Preliminary palynological analysis of a Holocene peat bog from Apakará-tepui ( Chimantá Massif , Venezuelan Guayana )

Preliminary palynological analysis of a Holocene peat bog from Apakará-tepui (Chimantá Massif, Venezuelan Guayana).This paper reports the preliminary palynological results, at a millennial scale, of a Holocene peat bog sequence, since around 8.0 cal kyr BP to the present, obtained in the summit of the Apakará-tepui (2170 m elevation), in the Chimantá massif, located in the neotropical Venezuelan Guayana. The early Holocene was characterized by a vegetation different to the present, in which trees and shrubs dominated and Myrica (Myricaceae) was the main element of the gallery forests around a permanent water body, as indicated by the continuous presence of Isoëtes (Isoëtaceae) in high percentages. Around the middle Holocene (5.3 cal kyr BP), a shift towards more herbaceous and non-flooded communities occurred, and the present day vegetation established. This has been interpreted as a shift from warmer and wetter climates to cooler and drier conditions. The first phase, from 8.0 to 5.3 cal kyr BP, falls within a warming phase widely documented worldwide, known as the Holocene Thermal Maximum. The millennial trends shown here will be refined with further studies at centennial to decadal time scales. These results support the hypothesis that the best sites to detect paleoenvironmental changes in the summits of the tabular Guayana mountains are close to altitudinal ecotones.


INTRODUCTION
Peat bogs in the summits of the table mountains (tepuis) from the neotropical Guayana Highlands (Fig. 1) have been considered a priori excellent locations to record late Pleistocene and Holocene paleoecological and paleoenvironmental changes, due to its pristine nature, which excludes human interference (Rull, 2007a;2010).So far, however, palynological studies conducted on these summits have not been able to find consistent regional patterns of climate/vegetation in time and space.Indeed, whereas most records from the summits of the Chimantá massif (Akopán-tepui, Amurí-tepui and Toronó-tepui) have shown heterogeneous patterns of change likely due to the dominance of local micro-and meso-climatic processes over general paleoclimatic trends (Rull, 2005a), other sequences (Churí-tepui and Guaiquinima massif) suggested vegetation shifts linked to temperature and moisture variations of potential regional extent (Rull, 2004;2005b).The main reason for this disagreement was assumed to be the different degree of sensitivity of the localities surveyed to past environmental changes, depending on their elevation.The better suited sites to document paleoecological shifts associated to climate changes were considered to be close to key altitudinal ecotones, and the highest mountaintops (Rull, 2005a).For example, in the Churí-tepui, the study site was close to the upper altitudinal boundary of grasslands dominated by Stegolepis Kloszch ex Körn (Rapateaceae), in their contact with the Chimantaea Maguire, Steyerm.& Wurdack (Asteraceae) shrublands (Rull, 2004).On the contrary, localities situated within the grasslands and far from this ecotone, as for example the other Chimantá sites, would be less useful to record paleoenvironmental changes (Rull, 2005a).A study in one of the highest Chimantá mountaintops (Eruoda-tepui) also showed remarkable vegetation constancy during the Holocene; therefore, ecotonal localities seem to be the preferred tepuian sites for paleoecological reconstruction (Nogué et al., 2009a).
This paper reports a preliminary palynological study in the Apakará-tepui (Chimantá massif), which is close to the mentioned grasslandshrubland altitudinal ecotone, at an elevation similar to the Churí-tepui, where a conspicuous vegetation shift was recorded during the late Holocene, and attributed to a temperature shift.The section embraces the last ~8.0cal kyr BP (calendar kilo-years before present) and the resolution is millennial, thus recording the main Holocene trends.Further studies at higher resolution will refine the vegetational and climatic history of the site.This type of studies are useful not only to increase the ecological and paleoecological knowledge of these singular neotropical ecosystems but also to improve biodiversity conservation strategies in the face of the predicted global warming (Rull & Vegas-Vilarrúbia, 2006;Nogué et al., 2009b;Rull et al., 2009;Vegas-Vilarrúbia et al., 2011).

Coring site
The coring site is located on the summit of the Apakará-tepui, in the Chimantá massif (5º 19' 21.68" N -62º 13' 34.14" W, 2170 m elevation).It lies within a relatively narrow valley modeled by an affluent of the Tírica river running to the SE (Figs. 1 & 2).The site is part of the Guayana Highlands (those situated above 1500 m elevation), under mesothermic ombrophylous climate with annual average temperatures between 12 and 18 ºC and very high rainfall (2500-3500 mm per year in average), without any true dry season.Additional moisture is provided by frenquent and dense mists.
Winds and thunderstorms are also frequent, especially when the Intertropical Convergence Zone (ITCZ) is around the region (March to November) (Huber, 1995a).
In the coring site, other shrubs as for example Paepalanthus chimantensis Hensold (Eriocaulaceae), and Sphagnum L. (Sphagnaceae) mosses, are worth mentioning.These mosses do not form real Sphagnum peat bogs, as in many other similar environments, but may be locally abundant, especially along water courses (Ahti, 1992) The Apakará coring site studied here lies close to the upper altitudinal boundary of the tepuian meadows dominated by Stegolepis ligulata, which is around 2200-2300 m (Huber, 1992b) (Cyrillaceae), as the main representatives.Stegolepis ligulata may occur slightly above the upper boundary of the meadows it dominates, but in small and scattered stands (Huber, 1987).

Sampling and pollen-analytical procedures
A 200 cm peat core called PATAM9-A07 was obtained with a Russian borer.Samples for pollen analysis were taken in the field at every 2 cm.In this preliminary analysis, samples are analyzed every 10 cm.A surface sample was taken in the coring point for comparison.Around 3 g of each of these samples were processed after addition of Lycopodium L. (Lycopodiaceae) tablets (batch nº 177745; 18,584 spores/tablet).The laboratory processing included KOH, HCl and HF digestions, and acetolysis; slides were mounted with silicone oil without sealing (Nogué et al., 2009a;Rull et al., 2010b).The palynomorphs considered were pollen, pteridophyte spores, algae and animal remains, and fungal spores.Counting was conducted until a minimum of 300 pollen and pterophyte spores (including a minimum pollen sum of 200), and the saturation of diversity (Rull, 1987).The pollen sum included all pollen types, except those from aquatic and semi-aquatic plants.Identification was based on local and regional pollen and spore floras (Absy, 1979;Roubik & Moreno, 1991;Salgado-Labouriau & Villar, 1992;Jiménez, 1996;Herrera & Urrego, 1996;Colinvaux et al., 1999;Rull, 2003;Burn & Mayle, 2008;Leal, 2010;López-Martínez et al., 2010), some general palynological catalogs (Erdtman, 1952;Tryon & Lugardon, 1991), and the Neotropical Pollen Database (Bush & Weng, 2006).Pollen diagrams were plotted and zoned using psimpoll 4.27, the Optimal Splitting by Information Content (OSIC) method, and the broken-stick model for significance testing (Bennett, 1996).Only pollen types over 0.5% were considered for zonation.Pollen grouping according to habits and vegetation types follows Huber (1992a, b;1995b) and the Flora of the Venezuelan Guayana (Steyermark et al., 1995(Steyermark et al., -2005)), which is also the taxonomic reference.AMS radiocarbon dating was performed in the Kek Crabon Cycle AMS Laboratory (University of California, Irvine) and Beta Analytic Inc., on nine samples consisting of plant macroremains.Age calibration and the age-depth model were performed with clam.R (Blaauw, 2010), using the IntCal09.14cdatabase (Reimer et al., 2009).

Lithostratigraphy and chronology
The core is mainly composed of dark-brown peat, with some clayey intervals at the base (Fig. 4).The very basal layer (199-190 cm) is of clayey peat, followed by a ~20 cm interval of dark organic peat (gyttja).The rest of the core, from ~170 cm onwards, is dark-brown fibrous/granular peat, except for a ~5 cm clayey layer around 140 cm.The agedepth model is based on the nine radiocarbon ages measured (Table 1).The best fit was obtained with a spline smooth model (Fig. 4).Peat accumulation rates changed two times along the sequence.At the base, they were of 0.24 mm/y but they abruptly increased to 0.55 mm/y around 6100 cal yr BP, coinciding with the lithostratigraphic change from gyttja to fibrous/granular peat.A subsequent lowering to 0.13 mm/y occurred around 5200 cal yr BP, which remained until the present.

Palynology
A total of 67 pollen and 15 pteridophyte spore types were found in this study.The average count was 359 (minimum 319, maximum 422), and the average pollen sum was 265 (minimum 231, maximum 305).The surface sample (Fig. 5) is dominated by herbs, mainly Poaceae, Cyperaceae and Xyris, followed by some ligneous elements, mainly Cyrilla Garden ex L. (Cyrillaceae) and Weinmannia L. (Cunoniaceae).The general pollen diagram (Fig. 5) represents the pollen and spore types over 0.5% of the pollen sum.This diagram was subdivided into 2 pollen zones:

APK-I (199-129 cm, 8010-5310 cal yr BP)
This zone is dominated by ligneous plants (trees and shrubs), reaching around 60% of the pollen sum.Among them, the more abundant are Myrica L. (Myricaceae), Cyrilla, Ilex and Weinmannia.It is noteworthy that Myrica rarely is below 10%, while it is almost absent at present.Clusiaceae (Tovomita-type) and Chimantaea are absent or less abundant than today.Among herbs, only Poaceae are worth mentioning, with percentages up to 40%, Brocchinia and Xyris, usually present today, are very scarce or absent in this zone.Cyperaceae are also scarcer than at present.Pterydophyte spores are also scanty, except for Isoëtes that reached values over 60% of the pollen sum but experienced a dramatic decrease from about 160 cm (5680 cal yr BP) until the top of the zone.Hymenophyllum/Trichomanes is almost absent, Huperzia Bernh.(Lycopodiaceae) shows a decreasing trend but of less magnitude than Isoëtes, and Pterozonium has similar percentages than today.This zone differs from the surface sample, which represents the present-day vegetation, in the higher abundance of Myrica, Isoëtes and Pterozonium, and the lower percentages of Clusiaceae (Tovomita-type), Weinmannia, Brocchinia, Poaceae and pteridophyte spores in general.

APK-II (129-0 cm, 5310-0 cal yr BP)
The more significant change with respect to the former zone is the remarkable decrease of Myrica, which reaches values similar to today (< 1%) and almost disappears at the top of the diagram.The other ligneous elements remain with percentages similar to the former zone, except for Clusiaceae (Tovomita-type) and Melastomataceae, which experienced a slight increase.There is a general increase of herbs (60-70%), which replace the ligneous plants in the dominance of the diagram.The more evident increase correspond to Poaceae (up to 50%), Brocchinia and Xyris also showed a consistent increase.Stegolepis appears occasionally without any definite trend.Cyperaceae also increase but it does not attain present-day values (14%).Among pteridophyte spores, the more outstanding shift is the dramatic decline of Isoëtes (~1%).Cyathea Sm. (Cyatheaceae) spores slightly increase and Pterozonium declines in a similar way that Myrica and Iosëtes, but at lower magnitudes.Hymenophyllum/Trichomanes increases and attains values similar to present day conditions.This zone was subdivided into two subzones, namely a and b, with the boundary at 69 cm (2990 cal yr BP).As compared to APK-IIa, subzone APK-IIb shows a slight increase in some minor ligneous and herbaceous elements such as Clusiaceae (Tovomita-type), Chimantaea, Weinmannia and Brocchinia, as well as the appearance of Bonnetia roraimae and Psychotria L. (Rubiaceae) (absent in the rest of the diagram), and the decrease of Poaceae.In general, the zone APK-II is more similar to the surface sample.

Paleoenvironmental interpretation
According to the former results, both the environment and the vegetation were different from the present ones between 8010 and 5310 cal yr BP.The high abundance of Isoëtes suggests a very wet environment, possibly permanent flooding, in the coring site.This moisture-loving genus of pteridophytes has three species in the Guayana region, one from the lowlands (Isoëtes triangula U. Weber, 50-200 m elevation) and two from the high-tepui summits (Isoëtes killipii C. V. Morton, 1900-2800 m elevation; and Isoëtes vermiculata Hickey, 2500-2700 m elevation) (Hickey & Smith, 1995).In the Chimantá massif, Isoëtes killipii grows in moist meadows, edges of pools, streams, wet crevices near waterfalls, between 1900 and 2800 m elevation (Huber, 1992a).In the Apakará diagram the presence of clayey sediments are indicative of standing water and support the interpretation of flooded vegetation or a small shallow pond (light availability is needed for Isoëtes to develop).Therefore, climates wetter than today could be inferred.This is supported by the increase of Pterozonium, a fern typical of humid tepuian environments (Lellinger, 1995), and the absence/scarcity of Hymenophyllum/Trichomanes, two genera of lithophilous ferns, in the Chimantá massif (Huber, 1992b).On the other hand, the dominance of Myrica (very scarce today) and the other forest elements indicates that these forests were different in composition than today.Myrica rotundata Steyerm.& Maguire (Myricaceae) is the only species of the family Myricaceae known from the Venezuelan Guayana (Miller, 2001).This species is typical from tropical mountains and, in the Cimantá massif, it is found in mixed Bonnetia forests and forested slopes between 1900 and 2200 m.The high abundance of Myrica pollen in contrast to today may be interpreted as more favorable environmental conditions for its development.
Since the Apakará locality studied here (2170 m elevation) is close to the upper boundary limit of Myrica rotundata, a climate warmer than today could be assumed for the early to mid-Holocene interval studied here.In summary, the more likely climate for the 8010-5310 cal yr BP interval is warmer and wetter than today.
The upper pollen zone, ranging from 5310 cal yr BP to the present shows no remarkable differences with respect to present-day vegetation and a similar climate can be deduced.In the absence of more detailed studies, the minor oscillations in several pollen and spore types inside this zone that justified the subdivision into two subzones, before and after 2990 cal yr BP, could be tentatively attributed to minor community reorganizations, which cause is not yet known.It is noteworthy that the typical Bonnetia gallery forests did not appear until this date (2990 cal yr BP), as indicated by the first appearance of the dominant species (Bonnetia roraimae) and Psychotria, growing in the shrubby stratum of these forests (Huber, 1992a).
Comparison with other tepuian localities is problematic as most of them go back to 5-6 cal kyr BP at most, and do not record the early Holocene.Furthermore, as stated before, local reorganizations dominate over potential regional trends thus preventing generalizations (Rull, 2005a).Other areas available for comparison are the Andes, the Cariaco Basin and Lake Valencia (Rull et al., 2010a).In the Andes, oxygen isotope records from the "Lagunas Verdes" lake cores showed a shift from wet climates during the early Holocene to drier ones in the middle Holocene (Polissar et al., 2006), coinciding with the Apakará record.The Lake Valencia sediments also recorded wetter conditions during the early-mid Holocene, where maximum lake levels were recorded, with a further shift to drier conditions during the mid-late Holocene (Curtis et al., 1999).The same trends were found in the Cariaco Basin using Titanium and Iron records, showing the occurrence of wetter climates during the early Holocene and a shift towards drier conditions since around 5.4 cal kyr BP (Haug et al., 2001), which matches almost exactly with the Apakará sequence.These moisture fluctuations have been attributed to the latitudinal migration of the Intertropical Convergence Zone (ITCZ), which is one of the main controls of the precipitation amount and regime in the Neotropics (Haug et al., 2001).
Concerning temperature, the assumedly warmer Apakará phase (8.0 to 5.3 cal kyr BP) falls within the so called Holocene Thermal Maximum (HTM), recorded in the Northern Hemisphere between about 11.0 and 5.0 cal kyr BP, and attributed to changes in orbital forcing and oceanic circulation (Renssen et al., 2009).The warming trend leading to the HTM (called Early Holocene Warming or EHW) was previously documented in two localities from the adjacent Gran Sabana midlands, around 800-900 elevation, coinciding with the first record of extensive savannas in the region (Rull, 2007b;Montoya et al., 2011).Contrastingly with the apparent paleoclimatic heterogeneity recorded in many Guayana records as compared with others from the Neotropics (Rull et al., 2010a), the Apakará sequence shows a convincing agreement in both temperature and hydrological balance, at a millennial scale.Further studies will show if this situation is maintained at decadal to centennial scales.The results obtained here are encouraging for the tepui summits as suitable locations to record regional and global paleoclimatic and paleoecological trends, and support the hypothesis of the altitudinal ecotones as the more sensitive sites (Rull, 2005a).

CONCLUSIONS
Millennial-scale pollen analysis of a Holocene sequence from the summit of the Apakará summit, in the Chimantá massif, has documented the occurrence of a significant vegetation change around 5.3 cal kyr BP, which is considered to have been a consequence of an environmental change from early-mid Holocene warmer and wetter climates to mid-late Holocene cooler and drier conditions, similar to present.The warmer and wetter phase (8.0 to 5.3 cal kyr BP) agrees with most neotropical records, and also with the Northern Hemisphere Holocene Thermal Maximum.During the second phase (5.3 cal kyr BP to present), the present climates and vegetation established, with minor fluctuations.Studies at centennial to millennial scales are needed for a detailed reconstruction of the Holocene trends.This study supports the potential interest of tepui summits as paleoenvironmentally sensitive sites.

Figure 1 .
Figure 1.Location map.A) Schematic map of northern South America indicating the position of the Chimantá massif (red dot).The approximate extension of the Guayana Shield is depcited as a brown area.B) radar image of the Chimantá massif (courtesy of ANAPRO Digital), indicating the Apakará coring site studied here (yellow star) and the other localities mentioned in the text.AP -Apakará, Er -Eruoda, To -Toronó, Am -Amurí, Ak -Akopán, Ch -Churí.

Figure 2 .
Figure 2. Aerial view of the Chimantá massif showing the valley of the Tírica river in the center, and the characteristic tepuian summits at both sides (Photo: V. Rull).

Figure 3 .
Figure 3.View of a Chimantaea mirabilis shrubland from the Apakará-tepui summit, close to the sampling site (Photo: V. Rull).

Figure 4 .
Figure 4. Stratigraphy and age-depth model of core PATAM9 A07.Dates are represented as horizontal bars idicating their respective confidence intervals.The smooth spline model (Blaauw, 2010) is repesented as a solid blue line and its confidence interval as a light brown band.The box indicates a significant acceleration in peat accumulation rates.

Figure 5 .
Figure 5. Pollen percentage diagram of core PATAM9-A07.Solid lines indicate x10 exaggeration.The surface sample is represented at the top of the diagram using horizontal bars.