contain

    either none, or a very low number of badly preserved

    pollen grains; therefore this core interval was excluded

    from analysis. Carbonized particles were not counted

    on the pollen slides because during heavy liquid

    separation many charcoal particles were eliminated

    and, therefore, such results are not meaningful.

    The pollen and spore data are presented in pollen

    diagrams as percentages of the pollen sum. The pollen

    sum excludes aquatic taxa, fern spores, moss spores,

    fungal spores, and algae. Species of Cyperaceae grow

    frequently in the savanna vegetation and were therefore

    included in the pollen sum. We arrived at the following

    ecological groups of fossil pollen: (1) shrubs and

    trees from the forest and gallery forest along rivercurrents, (2) savanna shrubs and trees, (3) savanna

    herbs, (4) aquatic taxa from lake shores, and (5) ferns.

    Pollen diagrams were plotted with the software

    TILIAGRAPH. For calculations we used TILIA and for

    the cluster analysis of terrestrial pollen taxa we used

    CONISS (Grimm, 1987).

    Stratigraphy

    The following lithological sequence was observed in

    the cores:

    Laguna El Pinal

    050 cm black fine detritus mud, compact, very rare

    plant remains, somewhat fine sandy

    5060 cm gray-greenish fine sand, very compact

    6064 cm gray-greenish clay with fine sand, very

    compact

    6472 cm light gray clay, compact, somewhat fine

    sandy

    Laguna Carimagua

    018 cm dark brown-dark gray fine detritus mud,

    medium compact, plant remains are very rare,somewhat fine sandy. At 18 cm depth: 1 cm thick

    wood fragment

    1830 cm dark brown-dark gray fine detritus mud,

    somewhat clayey, medium compact, plant remains

    are very rare, somewhat fine sandy

    3060 cm dark brown-dark gray fine detritus mud,

    somewhat clayey, medium compact, several grass

    roots

    6076 cm dark brown-dark gray clayey fine detritus

    mud, compact, plant remains are very rare,

    somewhat fine sandy

    7681 cm dark brown-dark gray clay with fine

    detritus mud, very compact, macroscopic plant

    remains not visible, somewhat fine sandy

    8195 cm white very fine silt, clayey, very compact.

    Chronological control

    The AMS radiocarbon dates of the two lake cores are

    listed in Table 1. The 6 AMS dates from Laguna El

    Pinal indicate glacial and Holocene lake deposits with

    low sedimentation rates. The basis of the profile falls

    during the Last Glacial Maximum (LGM), dated at

  • 8/6/2019 triar (1135)

    6/16

    466

    18,290 90 14C yr B.P. The lowermost 20 cm of the

    core represents 8000 yrs of the glacial period, suggesting

    a very low sedimentation rate. Gaps, probably becauseof the dry environment (ephemeral lake, see discussion

    of the results), are possible. The Late Glacial/Holocene

    transition is well represented, but mid and late Holocene

    deposits are apparently incomplete. Deposits of the

    last 1000 yrs are not represented in the core. Similar

    ages of AMS dates at 24 cm depth of 1065 14C yr B.P.,

    and at 2 cm depth of 1261 14C yr B.P. suggest that the

    topcore sediments are disturbed or absent. The

    stratigraphy does not show a clear indication for gaps

    in sedimentation.

    The 6 AMS dates from Laguna Carimagua indicate

    the deposit to cover the Holocene. The pollen record

    begins at 80 cm depth, dated 8270 14C yr B.P. The AMSdate from the sample of the base of the core at 94 cm

    depth is only 7830 14C yr old and might be contaminated

    during core recovery, or by younger carbon from

    decomposed rootlets. This date has been excluded for

    the age calculations of the pollen zones. The

    sedimentation rate in the uppermost part of the core is

    very low and may suggest that sediment accumulation

    is incomplete. The last c. 1400 yrs are not represented

    in the core.

    Laguna El Pinal

    The pollen diagram of core El Pinal shows the most

    abundant fossil pollen and spore taxa out of the 95

    different types which have been identified (Figures 5a

    & 5b). About 16 pollen types are still unknown. Pollen

    concentration and pollen influx values are low in the

    lower part of the core. Pollen concentration is higher

    in the middle part of the core. Highest concentrationand influx values are found in the uppermost core

    section. Charcoal particles are abundant in all pollen

    samples, but are not quantitatively documented for

    reasons mentioned earlier.

    Based on cluster analysis, 3 pollen zones were

    recognized in the pollen record of Laguna El Pinal:

    zone ELP-I (7260 cm, c. 18,29010,690 14C yr B.P.,

    6 samples), zone ELP-II (6049 cm, 10,6909340 14C yr

    B.P., 5 samples), and zone ELP-III (490 cm, 9340105014C yr B.P., 25 samples).

    The pollen record of Laguna El Pinal is char-

    acterized throughout by savanna herb pollen types (95

    80%), dominated by Poaceae, lower percentages ofCyperaceae, Asteraceae and several other herb taxa.

    Highest values of Poaceae (8592%) are found in

    zone ELP-I. The group of forest and gallery forest

    shrubs and trees show a low representation between 5

    and 17%. These shrub and tree percentages increase

    from 5% in zone ELP-I to 1015% in zone ELP-II.

    The group of forest and gallery forest shrubs and trees

    consists mainly ofMauritia-type, Alchornea, Are-

    caceae and Cecropia (highest percentages in zone

    ELP-I), Melastomataceae, Myrtaceae and Celtis

    (highest values in zone ELP-II), and some other taxa

    with low representation. The representation of the groupof savanna shrubs and trees,Byrsonima, Curatella and

    Didymopanax is very low, with values of < 12%.

    Fern spores, except forIsoetes, are found in low

    percentages.Isoetes spores are absent or rare in zone

    Table 1. List of AMS radiocarbon dates of the samples of cores Laguna El Pinal and Laguna Carimagua, Llanos Orientales, Colombia

    Lab. number Depth (cm) 14C yr B.P. 13C/12C r. Calendar age (cal B.P.)

    Laguna El Pinal

    UtC-5830 2 1065 36 27.1 979935

    UtC-5831 24 1261 36 27.9 12601162

    UtC-5832 48 9020 80 27.3 100359930

    UtC-5833 52 10280 50 27.8 1224711963

    UtC-5834 62 10790 60 25.6 1279712638

    UtC-5481 72 18290 90 22.6 20093 19704

    Laguna Carimagua

    UtC-6075 0 1357 36 26.7 12951269

    UtC-6076 11 4408 38 27.9 50405006, 49954872

    UtC-6077 32 4999 40 28.2 58475834, 57495700,

    56975663

    UtC-6078 57 5522 41 27.6 63886376, 63146288

    UtC-6079 80 8270 60 26.2 93739194, 91719152,

    91289093, 90689049

    UtC-4907 94 7830 60 26.7 86418637, 86048496

  • 8/6/2019 triar (1135)

    7/16

    467

    ELP-I, abundant in zone ELP-II and less frequent in

    zone ELP-III. Aquatic taxa are mainly represented by

    Sagittaria, Eichhornia and in low percentagesUtricularia, Typha, Cabomba and Nymphoides.

    Pollen grains of aquatics are rare or absent in zone

    ELP-I, increase to 7% in zone ELP-II, and to 20% in

    zone ELP-III. Cabomba, except for only one single

    grain, is only found in zone ELP-III. Algal colonies of

    Botryococcus are represented in all samples, but in the

    lower part of the core percentages are higher. Fungal

    spores show relatively low percentages.

    Laguna Carimagua

    The pollen diagram from Laguna Carimagua shows the

    most abundant fossil pollen and spore taxa out of the98 identified types, including 14 unknown types

    (Figures 6a & 6b). The downcore values of the pollen

    concentration are relatively stable, but increase

    slightly in the uppermost part of the core. Pollen influx

    is highest in the middle part of the core. Charcoal

    particles are abundant in all pollen samples, but are not

    quantitatively documented for reasons mentioned

    earlier.

    Based on the cluster analysis and major changes

    in the pollen assemblages, 3 pollen zones were rec-

    ognized in Laguna Carimagua: zone CAR-Ia (8057 cm,

    8270557014

    C yr B.P., 12 samples), zone CAR-Ib(5711 cm, 55703850 14C yr B.P., 23 samples), and

    zone CAR-II (110 cm, 38501360 14C yr B.P., 6

    samples).

    The pollen record of Laguna Carimagua is marked by

    a high representation of pollen of savanna herbs (80

    90%), primarily Poaceae and Cyperaceae. Percentages

    decrease from 80% in zone CAR-Ib to 50% at the end

    of zone CAR-II. The group of forest and gallery forest

    shrubs and trees, primarily represented by pollen grains

    of Mauritia-type, Mauritiella-type, Moraceae/

    Urticaceae, Melastomataceae,Alchornea and Cecropia

    show low percentages between 10 and 20%, but

    increase up 50% in zone CAR-II. Related to this

    ecological group there are a few differences between

    subzones CAR-Ia and CAR-Ib: slightly higher per-

    centages of the Mauritiella-type in zone CAR-Ia, and

    slightly higher percentages of taxa such as Mimosa,

    Acalypha and Macrolobium-type in zone CAR-Ib.

    Comparing pollen zones CAR-I and CAR-II, it is obvious

    that zone CAR-II is characterized by a high repre-

    sentation ofMauritiella-type, higher percentages of

    Mauritia-type,Alchornea, and Cecropia, and decreasing

    percentages of Poaceae,Eichhornia andIsoetes.

    Pollen grains of Byrsonima, Curatella, Didy-

    mopanax, which belong to the group of savanna shrubs

    and trees, are very low in the pollen record (< 12%).Aquatic taxa are mainly represented by Sagittaria and

    Eichhornia, but also by a few pollen grains from

    Ludwigia, Typha,Polygonum andNymphoides. Apart

    from the significantly low percentages ofEichhornia

    in zone CAR-II, aquatic taxa show little changes in the

    downcore representation. The curve forBotryococcus,

    as well as fungal spores, show clear fluctuations in the

    pollen diagram. Percentages of the aquatic fernIsoetes

    are highest in the lower part of the record and decrease

    in zone CAR-Ib to be almost absent in zone CAR-II.

    Reconstruction of paleoenvironmental change and

    discussion of the results

    From the LGM to the Late Glacial very little or no

    pollen grains of aquatic taxa are present, and sub-

    mersed aquatic Isoetes was absent in the Laguna El

    Pinal. This suggests that the lake existed mainly during

    the wet season, reflecting a dry environment with a

    marked long dry season. Poorly preserved pollen

    grains in the sediments representing this time period

    might be caused by such an ephemeral lake. The marked

    increase ofIsoetes spores and the presence of aquatic

    taxa such as Sagittaria points to the formation of a

    permanently shallow lake during the Late Glacialperiod from 10,690 14C yr B.P. onward. Precipitation

    was probably higher and the dry season shorter than

    during the previous period. It is interesting to mention

    that in the province of Arauca, some 300 km further

    north near the Venezuelan border, shallow lakes dry out

    completely during the dry season and do not contain

    any deposits yielding pollen. In Arauca annual pre-

    cipitation (about 1200 mm) is lower and the dry

    season (about 5 months) is longer than in region of the

    lakes El Pinal and Carimagua.

    The pollen record of El Pinal indicates the pre-

    dominance of grassland savanna during the recorded

    part of the last glacial and the Holocene. Although an

    unknown fraction of the carbonized particles was lost

    during sample preparation, the still high abundance of

    these particles throughout the sediment core indicates

    a high burning frequency. Savanna shrubs and trees were

    apparently not frequent in the grasslands. Comparing

    the proportions of the different ecological groups,

    forested areas surrounding lakes or occurring along

    small water courses were relatively rare. The pollen

    record shows a Late Quaternary savanna ecosystem

    with little change, but during the Late Glacial, around

  • 8/6/2019 triar (1135)

    8/16

    468

    Figure5a.Percentagepollendiagramo

    fcoreLagunaElPinal,LlanosOrientales,Col

    ombia.

  • 8/6/2019 triar (1135)

    9/16

  • 8/6/2019 triar (1135)

    10/16

    470

    Figure6a.Percentagepollendiagramof

    coreLagunaCarimagua,LlanosOrientales,Colombia.

  • 8/6/2019 triar (1135)

    11/16

    471

    Figure6b.Summarypollendiagramincludingpollenconcentrationandpolleninfluxr

    ecords,andtheclusteranalysisdendrogramo

    fLagunaCarimaguna.

    Mauritia-typecurvecolored

    blackinthesumo

    fforestandgalleryfo

    restshrubsandtrees.

  • 8/6/2019 triar (1135)

    12/16

    472

    10,690 14C yr B.P., the area of forest and/or gallery

    forest along the rivers increased probably about 50%,

    reflecting a change to wetter climatic conditions. Incomparison to the extensive savanna grassland at that

    time, areas of forest were still very small. At the Late

    Glacial to Holocene transition, the pollen record from

    El Pinal (zones ELP-II to ELP-III) shows a slight

    change in forest composition: reduced presence of

    Melastomataceae, Myrtaceae, and Celtis, and higher

    presence ofMauritia-type and Alchornea. More

    abundant Mauritia palm vegetation suggests wetter

    conditions during the Holocene than during the Late

    Glacial.Mauritia, as well asMauritiella, can be seen

    as an indicator of higher moisture availability (Kahn

    & De Granville, 1992; Henderson, 1995). The decline

    ofIsoetes and the occurrence of the sea rose Cabomba(Cabombaceae; an aquatic with rhizomes, submersed

    stem, and floating leaves; Maas & Westra, 1993) in the

    lake also indicates higher lake levels during the

    Holocene.

    The Holocene part of the Laguna Carimagua record

    is similar to the record of El Pinal, and both lakes show

    a well documented and similar history of the regional

    ecosystem. Except for the late Holocene period, there

    are only minor changes in the vegetation of the savanna

    ecosystem during the Holocene. The cluster analysis

    dendrogram indicates a slight change around 5570 14C

    yr B.P., at the transition from zone CAR-Ia to CAR-Ib.The lower abundance of the palm Mauritiella in zone

    CAR-Ib may suggest slightly drier conditions, but the

    slightly higher presence of several other taxa of the

    forests and gallery forests may indicate wetter con-

    ditions than in zone CAR-Ia. Hence we cannot arrive at

    a conclusive interpretation for these changes.

    A marked change occurred around 3850 14C yr B.P.

    (interpolated age), primarily by the sudden high

    abundance ofMauritiella, but also some other taxa

    with an increased representation, such as Mauritia,

    Alchornea and Cecropia. This change is related to

    wetter conditions, i.e. an increase in precipitation and/

    or a shorter dry season. The increase of the pioneer

    and disturbance indicator Cecropia points to an

    increase in human impact on the vegetation after 385014C yr B.P. (zone CAR-II). Interestingly enough, this

    vegetational change is not recorded in Laguna El Pinal.

    Pollen analysis of additional cores from this area are

    in preparation and may provide a more substantiated

    reconstruction of the beginning of human impact in

    this part of the Llanos Orientales.

    The present-day vegetation around the lake consists

    of dense stands ofMauritia, but this is not reflected

    in the pollen record. Some disturbance of the upper

    core section was already inferred from the AMS dates,

    and the absence of a clearMauritia peak in the pollenrecord may be another argument to suspect that the

    record for the last thousand of years is incomplete. We

    conclude that at least the last 1000 yrs are incomplete,

    or missing, in both lake records. Perhaps the organic

    rich sediments experienced a high level of decom-

    position because of the shallow water.

    Comparison with other pollen records of the

    Llanos Orientales

    The distance from our study lakes to Laguna Angel

    (Behling and Hooghiemstra, 1998) is about 100 km

    and about 200 km to Laguna Sardinas (Behling andHooghiemstra, 1998), and about 250 km to Laguna de

    Agua Sucia (Wijmstra & Van der Hammen, 1966)

    (Figure 1). The setting of Laguna Angel (428N,

    7034W, 200 m a.s.l.) is comparable to the lakes in

    this study: located between two larger rivers at some

    45 km distance. Today, the forested area near Laguna

    Angel and Laguna Sardinas is larger than in the region

    of El Pinal and Carimagua. Laguna Sardinas (458N,

    6928W, 80 m a.s.l.) is an isolated lake at the border

    of a flood plain area and close to large areas with gallery

    forest along a complex river system. Therefore, the

    setting of Laguna Sardinas is different compared to thelakes Angel, El Pinal and Carimagua.

    The record of Laguna Angel (5 pollen zones) starts

    at 10,030 14C yr B.P., and that of Laguna Sardinas (6

    pollen zones) at 11,570 14C yr B.P. The formation of

    permanent lakes during the Late Glacial, as in Laguna

    El Pinal, occurs at the same time and suggests wetter

    climatic conditions than during the preceding full

    glacial.

    The palynological records from Lagunas Angel and

    Sardinas (Figure 7) document, since the Late Glacial,

    the predominance of grassland savanna with a small

    contribution of savanna shrubs and trees, such as

    Curatella,Byrsonima and Waltheria. The proportion

    of grassland savanna, in comparison to forest and

    gallery forests, is relatively constant after c. 11,00014C yr B.P. This relatively stable history of the savanna

    ecosystem is supported by the newly studied sites,

    but proportions of forest and/or gallery forest are

    markedly lower in the region of El Pinal and Carimagua

    than in region of Lagunas Angel and Sardinas. Further,

    the floristic composition of the gallery forest in the

    area of Lagunas Angel and Sardinas is subject to

    considerable change throughout the records, which

  • 8/6/2019 triar (1135)

    13/16

    473

    Figur

    e7.SummarypollendiagramsofLagunaAngelandLagunaSardinas(afterBehlingandHo

    oghiemstra,1998).

  • 8/6/2019 triar (1135)

    14/16

    474

    contrasts to our newly studied sites. In the Laguna

    Sardinas record, Alchornea significantly increases

    during zone SAR-IIa (10,68010,07014

    C yr B.P.).Alchornea also occurs more frequently in the gallery

    forests during zone SAR-IIb (10,0709390 14C yr

    B.P., the beginning of the Holocene), than in other

    pollen zones. Further, the slightly higher presence of

    theMauritia palm in the pollen zones SAR-I to SAR-

    IIb may indicate wetter climatic conditions at the

    transition from the Late Glacial to the Holocene. Such

    vegetational changes are not found in the pollen record

    of Laguna El Pinal, but other changes in El Pinal, e.g.

    expansion of forest, also suggest wetter conditions

    since c. 10,700 14C yr B.P. Sediments from Laguna

    Sardinas, located at the edge of the floodplain area,

    might be better suited to register signals of climatechange, expressed by changes and/or expansion of

    gallery forests, than Laguna El Pinal, which is located

    in an almost unforested grassland savanna with little

    gallery forest.

    A similar situation also prevails in the Holocene.

    During the early and middle Holocene zones ANG-II

    (97305260 14C yr B.P.) and SAR-III (93906390 14C

    yr B.P.), savanna herbs are slightly more abundant than

    during preceding periods. This period apparently re-

    presents the maximum extension of grassland savanna

    (Figure 6). There is no, or little, presence ofMauritia

    during this period. This pollen evidence reflects thedriest period in the region of Lagunas Angel and

    Sardinas, probably related to low rainfall rates and/or

    an extended annual dry season. Such a vegetational

    signal is not found in our newly studied sites.

    The period represented by zones ANG-III (5260

    3890 14C yr B.P.) and SAR-IV (63903680 14C yr B.P.),

    shows a decrease of savanna herbs and an increase of

    forest and gallery forest taxa. In the floodplains near

    Laguna Sardinas, gallery forests expanded and the

    surroundings of the lake became more forested. Wetter

    climatic conditions, leading to a higher soil moisture

    content, may have been an important factor to make this

    vegetational change possible (Ross et al., 1992). This

    climatic signal is not found in El Pinal, whereas in

    Carimagua, only minor changes have been recorded

    around 5570 14C yr B.P. without clear climatic impli-

    cations. In view of the data of the other sites, it might

    be plausible that the minor changes in the Carimagua

    record indeed are related to a change to a wetter climate,

    as evidenced by the records of Laguna Angel at 526014C yr B.P. and Sardinas at 6390 14C yr B.P.

    The late Holocene period, represented by the zones

    ANG-IV (3890 to c. 2000 14C yr B.P.) and SAR-V

    (3680 14C yr B.P. to modern), is incomplete or missing

    in both of these sites. The same situation holds for our

    two new sites, probably because of the same reasons(high level of decomposition of the organic deposits).

    The pollen records of Lagunas Angel and Sardinas show

    that the most significant change occurred in the

    uppermost pollen zone, primarily reflected by the

    marked increase ofMauritiella in the area of Laguna

    Angel and Mauritia in the area of Laguna Sardinas.

    These changes have been interpreted as a result of an

    increased human impact on the vegetation by an

    intensified fire regime under wetter climatic con-

    ditions (Behling and Hooghiemstra, 1998). A similar

    environmental change is reflected in the record of

    Laguna Carimagua around 3850 14C yr B.P., by an

    strong increase of stands ofMauritella palms and theincreased presence ofMauritia. An increased presence

    of palms is also reported from other sites in northern

    South America (Rull, 1991, 1992; Behling and Costa,

    1997).

    The first pollen record from the Colombian savannas,

    Laguna de Agua Sucia at 260 m elevation (Wijmstra &

    Van der Hammen, 1966), lies at about 250 km distance

    in the southeastern part of the Llanos Orientales

    (Figure 1). This site lies in a region with higher

    precipitation and a shorter dry season than is the case

    in the region of the lakes Sardinas, Angel, El Pinal and

    Carimagua. Laguna de Agua Sucia is also at shorterdistance to the Amazonian rain forest. The extrapolated

    radiocarbon age at the base of this 5 m long core is

    about 5100 14C yr B.P. Vegetational composition

    (more shrubby savannas), and vegetational change

    during the last 5000 yrs are somewhat different, but

    inferred climatic changes are comparable to those of

    the region of the central and central-eastern part of the

    Llanos Orientales (Behling and Hooghiemstra, 1998).

    Conclusions

    The new pollen record of Laguna El Pinal extends the

    environmental history of the savannas of the central

    Llanos Orientales back to the LGM. Grassland savanna

    vegetation, dominated by Poaceae, and with very little

    occurrence of woody taxa such as Byrsonima and

    Curatella, characterized the landscape in this area

    since the LGM. The proportions of gallery forest along

    river courses and/or forested areas in between river

    systems have been always very small. During the Late

    Quaternary, the floral composition of the savannas was

    relatively stable in the studied areas.

  • 8/6/2019 triar (1135)

    15/16

    475

    From the LGM until 10,690 14C yr B.P., Laguna El

    Pinal was a shallow ephemeral lake, reflecting the

    driest climatic conditions of the last 18,000 yrs: theannual precipitation must have been low and the annual

    dry season long. Laguna El Pinal developed into a

    permanent shallow lake during the Late Glacial, around

    10,690 14C yr B.P. At other places in the Llanos

    Orientales, permanent lakes also came into existence

    during the same period (Laguna Angel around 10,03014C yr B.P., Laguna Sardinas around 11,570 14C yr B.P.),

    indicating a change to wetter climatic conditions. The

    increase of the forested area in the region of Laguna

    El Pinal supports this interpretation. The vegetational

    change in the area of Laguna Sardinas, i.e. an increased

    presence ofMauritia andAlchornea, during the Late

    Glacial, has not been observed in the area of El Pinal.This is probably because of the greater distance to

    major river systems with gallery forests of the last

    mentioned site.

    Higher lake-levels in Laguna El Pinal, and the start

    of the pollen record in Laguna Carimagua at 8270 14C

    yr B.P., are indicative of wetter conditions during the

    Holocene than during full glacial times. Minor changes

    in the vegetation of the area of Laguna Carimagua at

    5570 14C yr B.P. may be considered comparable with

    an increase of the forested area in the record of

    Laguna Angel at 5260 14C yr B.P., and in the record of

    Laguna Sardinas at 639014

    C yr B.P., and reflects mostpossibly a change to wetter climatic conditions. The

    early Holocene was somewhat drier than the following

    period until about 3850 14C yr B.P. The late Holocene

    period, as far as recorded, is characterized by an

    increase of the stands of the palm Mauritiella in the

    surroundings of Laguna Carimagua. A marked increase

    of palms in the regional vegetation was also observed

    in the areas of Lagunas Angel and Sardinas, and in

    other regions in tropical South America. This general

    vegetational change is interpreted as reflecting in-

    creased human impact on the vegetation under the

    wettest Holocene climate regime.

    Acknowledgments

    The authors acknowledge Pedro Botero (IGAC) and

    Carlos Botero for organizing the impressive coring

    expedition and continuous valuable assistance in this

    remote area of Colombia. Guido van Reenen (Am-

    sterdam) assisted during part of the fieldwork. Elly

    Beglinger and Annemarie Phillip (Amsterdam) are

    thanked for preparing the pollen samples. Antoine

    Cleef (Amsterdam) provided valuable information

    during the evaluation of the data. Klaas van der Borg

    (Utrecht) is acknowledged for the radiocarbon dates.The director of the Tropenbos-Colombia office, Juan

    Saldarriaga, is thanked for the hospitality and con-

    tinuous logistic support during field expeditions. The

    Embassy of The Netherlands in Bogot is thanked for

    assistance with sample transport. We thank John P.

    Smol and two unknown reviewers for constructive and

    valuable comments on the manuscript. We acknowledge

    The Netherlands Foundation for Scientific Research

    (NWO/GOA) for financial support (project number

    750.195.10 to H. Hooghiemstra).

    References

    Behling, H., 1993. Untersuchungen zur sptpleistoznen und

    holoznen Vegetations- und Klimageschichte der tropischen

    Kstenwlder und der Araukarienwlder in Santa Catarina

    (Sdbrasilien). Dissertationes Botanicae, 206: J. Cramer, Berlin

    Stuttgart, 149 pp.

    Behling, H. & M. L. da Costa, 1997. Studies on Holocene tropical

    vegetation, mangrove and coast environments in the state of

    Maranho, NE Brazil. Quat. S. Am. Antarctic Pen. 10: 93118.

    Behling, H. & H. Hooghiemstra, 1998. Late Quaternary paleoecology

    and paleoclimatology from pollen records of the savannas of the

    Llanos Orientales in Colombia. Palaeogeography, Palaeo-

    climatology, Palaeoecology 139: 251267.

    Blydenstein, J., 1967. Tropical savanna vegetation of the Llanos of

    Colombia. Ecology 48: 115.

    Clapperton, C., 1993. Quaternary geology and geomorphology of

    South America. Elsevier, Amsterdam, 779 pp.

    Cole, M. M., 1982. The influence of soils, geomorphology and geology

    on the distribution of plant communities in savanna ecosystems.

    In: Huntley, B. J. & B. H. Walker (eds), Ecology of Tropical

    Savannas. Ecological Studies, 42: 145174. Springer, Berlin.

    Cole, M. M., 1986. The savannas: biogeography and geobotany.

    Academic Press, London, 438 pp.

    Cuatrecasas, J., 1989. Aspectos de la vegetacin natural en Colombia.

    Perez-Arbelaezia, II (8), 155283. Jardn Botnico de Bogot

    Jose Celestino Mutis.

    Faegri, K. & J. Iversen, 1989. Textbook of pollen analysis. (4th ed.);

    John Wiley & Sons, Chichester, 328 pp.

    Furley, P. A., 1992. Edaphic changes at the forest-savanna boundary

    with particular reference to the neotropics. In: Furley, P. A. et

    al. (eds), Nature and Dynamics of Forest-Savanna Boundaries.

    Chapman & Hall, London, 91117.

    Furley, P. A., Proctor, J. & J. A. Ratter, (eds), 1992. Nature and

    Dynamics of Forest-Savanna Boundaries. Chapman & Hall,

    London, 616 pp.

    Gentry, A. H., 1993. Woody plants of northwest South America.

    Conservation International, Washington, DC, 895 pp.

    Grimm, E. C., 1987. CONISS: A Fortran 77 program for stratigraphically

    constrained cluster analysis by the method of the incremental

    sum of squares. Computers and Geosciences 13: 1335.

    Henderson, A., 1995. The palms of the Amazon, Oxford University

    Press, New York, 362 pp.

  • 8/6/2019 triar (1135)

    16/16

    476

    Herrera, L. F. & L. E. Urrego, 1996. Atlas de polen de plantas tiles y

    cultivadas de la Amazonia colombiana (Pollen atlas of useful

    and cultivated plants in the Colombian Amazon region).

    Estudios en la Amazonia Colombiana, XI: 462 pp. Tropenbos-Colombia, Bogot (also published in The Quaternary of

    Colombia, 24 (ed. H. Hooghiemstra), Amsterdam).

    Hooghiemstra, H., 1984. Vegetational and climatic history of the high

    plain of Bogot, Colombia: a continuous record of the last 3,5

    million years. Dissertationes Botanicae, J. Cramer, Vaduz 79:

    368 pp.

    Huber, O. 1987. Neotropical savannas: their flora and vegetation.

    Tree 2: 6771.

    Hueck, K. & P. Seibert, 1972. Vergetationskarte von Sdamerika.

    Fischer, Stuttgart, 71 pp + map.

    Hueck, K., 1966. Die Wlder Sdamerikas. Fischer, Stuttgart, 422 pp.

    Kahn F. 1987. The distribution of palms as a function of local

    topography in Amazonian terra-firme forests. Experientia 43:

    251259.

    Kahn, F. & J.-J., de Granville, 1992. Palms in forest ecosystems of

    Amazonia. Springer, Berlin, 226 pp.

    Maas, P. J. M. & L. Y. T. Westra, 1993. Neotropical plant families. A

    concise guide to families of vascular plants in the Neotropics.

    Koeltz Scientific Books, Koenigstein, Germany, 289 pp.

    Markgraf, V., 1993. Climatic history of Central and South America

    since 18,000 yr B.P: Comparison of pollen records and model

    simulations. In: Wright, H. E. Jr., Kutzbach, J. E., Webb III, T.

    Ruddiman, W. F., Street-Perrott, F. A. & P. J. Bartlein (eds).

    Global Climates Since the Last Glacial Maximum. University of

    Minnesota Press, Minneapolis Londen, 357385.

    Medina, E. & J. F. Silva, 1990. Savannas of northern South America: a

    steady state regulated by water-fire interactions on a background

    of low nutrient availability. J Biogeogr 17: 403413.

    Pinto-Escobar, P., 1993. Vegetacin y flora de Colombia. Editora

    Guadalupe, Bogot, 72 pp.

    Ross, S. M., Luizao, F. J. & R. C. C. Luizao, 1992. Soil conditions and

    soil biology in different habitats across a forest-savanna

    boundary on Maraca Island, Roraima, Brazil. In: Furley, P. A. et

    al. (eds), Nature and Dynamics of Forest-Savanna Boundaries.

    Chapman & Hall, London, 145170.

    Roubik, D. W. & J. E. Moreno, 1991. Pollen and spores of Barro

    Colorado Island. Missouri Botanical Garden, 36, 270 pp.

    Rull, V., 1991. Contribucin a la paleoecologia de Pantepui la Gran

    Sabana (Guayana Venezolana): clima, biogeografa, y ecologa.Scientia Guaianae, 2: 1333.

    Rull, V., 1992. Successional patterns of the Gran Sabana (southeastern

    Venezuela) vegetation during the last 5000 yrs, and its responses

    to climatic fluctuations and fire. J. Biogeogr. 19: 329338.

    Sarmiento, G. & M. Monasterio, 1975. A critical consideration of the

    environmental conditions associated with the occurrence of

    savanna ecosystems in tropical America. In: Golley, F. B.& E.

    Medina (eds), Tropical Ecological Systems; Trends in Terrestrial

    and Aquatic Research. Springer, Berlin, 223250.

    Sarmiento, G., 1983. The savannas in tropical America. In: Bourlire,

    F. (ed), Tropical Savannas. Elsevier, Amsterdam, 245288.

    Schubert, C. 1988. Climatic changes during the Last Glacial Maximum

    in northern South America and Caribbean: a review. Interciencia

    13: 128137.

    Snow, J. W., 1976. The climate of northern South America. In:

    Schwerdtfeger, W. (ed.), World Survey of Climatiology, 12.

    Climates of Central and South America. Elsevier, Amsterdam,

    295403.

    Tissot, C. & C. Marius, 1992. Holocene evolution of the mangrove

    ecosystem in French Guiana: a palynological study. In: Singh,

    K. P. & J. S. Singh (eds), Tropical Ecosystems: Ecology and

    Management. Wiley Eastern Limited, New Delhi, 333347.

    Van der Borg, K., Alderliesten, A., Harnton, C. M., De Jong, A. F. & N.

    A. Van Zwol, 1987. Accelerator mass spectrometry with 14C and10Be in Utrecht. Nuclear Instrumental Methods B 29: 143145.

    Van der Hammen, T., 1963. A palynological study on the Quaternary

    of British Guyana. Leidse Geologische Mededelingen 29: 126

    168.

    Vareschi, V., 1980. Vegetationskologie der Tropen. Ulmer, Stuttgart,

    293 pp.

    Wijmstra, T. A., 1971. The palynology of the Guiana coastal

    basin. Ph.D. Dissertation, University of Amsterdam, 63 pp.

    Wijmstra, T. A. & T. Van der Hammen, 1966. Palynological

    data on the history of tropical savannas in northern South

    America. Leidse Geologische Mededelingen 38: 7190.