Holocene fire history of Laguna Angel
embedded in the vegetation development in the
savannas of the Llanos Orientales, Colombia
Name: Jet Vessies
Student Number: 10543619
Draft: 13-07-2018
Supervisors: Henry Hooghiemstra and William Gosling period: April 2018 - July 2018
EC points: 18
Course: Bachelorproject Bèta Gamma - Earth Sciences
Institute for Ecosystem and Biodiversity Dynamics, Department of Ecosystem and Landscape dynamics, University of Amsterdam
Abstract
The presence and extent of human impact on tropical ecosystems in South America is highly debated. In the Colombian savanna of the Llanos Orientales, lake sediments have been collected and studied for pollen in 1996. Many of these sediments showed an increase of Mauritia and Mauritiella palms between 3800 - 2500 cal yr BP. To test whether this increase is human driven or a consequence of wetter climate, this research shows a charcoal analysis of the lake sediment from Laguna Angel (4˚430 ‘N, 70˚542 ‘W). Macro charcoal particles (>160 μm) are collected at 4 cm intervals over 192 cm core length and processed following standard methods. Counts of charcoal fragments [grains/cm 3 sediment], surface area [cm2charcoal/cm3sediment] and volume [cm3 charcoal/cm3 sediment] have been calculated. 8 AMS dates (between 112 - 4 cm depth) have been recalibrated with BACON. Five charcoal zones have been recognized visually. Zone 2,3 and 4 (11610 - 4140 cal yr BP; 98 - 10 cm) showed high charcoal volume. These zones correspond with drier climate found in the environmental reconstruction by pollen from Laguna Angel. Charcoal decreased (zone 5) while climate became wetter again after 4000 cal yr BP. The Cariaco Basin, 1000 km northeast of Laguna Angel, shows wet climate in charcoal zone 2,3 and 4 and drier climate in zone 5. So, climate based on the Cariaco Basin showed opposite results from climate based on pollen analysis. Difference between north and south hemisphere response to the movement of the intertropical convergence zone could be the explanation of these opposite results. To conclusively answer the question whether the increase in Mauritia is human driven or a consequence of wetter climate, several Llanos Orientales lake sediments need to be counted for charcoal.
Content
1. Introduction 3
2. Environmental setting of Laguna Angel 7
2.1 Geography and geomorphology 7
2.2 Modern Climate 8
2.3 Climate history 8
2.4 Vegetation history 12
2.5 Charcoal and fire history 12
3. Methods and Data 14
3.1 Site description 14
3.2 Processing and charcoal analysis 14
4. Results 16
4.1 Chronology 16
4.2 Charcoal stratigraphy and zonation 17
5. Fire history embedded in the climate-driven vegetation history 19
5.1 Chronology 19
5.2 Fire history and zonation 19 5.2.1 CHAR - I (15,970 - 11,610 cal yr BP; 192 - 98 cm) 19 5.2.2 CHAR - II (11,610 - 10,540 cal yr BP; 98 - 70 cm) 20 5.2.3 CHAR - III (10,540 - 6410 cal yr BP; 72 - 26 cm) 20 5.2.4 CHAR - IV (6410 - 4140 cal yr BP; 26 - 10 cm) 20 5.2.5 CHAR - V (4140 - 2040 cal yr BP; 10 - 0 cm) 20 5.3 Fire and vegetation history. 21
6. Discussion 26
6.1 Chronology 26
6.2 Fire history 26
6.3 Environmental reconstruction from pollen and charcoal 27
6.4 Hypothesis testing 28 6.4.1 Human Impact 28 6.4.2 Climate variability 29 7. Conclusion 31 8. Acknowledgements 32 9. Appendices 32 Appendix A: Data 32
Appendix B: Pollen diagram 32 Appendix C: Sample Preparation Protocol 34
1. Introduction
The first humans entered the South American continent around 15,000 years ago (Cramon-Taubadel, Strauss & Hubbe, 2017). Because little is known about these first humans dispersal, their influence on tropical ecosystems is unknown (Goldberg, Mychajliw, & Hadly, 2016). Ancient people in the Neotropics have probably putted parts of nature to their hands, but the extent and presence of their potential modifications is highly debated (Goldberg et al., 2016; McMichael et al., 2012; Levis et al., 2017; Rull & Montoya, 2014; Smith & Mayle, 2018). Within this debate, most of the paleoecological literature is focusing on Amazonian moist evergreen tropical forests and the Andes, while little is known about Amazonian savanna regions.
Savanna ecosystems are mostly covered with herbaceous species to a mixture of grasses and scattered trees (Kricher, 2011). Fire and seasonal precipitation are key for savanna regions and most of the species found are adapted to survive in dry season (Kricher, 2011). For example by developing a corkey bask as a protection for the frequent burning. However, the frequent burning influences ecosystems by reducing the amount of shrubs and trees (Andersen, Cook & Williams, 2003).
Archeological finding can be used to investigate whether neotropical people have been in a certain area (Gassón, 2002). However, there is very little archeological evidence found in some regions (Gassón 2002, Berrio & Hooghiemstra, 2007). Fossil pollen and fire history reconstructions from stratigraphic lake sediments can also tell this history and specifically the impact of indigenous people on the vegetation.
Palm species have been widely used by neotropical societies through history on a global scale, for example Mauritia and Mauritiella palms (Rull & Montoya, 2014). These Amazonian palm species were intensively used during the late Holocene by neotropical communities for food, fibre and housing material (Heckenberger & Neves, 2009; Rull & Montoya, 2014). An increase in Mauritia or Mauritiella found in pollen reconstructions can point towards past people, as Rull & Montoya (2014) have shown in the Gran Sabana tropical savanna in Venezuela. However, an increase in Mauritia or Mauritiella can also be explained by climatic changes (Behling & Hooghiemstra, 1998;1999; Berrio & Hooghiemstra, 2007).
Fire reconstructions can add evidence for human impact to changes in the vegetational reconstruction, as Rull & Montoya (2014) have shown in their research on the Gran Sabana in Venezuela (Whitlock & Larssen, 2001). Modern South American farmers use slash-and-burn agriculture in savanna regions (Iriarte et al., 2012). With this technique, savannas are burned down frequently to increase sight and have fresh green grass for cattle to graze (Rull & Montoya, 2014). Neotropical communities might have used this as well, so an increase in charcoal abundance can be
Such a fire history still needs to be discovered in the savannas of the Llanos Orientales in Colombia. Sediment cores in this area were collected in 1996 by Behling & Hooghiemstra (1998; 1999) (Appendix B). The pollen records of these cores showed a remarkable increase in the Mauritia and Mauritiella palm around 3800 yr BP (before present). This increase was interpreted as an indicator of neotropical settlement (Marchant et al., 2001). Also, little archeological information is found to support this idea to prove human influence on the palms’ increase (Gassón, 2002). In Maporita, Casanare, lithic artifacts were found and dated to 3620 14C BP ± 50 years. And around 3000 or 4000 years ago, there is a transition from hunting and plant-collecting towards the development of basic agriculture (Gassón, 2002). Charcoal as a proxy for fire history can help to give an answer to the question whether this increase in MauritiaandMauritiella in the Llanos Orientales is indeed induced by humans, or a response to wetter climate (Behling & Hooghiemstra, 1998;1999; Berrio & Hooghiemstra, 2007).
However, charcoal as a proxy for fire can be interpreted in two ways. First, as human impact increasing fire. That’s for example how McMichael et al. (2012a), interpret the presence of charcoal in moist evergreen tropical forests. High precipitation rates in these forests cause the absence of natural fire regimes, which indicates a direct relation between charcoal and human impact. Montoya & Rull (2011) showed that charcoal increased from 2000 yr BP and onwards in a tropical dry seasonal savanna in Venezuela; the Gran Sabana. So also in savanna regions, this relation between Mauritia and charcoal is found. Secondly, climate change can force fires to stop faster or happen more often. A longer wet season and therewith wetter climate increase biomass, which is a fuel for fires in dry season (Daniau et al., 2012). Drier climate can contraliraly decrease biomass, but with strong dry season, still give high values in a charcoal record. However, following Daniaeu et al. (2012) this is more unlikely.
In this research, a fire reconstruction is made of Laguna Angel (4˚430 ‘N, 70˚542 ‘W), a lake in the East of the Colombian Llanos Orientales (Figs 1 and 2). In this lake sediment record, a distinct increase in Mauritiella is found from 3800 yr BP (Behling & Hooghiemstra, 1998). With this record, the following question will be answered. How does the increase in the Mauritiella palm in the Colombian Llanos Orientales savanna between c. 3800 and 2500 cal yr BP relate with fire history based on charcoal analysis? With the following subquestions. 1) What trend is found in fire history from Laguna Angel over the last 10,000 years? 2) Does the increase in palms ( Mauritia, Mauritiella) in the Colombian Savannas between ca.3800 and 2500 cal yr BP relate with this fire history? 3) How can the achieved results be explained and supported by other evidence?
Because of the different possible interpretations of the data, the following hypothesis of this research are set up (Fig. 3).
H0. Charcoal in Laguna Angel will not have any relationship with either climate or the abundance of the Mauritiella palm.
HA1. Charcoal in Laguna Angel increases with human settlement, simultaneously with the increase of cultivated of Mauritiella.
HA2. The reconstruction of fire intensity of Laguna Angel is related to drier climate. HA3. The reconstruction of fire intensity Laguna Angel is related to wetter climate.
In this thesis, the research area will be described as a theoretical framework in chapter 2, after which the methods and data (chapter 3) are explained in detail. The results (chapter 4) will be shown, where after these will be interpreted (chapter 5) and discussed (chapter 6) and ended with a conclusion (chapter 7).
Figure 1: Map showing the geographical location Laguna Angel, and other studied lake sediments in the savannas of the Llanos Orientales after Behling & Hooghiemstra (1999).
Figure 2. Photograph of Laguna Angel. (photograph: H. Hooghiemstra)
2. Environmental setting of Laguna Angel
2.1 Geography and geomorphology
The Llanos Orientales is a savanna ecosystem covering parts of Colombia and Venezuela. It is the largest savanna in northern South America and covers around 1,243,200 km 2(Berrio & Hooghiemstra, 2007). In Colombia, the Llanos Orientales covers the Departementos Meta and Casanare and extent from the Eastern Cordillera in the west, the eastern coast of Venezuela in the north, the northern Amazon Basin in the south and the Orinoco river in the west (Berrio & Hooghiemstra, 2007; Flantua et al., 2007). Which is about 500,000 km 2 (Flantua et al., 2007). The savanna is rather flat and between ca. 180 m and 600 m altitude, but includes some areas with a higher altitude (Behling & Hooghiemstra, 1999; Flantua et al., 2007). It is cut through by several small and big rivers and streams, all eventually draining into the Orinoco River. The main rivers are the Rio Meta, Rio Vichada, Rio Guiavara and Rio Inírida. Nowadays, around these rivers and lakes, gallery forests are found (Behling & Hooghiemstra, 1998;1999). Mauritia is one of the main riparian palms found there, forming the typical densely distributed palm swamps of the Orinoco Basin (Rull & Montoya, 2014). Mauritia and Mauritiella palms are highly similar, but Mauritiella is slightly shorter than Mauritia (Rull & Montoya, 2014), The rest of the savanna and the higher altitude areas are mostly covered with poaceous vegetation, herbs and shrubs. Within this savanna ecosystem, Laguna Angel is located in the eastern part of the Colombian Llanos Orientales (4˚430 ‘N, 70˚542 ‘W) (Fig 1.). It is found between two major rivers on a 200 m elevation. Laguna Angel is fully surrounded with gallery forest, in which Mauritia palm swamps occur abundantly (Fig 2). Pollen analysis (Appendix B) and stratigraphic results (table 1) and radiocarbon dates have been published by Behling and Hooghiemstra (1998).
Depth [cm] Description
0 - 10 brown fine detritus mud, medium compact, somewhat fine sandy dark
10 - 86 brown fine detritus mud, medium compact, rootlets, somewhat fine sandy
86 - 95 transition to grey brown clayish fine detritus mud, more fine sandy
95 - 101 white very fine sand
101 - 110 light grey clay, compact
110 - 150 light grey clay, soft
126 - 132 some dark grey layers
2.2 Modern Climate
The climate in the Llanos Orientales is seasonal and is mainly related to the annual migrating intertropical convergence zone (ITCZ). The ITCZ is the zone where northeastern southeastern trade winds converge and where air rises (convection) so rainfall is produced. In its northernmost position, the ITCZ causes high precipitation (González et al., 2008). In this rainy season from April to November, the climate is warm and humid covering 90% of the total yearly precipitation (Behling & Hooghiemstra, 1998). In its southernmost position, the ITCZ causes strong northeastern trade winds and dry conditions in the Llanos Orientales. In this dry season from December to March climate is warm and dry and there is evidence for a higher occurrence of lightning strikes (Behling & Hooghiemstra, 1998; Price, 2009). The mean annual temperature (MAT) is 26.5 ˚C. There is an increase in precipitation towards the Colombian Amazon in the south (Mean Annual Precipitation: from 1200 mm northeast in the savanna to 2800 mm southwest of the Llanos orientales), and an increase in length of the dry season (2 to 5 months) towards the Venezuelan border in the North. (Flantua et al., 2007).
2.3 Climate history
Most of the literature on past climate in the Llanos Orientales is based on pollen analysis from the samples taken in 1996 by Behling and Hooghiemstra. Pollen reconstructions from several lakes in the Llanos Orientales are summarized by Berrio & Hooghiemstra (2007). General climate conditions are described as followed (Fig. 4): From ca. 10,000 - 6000 cal yr BP, climate is very dry. From ca.6000 - present climate becomes wetter. From 3800 and onwards, an increase in Mauritia and Mauritiella palms is found as an indicator of either wetter climate or human impact (read more about these vegetational changes under 2.4). The ITCZ is the main cause of seasonal precipitation and variability in the Llanos Orientales ecosystem (Hooghiemstra & Berrio, 2007). Over a longer time span, the ITCZ shows an orbitally driven cycle, i.e. the precession cycle of ca. 21,000 years (Haug et al., 2001).
The pollen from Laguna Angel are not independent climate proxies to compare charcoal abundance from the same lake with. Pollen do namely influence biomass and so the charcoal record. Independent proxies in the area around the Llanos Orientales are for example Titanium erosion from the Cariaco Basin (10°N) (Haug et al., 2001), changes in δ 18O isotope ratio from the Cueva del Tigre Perdido (6°S) (van Breukelen et al., 2008) and changes in δ 18O isotope ratio from the Santiago Cave (3°S) (Fig. 5) (Mosblech et al., 2012; Bender, 2013). These changes found in Titanium erosion or δ 18O isotope ratio agree with the movement of the ITCZ (Haug et al., 2001). Important to notice is the antiphase precipitation abundance between the northern and southern hemisphere causing the negative correlating independent proxies (Fig. 5) (Cruz et al, 2009). From ca.10,000 - 6000 cal yr BP
the ITCZ is moving north, causing more precipitation in the northern hemisphere (Haug et al., 2001; Cruz et al., 2009), where after it moves southwards again (ca.6000 - 250 cal yr BP). This southern directed movement is considered to be causing drier climate on northern hemisphere South America (Fig. 6) (Wanner et al., 2008).
The Cariaco Basin (offshore Venezuela) is an anoxic marine basin based on changes in titanium as an independent proxy for precipitation from the Holocene, giving a northern hemisphere signal. The main driver of the precipitation changes in this basin is the ITCZ and its seasonal and orbitally driven changes (Peterson & Haug, 2006; González et al., 2008). A decrease in titanium is interpreted as drier climate and more southern position of the ITCZ, because terrestrial erosion brings titanium towards rivers and eventually the ocean and when precipitation and river discharge increase, so does the erosion of titanium and the amount of it in the Cariaco Basin (González et al., 2008). So, an increase in titanium is interpreted as wetter climate, strong northeastern trade winds and a shift towards a more northern position of the ITCZ (González et al., 2008). The Cariaco Basin shows low precipitation during the end of the last ice age with very low values during the Younger Dryas (13000 - 11500 cal yr BP) (Haug et al., 2001). An increase in precipitation is seen in the beginning of the Holocene; the ‘Holocene Climatic Optimum’, between 10,000 and 5500 cal yr BP (Haug et al., 2001). From 5500 to present, climate becomes drier again, with a period with high fluctuation between 4000 and 2000 cal yr BP (Haug et al., 2001). Interesting is that precipitation based on the Cariaco record and precipitation inferred from pollen analysis in the Llanos Orientales are showing a negative relation (Figs. 5 and 6).
The Santiago Cave (south Ecuador) and Cueva del Tigre Perdido (Northeast Peru) are speleothem records based on changes in δ 18O isotopes (van Breukelen et al., 2008, Mosblech et al., 2012). A standard ratio between 16O and 18O isotopes is compared with the found ratio in the speleothem stalactites and stalagmites (Bender, 2013). Whenever this ratio is higher than standard, more 18O is found. This indicates more heavier isotopes in the precipitation, interpreted as drier climate and vise versa (Bender, 2013). These two speleothem records are located on the southern hemisphere and show a southern hemisphere precipitation system,which is the opposite of the Cariaco Basin precipitation.
Figure 4: Synthesis of environmental change from savanna lakes located on a transect from east (left) to west (right) shown on a linear time scale. SAR; Laguna Sardinas, ANG; Laguna El Angel, AP; Laguna El Piñal, CAR; Laguna Carimagua, CV; Laguna Chenevo, LL; Laguna Loma Linda, MAR; Laguna Margaritas (from Berrio & Hooghiemstra, 2007).
Figure 5. Precipitation reconstruction based on speleothems in South America with δ 18O or Titanium as a proxy (Haug et al., 2001; van Breukelen et al., 2008; Mosblech et al., 2012; Bernal et al., 2016). Right in the graphs represents wet climate, Left in the graph represents dry climate.
Figure 6. Southward migration of the ITCZ during the Holocene and induced global climate changes between 6000 and 250 cal yr BP. After Wanner et al., 2008.
2.4 Vegetation history
From the pollen analysis (Appendix B) by Behling & Hooghiemstra (1998; 1999), the following past climate can be inferred. Between 10,680 - 10,070 cal yr BP (zone ANG-1), there was found an increase in Mauritia as a result of higher moisture availability. This moisture availability was likely to be a result of colder Lateglacial conditions and related decrease in evaporation, rather than increased precipitation. During 9730 - 5620 cal yr BP (zone ANG-2a and zone ANG-2b), there is little presence ofMauritia.This absence of Mauritia and other pollen reflect the driest period of the record, caused by low rainfall rates and an extended annual dry season. Cyperaceae is highly available, which could be the aquatic type living on shallow depths, or the grassy taxa, found in dry open areas (Edwards et al., 2003; González et al., 2008). High availability of Cyperaceae is an indicator of dry climate (Edwards et al., 2003). Between 5260 - 3890 cal yr BP (zone ANG-3) does Mauritiellaoccur for the first time around Laguna Angel. There is a vegetational change towards wetter climate conditions and a higher soil moisture content (Ross et al., 1992). This vegetational change is clearly distinguishable in Laguna Angel. Following Behling and Hooghiemstra (1998), the pollen record between 3890 - 2000 cal yr BP (zone ANG-4) is incomplete or partially missing, probably due to high level of decomposition of the organic deposits. Though, there is a marked increase of Mauritiella, which could be a result of increased human impact. Hopefully, which can be given more evidence for by the charcoal analysis that’s going to be done in this research. Around 2500 cal yr BP there is a decline in Mauritia, which could have been caused by an ENSO intensification (Haug et al., 2001).
2.5 Charcoal and fire history
Charcoal has been widely used as a proxy for the fire history and human impact in tropical rainforest regions. However, charcoal hasn’t been used much in savanna ecosystems as a proxy for human impact. Nonetheless, Montoya and Rull (2011) used it in their research in the Gran Sabana savanna and concluded a positive relation between the presence of Mauritia and Mauritiella and charcoal abundance. Tropical savannas experience mild burns and major fires every few years (Kricher, 2011). These natural fire regimes as a result of frequent lightning strikes can complicate distinguishing a relationship between people and fire (Kricher, 2011; Armenteras, Romero, & Galindo, 2005). Lightning strikes can cause fires to start. During drier climate, there is a higher occurrence of lightning strikes than during wetter climate (Price, 2009). When fires are suppressed, competition between plant taxa will actively change the composition of the savannas (Kricher, 2011). Fires nowadays are approximately by 20% caused naturally, while 80% is of human origin (Armenteras et al., 2005). Besides these natural fires, fire abundance can increase during wetter climate, corresponding with the first alternative hypothesis (Fig. 3). During wetter climate, more
biomass is produced, which can easily catch fire (Rull & Montoya, 2014). It might be hard to distinguish between the effects of wetter climate and potential anthropogenic effects. So, both the increase in Mauritia suggesting the presence of pre-Columbian societies and the seemable more independent proxy charcoal are debatable. With this knowledge, it might be hard to draw an unbiased and one-sided conclusion.
3. Methods and Data
3.1 Site description
Laguna Angel (4°4305’N, -70°5416’W) is located in the northeastern part of the Colombian Llanos Orientales. It is found isolated from other hydrological features and between two major rivers (at a 5 to10 km distance) at 200 m elevation. It is fully surrounded with gallery forest, where Mauritia palms are abundant (Fig. 2) (Behling & Hooghiemstra, 1998). The lake is isolated, ca. 300m in diameter and had a water depth of 150cm during coring in the dry season, which was the end of February 1996 (Behling & Hooghiemstra, 1998). The sediment is cored in the middle of the lake with a Livingstone piston sampler , has a 193 cm length and is 5cm in diameter (Behling & Hooghiemstra, 1998). The sediment core was kept in a cold dark room (4℃). Pollen analysis (Appendix B), stratigraphic results (table 1) and radiocarbon dates are published by Behling and Hooghiemstra (1998).
Laguna Angel was preferred above the other lakes cored in the Llanos Orientales because of the location in the middle of the Llanos, the size (length) of the core and the clear change in vegetation and specifically the palm Mauritiella (Berrio & Hooghiemstra, 2007).
3.2 Processing and charcoal analysis
Fortyseven samples were extracted on a 4 cm interval from 0 - 192 cm, with extra samples on 2, 4.5, 5, 5.5, 6, 6.5, 7, 9 and 10 cm core depth with a 1.3 cm 3 volume for charcoal analysis. This 4 cm interval is the same interval as used for Pollen Analysis in 1996 and includes approximately 400 years of sediment accumulation per sample (Behling & Hooghiemstra, 1998). For pollen analysis was only cored up til 96 cm (in 1998) (Behling & Hooghiemstra, 1998). There is chosen for a deeper charcoal record because of potential interesting findings in the second meter of the core. Standard laboratory techniques were used to identify charcoal particles, see full Sample Preparation Protocol in appendix A (Whitlock & Larsen, 2001). Weak hydrogen peroxide (3% H 2O2) was used to bleach organic material in the sediment samples to make charcoal recognition under the microscope less complicated (Appendix C) (Rhodes, 1998). By sieving, a macro charcoal fraction was quantified for particles >160 μm, which is used as a proxy for local fires (Whitlock & Larssen, 2002).
Aeolian transport of charcoal pieces causes small charcoal particles can origine from further away (regional or extralocal fires), mostly downwind of the site (Whitlock & Larsen, 2001; Patterson, Edwards & Maguire, 1987). Another possibility is that the small fraction originates from burned herbs. High amounts of herbs and little shrubs and trees in the savanna make this likely (Fig. 7). The bigger fraction (>160 μm) represents more local fires at the study site and the burning of woody material instead of herbs (Clark & Patterson, 1997; Gosling et al., 2017; Whitlock & Larsen, 2001;
Patterson et al., 1987). Also, the amount of charcoal found in a subsample is dependent of the size of the lakes, because bigger lakes catch more charcoal from the area (Whitlock & Millspaugh, 1996).
The >160 μm fraction is evaluated under an Olympus stereo microscope with 7 - 40x magnification. Following Whitlock & Larsen (2001), charcoal particles are opaque, angular and mostly planar black fragments, but are easily confused with other black particles in sediments. It is breakable and will fracture rather than impress or impale under slight pressure (Whitlock & Larsen, 2002).
Particles [amount of particles/ 1.3 cm 3] were counted and a more standardized unit [amount of particles/cm cm3] was calculated. Surface area [mm2/cm3] is calculated with use of ImageJ (Rasband, 2011). From the surface area, cubic cm (cm3) is calculated following Weng (2005) to decrease the influence of tiny particles with the following formula: CA 3/2. C has to be calculated from known charcoal volume in the sample, which is unknown for Laguna Angel. So C is taken as 1, which is substantiated by literature (Weng, 2005).
Behling & Hooghiemstra (1998) selected 8 core bulk samples for accelerator mass spectrometry (AMS) for 14C dating at the University of Utrecht (Van der Borg et al., 1987). These AMS dates are recalibrated and interpolated with Bayesian ACcumulatiON (BACON) in R, using intcal13 as 14C calibration curve (Blaauw & Cristen, 2011;2013; Marsh et. al., 2018). Because Behling & Hooghiemstra assume that Laguna Angel does not include the ages 2000 - 0, a surface sample of 2000±1000 has been added for calibration. All calibrated ages have been rounded to multiples of 10.
For interpretation, visualisations are made in C2 (Juggins, 2007). For the comparison with pollen, data from Behling & Hooghiemstra (1998) is used with their choice of CONISS and pollen sum. This pollen data was available in the Latin American Pollen Database (LAPD) and downloaded from the Neotoma Paleoecology Database. Because of time limitations, no statistical analysis is performed and only descriptive conclusions are drawn based on visual changes. A link to the raw data and metadata are added to Appendix A.
4. Results
4.1 Chronology
AMS dates derived from Behling and Hooghiemstra (1999) are recalibrated and interpolated with BACON (Blaauw & Cristen, 2011;2013) (Table 2; Fig 8). The accumulated rate, memory and calibrated ages from BACON recalibration are shown in this figure. The accumulation rate is right skewed and the mean sediment rate is 1 cm every 100 years. The deepest part of the core (112 - 192 cm) is extrapolated because no big enough carbon particles were available. The deepest AMS dated particle is found on 112 cm (12,380 cal yr BP). Between 12,380 and 5620 cal yr BP, sedimentation rate is fast. From 5610 cal yr BP (22 cm) to 9490 cal yr BP (40 cm) the sediment accumulation rate decreases. From 2560 (4 cm) cal yr BP until 5610 cal yr BP (22 cm), sedimentation rate increases (the graph flattens). The top part of the core between 2100 cal yr BP and 2560 (0 - 4 cm) shows a slow sediment accumulation rate.
Laboratory number
Depth (cm)
14C yr BP δ13C Calibrated age weighted mean [cal
BP] following BACON (Blaauw & Cristen, 2011;2013) Surface 0 2000±1000 2097.7 UtC - 5472 4 2.451±30 -26.2 2557.5 UtC - 4950 9.5 3,651±43 -25.0 3985.3 UtC - 4951 10 4,830±60 25.1 4129.8 UtC - 5473 22 4.864±38 -26.0 5606.9 UtC - 4952 40 8.450±38 -25.1 9486.6 UtC - 4953 70 9.320±70 -24.3 10612.7 UtC - 4954 95.5 10.070±60 -24.7 11609.8 UtC - 4955 112 12.880±80 -22.1 12377.3
Table 2; List of AMS radiocarbon dates and sample specific data from core Laguna Angel. We argue that the date at 10 cm core depth is erroneous and therefore rejected to develop the age model (after Behling & Hooghiemstra, 1998).
Figure 8: Age model for core Laguna Angel based on calibrated radiocarbon ages with BACON (Blaauw & Cristen, 2011; 2013)
4.2 Charcoal stratigraphy and zonation
The amount of charcoal fragments [grains/cm3 sediment], surface area [cm2 charcoal/cm3 sediment] and volume [cm3 charcoal/cm3 sediment] have been plotted both on a linear depth and linear time scale to prevent that interpretations are made from changes in sedimentation rate rather than the calibrated age. In the plot over depth, there are evenly distributed samples, with extra samples in the upper 10 cm. In Fig. 9, charcoal amount, surface area and volume are plotted. There are many samples between 9500 - 16,000 cal yr BP, fewer between 9500 - 4400 cal yr BP, a small increase between 4400 - 3000 BP, after which the sample density decreases rapidly.
The results have been divided into 5 different zones based on charcoal volume [cm 3charcoal/cm3 sediment], with a threshold value of 50 cm 3/cm3. CHAR - I (15,970 to 11,610 cal yr BP) covers the lower 92 cm with 24 samples (98 - 192 cm), this zone is characterized by the absence of charcoal. CHAR - II (11,610 to 10,540 cal yr BP) is covering the depth of 70 to 98 cm with 7 samples. This zone is characterized by medium charcoal volume between 13 and 43 cm 3/cm3. CHAR - III (10540 - 4140 cal yr BP) is found between 26 and 70 cm. It is defined by high charcoal volume > 50 cm 3/cm3 except one sample on 52 cm with 44 cm 3/cm3(9940 cal yr BP). CHAR - IV (6410 to 4140 cal yr BP) is covering 6 samples between 10 and 26 cm. The zone is characterized by decreasing amounts of charcoal. The upper 10 cm of the core reflects CHAR - V (4140 to 2040 cal yr BP) with 10 samples.
Figure 9. Distribution of charcoal throughout the 190 cm long sediment core. The upper figure shows charcoal data plotted on a linear depth scale with time as secondary axis. The lower figure shows charcoal plotted on a linear time scale with depth as a secondary axis.
5. Fire history embedded in the
climate-driven vegetation history
5.1 Chronology
The new age model made with BACON seems to correspond with climate conditions found in the Holocene (Blaauw & Cristen, 2011; 2013). Sedimentation rate is related to the dry and wet climate conditions that Berrio & Hooghiemstra (2007) found in the comparison they made within lake sediment pollen (including Laguna Angel). A higher amount of bare soil and less deeply rooting vegetation during the dry beginning of the Holocene causing more run off during wet season and more aeolian erosion during dry season could have caused this higher sedimentation rate (Berrio & Hooghiemstra, 2007). Also, dry climate and high wind speeds in the end of the Pleistocene caused deposition of sandy sediment and dune formation (Carr et al., 2016). Sedimentation rate decreased as a result of climate that became wetter after this dry period (Berrio & Hooghiemstra, 2007). Because the chronology from BACON and climate conditions correspond, we can confirm the validity of the chronology. More critical notes on the methods will be given in section 6.2.
5.2 Fire history and zonation
5.2.1 CHAR - I (15,970 - 11,610 cal yr BP; 192 - 98 cm)
This zone is characterized by the absence of charcoal. During the end of the Pleistocene, climate was drier in the northern part of South America (Haug et al., 2001; Behling & Hooghiemstra, 1998; 1999, Berrio & Hooghiemstra, 2007; Peterson & Haug, 2006). According to climate based on the Cariaco Basin, the ITCZ was located more southwards (more on this in section 2.3), which could have lead to this drier period with a longer dry season and increased wind speed in the Llanos Orientales (Fig. 6) (Carr et al., 2016, Haug et al., 2001). This dry period caused less riparian vegetation surrounding lakes in the Llanos Orientales (Behling & Hooghiemstra, 1998;1999; Berrio & Hooghiemstra, 2007). Less riparian vegetation and higher wind speed caused that sedimentation rate increased and grain size got bigger, (Fig. 8; Tables 1 and 2). This might also have led to the formation of dunes by aeolian transport (Carr et al., 2016).
With this information, there are three explanations for the absence of charcoal. Firstly, it could indicate that drier climate caused the availability of less biomass to be burned naturally (Daniau, 2012). So no charcoal is found. Secondly, when wind speed is high, charcoal particles are transferred
in the macro charcoal fraction (> 160 μm). Thirdly, it could be related to extrapolation of depth 192 to 112 cm because the deepest AMS - dated particle is on 112 cm. It can be that the sediment between 192 and 112 cm is deposited very fast; causing the extrapolated time dates to be wrong (see section 6.1).
5.2.2 CHAR - II (11,610 - 10,540 cal yr BP; 98 - 70 cm)
This zone represents the early Holocene with an increasing amount of charcoal. The early Holocene is characterized by increasing precipitation after the last ice age and the ITCZ being in its northernmost position (Fig. 6) (Wanner et al., 2008; Haug et al., 2001). Sedimentation rate decreased a little (Fig. 8). Stratigraphic results show a change from sandy material towards grey brown and clayish mud, which could be a result of changes in erosion. More muddy sediment can be a sign of lower energy levels in the sedimentary environments, which could indicate a bigger lake. Most likely, charcoal has increased because of the higher availability of biomass and the decrease of wind speed leading towards less fractionation of big charcoal particles (Daniau et al., 2012).
5.2.3 CHAR - III (10,540 - 6410 cal yr BP; 72 - 26 cm)
Within CHAR - III, charcoal volumes are the highest and covers most of the Holocene Climatic Optimum (Haug et al., 2001). In this period, temperatures and precipitation were high as a result of a more northern location of the ITCZ (Haug et al., 2001; Wanner et al., 2008). High precipitation induced high biomass production which can have caused high amounts of charcoal. But also the low wind speeds and less fractionation in this period, leading towards bigger charcoal particles (at least over 160 μm).
5.2.4 CHAR - IV (6410 - 4140 cal yr BP; 26 - 10 cm)
CHAR - IV is characterized by decreasing charcoal volume. The ITCZ is moving south with changes in solar insolation (Fig. 6) (Wanner et al., 2008). This transition to the south is causing a drier climate in northern hemisphere South America. Decreasing charcoal volume can be related with drier climate because less precipitation causes less biomass production and less vegetation to be burned (Daniau et al., 2012).
5.2.5 CHAR - V (4140 - 2040 cal yr BP; 10 - 0 cm)
In this zone, close to zero charcoal volume is visible. The southern position of the ITCZ that causes drier climate in northern hemisphere South America, could have led to the low charcoal volume in this zone (Wanner et al., 2008). In this zone, also the increase in Mauritiella is found. Mauritiella and charcoal are opposite: when charcoal had indeed increased, this would be a sign of human impact or wetter climate in the area. Sedimentation rate is very slow (Fig. 8) and sediment type
changed towards more fine sandy brown fine detritus mud (Table 1). These could be related to changes towards wetter climate.
5.3 Fire and vegetation history.
In the previous section, the fire history has only been compared with the independent proxy for climate from the Cariaco Basin and the most accepted movements of the ITCZ with related northern hemisphere South America climate conditions (Haug et al., 2001; Wanner et al., 2008). Cariaco seems the best proxy to compare Laguna Angel with because it is the closest independent proxy on the northern hemisphere on a ca.1000 km distance. Besides that, it is partially influenced by erosion from the Orinoco River, the river where all the precipitation from the Llanos Orientales ends up. Charcoal and precipitation based on the Cariaco Basin look similar with wetter climate inducing more biomass and therefore an increase in charcoal. Even though charcoal and the Cariaco Basin look similar, precipitation inferred from pollen analysis in the Llanos Orientales and Laguna Angel are giving opposite results from precipitation based on the Cariaco Basin (see section 2.3). It shows more relation with the Cueva del Tigre Perdido and Santiago Cave speleothems. However, these are located on the southern hemisphere and should be showing an antiphase signal from the northern hemisphere Laguna Angel (Cruz et al., 2009). This interesting difference strongly changes the interpretations made in section 5.2. These will be further discussed in this section, where after in chapter 6.5, the opposite climate between Laguna Angel and the Cariaco Basin will be discussed.
So in this section, fire abundance will also be compared with pollen data published by Behling & Hooghiemstra (1998) (Figs. 10 and 11). This paper describes 4 representative zones derived from CONISS (ANG - I to ANG - IV). These ANG - zones can be compared with the charcoal zonation (CHAR - I - IV). For this comparison, new calibrated BACON ages have been used, instead of the radiocarbon ages by Behling & Hooghiemstra (1998).
CHAR - I (15,970 - 11,610 cal yr BP; 192 - 98 cm) is not included in the pollen record, because no pollen samples were taken on these depths (Behling & Hooghiemstra, 1998). In CHAR - I, no charcoal is found. This absence of pollen and charcoal could be a result of extrapolation between 192 and 112 cm, because the deepest AMS-dated particle is on 112 cm. It can be that the sediment between 192 to 112 cm is deposited very fast; causing the extrapolated time dates to be wrong (further explanation in section 6.1).
CHAR - II (11,610 - 10,540 cal yr BP; 98 - 70 cm) does partially correspond with ANG - I (11,770 - 11,290 cal yr BP; 94 - 84 cm). ANG - I is characterized by low pollen concentration, which could be
an indicator of dry climate. Around 90 cm (11,590 cal yr BP), Poaceae is decreasing while charcoal increases. These might be related, as drier climate and increase lightning can have burned these grasses.Isoetes, an aquatic fern preferring shallow water, and Cyperaceae are also high in this period. Cyperaceae can either live in shallow water depths.Decreased water levels tend to stimulate biomass production of the taxa over time (Edwards et al., 2003). Cyperaceae is also a grassy type, found in dry open areas, such as the savanna (González et al., 2008). The increase in these two types of Cyperaceae and the increase in Isoetes can be related to this drier weather. ANG - I seems to be mostly related with dry climate as found in the end of the Pleistocene found in the Cariaco Basin (Haug et al., 2001). In this dry climate with little vegetation, little biomass was available to get burned, as found in the charcoal record.
CHAR - III (10,540 - 6410 cal yr BP; 72 - 26 cm) overlaps half of ANG - IIa (11,290 - 9940 cal yr BP; 84 - 52 cm) , and all of ANG - IIb (9940 - 5920 cal yr BP; 24 - 52 cm). If the beginning of CHAR - III would have been on 84 cm (as discussed under 6.2) , CHAR - III and ANG - II would have matched more closely. In CHAR - III, charcoal volume is above 50 cm 3/cm3, with the highest amounts found in the record. Over all, ANG - II is characterized by taxa that reflect dry climate (Berrio & Hooghiemstra, 2007; Behling & Hooghiemstra, 1998). Cyperaceae is relatively high, corresponding with dry climate (Edwards et al., 2003). However, there is no increase in Isoetes as in ANG - I, which does not reflect the idea of dry climate. In ANG - IIa, pollen concentration and species richness are high, related to wetter climate. Typha is present during this ANG - IIa period, but just with some few pollen grains. This dry climate inferred from pollen is the opposite from wetter climate found in the Cariaco Basin proxy (Fig. 12) (Haug et al., 2001).
If we assume that high amounts of charcoal is a result of high biomass availability, charcoal shows wet climate (Daniau et al., 2012). Because the pollen show dry climate, charcoal and pollen do not relate under that assumption. However, it could be that savanna ecosystems always show a high enough biomass availability to get burned. And with the increase of lightning during drier periods, charcoal abundance would increase naturally with drier climate (Price, 2008).
CHAR - IV (6410 - 4140 cal yr BP; 26 - 10 cm) covers ANG - III (5920 - 4370 cal yr BP; 24 - 12 cm) Charcoal in ANG - III still shows some charcoal volume. Precipitation based on pollen shows that ANG - III is the transition zone from dry towards wetter climate. Within the forest and gallery forest shrubs and trees, the first Mauritiella occurs, Araceae increases and Myrsine appears. Within the savanna shrubs and trees, Byronima and Didymopanaxincrease. The savanna herbs change as well. Cyperaceae is still high, but decreases rapidly, Lamiaceae increases. None of the aquatics show significant changes and show a relatively uniform representation. The ferns show increase in big
monolete psilate (> 50 μm) and Selaginella. Both of these do only occur with some spore grains. Over all, pollen show a wetter climate. This does not correspond with the Cariaco Basin, where precipitation decreases and stabilizes around 2000 BP (Haug et al., 2001).
CHAR - V (4140 - 2040 cal yr BP; 10 - 0 cm) is covering nearly the same area as ANG - IV (4370 - 2040 cal yr BP; 12 - 0 cm). This zone shows a charcoal volume between 1 and 0 cm 3/cm3. In ANG - IV big changes in vegetation composition are happening. Mauritiella is really high, which can either be climate or human impact. Curatella is increasing, but only up to 1%. Poaceae suddenly decreases and slowly increases again.Cecropia, an indicator of disturbance, increases, related to wetter climate conditions, increased fire abundance or human impact.
Overall, pollen and charcoal show similar precipitation. When precipitation is low based on environmental reconstruction from pollen, charcoal increases. When precipitation is high, charcoal decreases. As explained previously, charcoal does not need to follow higher precipitation with increased biomass abundance, even though Daniau et al. (2012) indicate that from Gran Sabana savanna fire systems. The decrease in charcoal can therefore be explained as a result of wetter climate and possibly decreased lightning. However, the decrease in charcoal can actually be a result of human impact, as will be explained in section 6.4.
Figure 10. Pollen and charcoal in core Laguna Angel plotted on a linear depth scale. Green marked taxa represent the most abundant pollen. Background grey and white represent the charcoal zones (CHAR - II, CHAR - III and CHAR - IV). Red crosses represent the evaluated charcoal samples. Samples from 96-193 cm core depth are not shown as samples do not contain charcoal nor pollen (Behling & Hooghiemstra, 1998).
Figure 11. The main fossil pollen taxa in the pollen spectra compared with charcoal volume, pollen concentration and species richness. Background grey and white represent the charcoal zones (CHAR - II, CHAR - III and CHAR - IV). Red crosses represent the evaluated charcoal samples. From 96-193 cm depth, samples are omitted because charcoal nor pollen were found and made visualisation less
6. Discussion
6.1 Chronology
The BACON recalibrated chronology (as explained under 5.1) is related to Holocene climate conditions. This seems to fit nicely, but there are some inaccuracies which should be taken under consideration. Firstly, the oldest AMS dated sample at 112 cm depth, while the core is 193 cm deep. Between 193 and 121 cm, the ages are extrapolated. So, no chronological conclusions can be drawn from this part of the core. To do so, some samples from the lower part of the core should be dated with AMS, but the lower part of the core does not contain datable organic matter (Behling & Hooghiemstra, 1998). Secondly, the sedimentation rate seems to be very slow in the upper part of the core (from 10 cm and onwards). At 4 cm the age is 2560 cal yr BP, which would also be a very slow sedimentation rate. That is why Behling & Hooghiemstra (1998) assumed that the upper 2000 years may be missing. Thirdly, the sample on 22cm might not have the right AMS date (Behling & Hooghiemstra, 1998), but the date from 10 cm has been excluded with BACON recalibration because of the high 1180 years difference within 1 cm. This can also be explained by a possible hiatus in the record. The sedimentation accumulation rate does not need to be continuous.
6.2 Fire history
Without comparing charcoal with pollen, charcoal abundance is a result of wetter and drier climate conditions from the Cariaco Basin. This is in line with the theory that wetter climate increases biomass and so fuel to burn for charcoal (Daniau et al., 2012). The discussion of the comparison with pollen will be made in section 6.3. There are some points that need to be discussed that can have changed the outcome of the charcoal data. First, the zonation of the charcoal is based on visually determined changes with a threshold of 50 cm 3/cm3 in the charcoal volume record. This zonation shows the biggest changes, but could also have been taken on 40 cm 3/cm3. This threshold would have changed CHAR - III from 26 - 7 cm to 26 - 84 cm. and isn’t chosen because of the relatively low charcoal volumes on 72, 76 and 80 cm. Second, zonation is based on charcoal volume, because that is the most commonly used type of charcoal unit, based on Weng (2005). If charcoal amount was used, zone CHAR - II would not have been recognized. Third, in the formula calculating charcoal volume from charcoal surface, the constant C has been taken 1 because it is unknown (Weng, 2005). This might have slightly altered the charcoal volume data. Surface area [cm2] is calculated with imageJ software. Fourth, because of the high amount of small charcoal pieces, it cannot be overseen that some pieces might have been lying on top of each other during surface area calculation with imageJ, so surface area reflects a minimum.Fifth, time limitations caused that no statistical analysis is performed
and only descriptive conclusions are drawn based on visual changes. Despite complications of the chosen methods the achieved results are still representing a local signal from charcoal from Laguna Angel in the Llanos Orientales. It can be a good start for more research within the reconstruction of fire histories. And future research should then take into account and add the following points.
Besides Laguna Angel, there are other charcoal counts available from the Llanos Orientales lake sediment cores, but these are not published yet (Berrio, personal communication, June 2018). Adding this data to the charcoal data from Laguna Angel, would give a more complete fire history of the Llanos Orientales. In addition, micro charcoal (75 - 160 μm) has not been counted yet. Micro charcoal reflects a more regional fire history. It can be the result of big pieces of charcoal getting fractured by aeolian transport and charcoal from poaceous vegetation could be included more in this fraction than woody vegetation. However, this fraction can also catch too much charcoal from further away, which could give chaotic results. Also, many papers show similarities in micro charcoal and macro charcoal (Gosling et al., 2017; Montoya & Rull, 2011).
Lastly, type of vegetation, woody or grassy, can have influence on the temperature on which it gets burned. In future research, FTIR (Fourier-transform infrared spectroscopy) might help find the temperature of burned particles and so tell us more about the nature of charcoal pieces.
6.3 Environmental reconstruction from pollen and charcoal
Climate inferred from pollen and climate inferred from the Cariaco Basin show a negative relation (Fig. 12). There are some interesting thoughts about the environmental reconstruction that should be taken into consideration.
First, many of the fossil pollen taxa are just build up from single grains (see appendix B) and make out a very small part of the pollen sum, which with statistical analysis probably would not be enough to be mentioned. Secondly, no statistical tests are done with the data. It could be that some of the described taxa under section 5.3 are not significantly present. Thirdly, some of the zones described in Behling & Hooghiemstra only include 4 samples. To draw more thorough conclusions from these samples, more samples should be taken in these areas. Fourthly, only counting pollen can give an incomplete representation of the environmental history. Therefore, including diatoms and phytoliths in future research would be useful. Fifthly, it is hard to add several lake sediments to the data, because most of the lowland Colombian sedimentary basins are included in the riverine system (Marchant et al., 2001). Lastly, counting phytoliths in several lake sediments from the Llanos Orientales might show new insight into the environmental history and therewith the climatic history or show evidence for human impact.
6.4 Hypothesis testing
In the introduction, four hypotheses were stated to answer the research question (Fig. 3). Comparing the results with these hypotheses, the second hypothesis would be the best fitting one: charcoal based fire abundance coincides with the pollen-based driest intervals. To underpin this hypothesis, the impact of human impact and climate variability on the results of Laguna Angel are discussed in this section.
6.4.1 Human Impact
The first alternative hypothesis suggests an increase in charcoal occurring simultaneously with the increase of Mauritia around 3800 cal yr BP. The results from Laguna Angel do not look similar as this hypothesis. However, there is a paper that shows similar results as in Laguna Angel; a decreasing amount of charcoal with expected arrival of humans, and attributes these results to human impact. Iriarte et al. (2012) describe that pre-Columbian people in French Guiana limited fire for agriculture. This limitation of fire is debating the common view that anthropogenic fires have been a key feature in neotropical savanna ecosystems (Iriarte et al., 2012). Whether fire limitation could have happened in the Llanos Orientales is unclear. French Guiana farmers used ‘raised-field’ agriculture, which is only suitable in seasonally flooded areas. Lake levels increase during wet season in the Llanos Orientales, but it is unclear whether this would be enough to practice raised-field agriculture.
Other areas in northern South America do have charcoal records and other proxies for human impact, such as archeological records. McMichael et. al (2012b) used charcoal, phytolith and geochemical data from 247 soil cores at 55 locations in Western Amazonia, south of the Llanos Orientales to find evidence for human impact on the ecosystem. Even though other papers suggested high neotropical human impact on ecosystem functioning, they only found signs of small settlement without large scale forest clearing or agriculture. Slash-and-burn was less thorough than found in modern communities. Goldberg et al. (2016) modelled radiocarbon ( 14C) dates from archaeological records in South America and found an exponential growth of the human population from 5.5 ka to 2 ka with Kernel distribution. Marchant et al. (2001) found that human impact is likely to be the origin of anthropogenic vegetation disturbance around 3000 cal yr BP in Colombia (including sites from the Llanos Orientales). However, they conclude that human settlement is highly localized.
Charcoal is found to be a good indicator of past people in the Venezuelan Gran Sabana (Montoya & Rull, 2014). There, presence of Mauritiella palms and charcoal look similar over time. Because Montaya & Rull (2011) show a positive relation between Mauritiellaand charcoal, and Laguna Angel shows a negative relation, it seems unlikely to draw the conclusion that human impact has influenced Laguna Angel. However, Montaya & Rull (2011) show little to no charcoal when people were not
expected, in contrast to Laguna Angel. Also, their environmental reconstruction corresponds better with the Cariaco Basin.
Additionally, there is little evidence of archeological artifacts in the area, mainly because it is under explored by archeologists. So human settlement in the Llanos Orientales cannot be proved yet.
6.4.2 Climate variability
From the charcoal results, human impact seems unlikely. But the results show nice overlap with hypothesis 2. And so, the changes in Mauritiellamight be climate driven too. Laguna Angel is located 1000 km southwest of the Cariaco Basin (10°N). This anoxic marine Basin based on changes in titanium is an independent proxy for precipitation from the Holocene, giving a northern hemisphere signal, as explained in section 2.3. The Cueva del Tigre Perdido speleothem (6°S), based on δ 18O shows the southern hemisphere Holocene record. The Santiago Cave (3°S) shows the Holocene - Pleistocene southern hemisphere record based on δ18O. In figure 12, pollen analysis DCA from Laguna Angel and charcoal volume are compared with the Cariaco Basin precipitation and two speleothem records, based on δ18O.
Laguna Angel is located 4°N. Because moisture from the ITCZ is thought to arrive from the northern hemisphere, it would be expected that the pollen DCA dry and wet climate look similar as the Cariaco Basin precipitation. However, the graphs show exactly opposite results: where the Cariaco Basin suggests dry (wet) conditions, pollen inferred climate shows wetter (drier) climate conditions around Laguna Angel. This would indicate that pollen from Laguna Angel show a southern hemisphere signal, because pollen DCA matches with Cueva del Tigre Perdido and the Santiago Cave. Because the ITCZ brings moisture from the north (Haug et al., 2001), this is an unexpected signal. This difference leaves a peculiar difference between pollen, charcoal and independent precipitation data from Cariaco. The possible southern hemisphere signal at Laguna Angel could be explained by papers that describe a difference between eastern and western climate in South America (Cruz et al., 2009; Cheng et al., 2013). This could be the result of moisture arriving from the south instead of from the northern hemisphere and ITCZ. Cheng et al. (2013) describe high fluctuations in the climate of western Amazonia compared with the east, which might have resulted in the difference between Cariaco and Laguna Angel. Beside that, there is also still the antiphase between northern and southern hemisphere climate (Cruz et al., 2009).
Over all is the following explanation the most plausible. In Laguna Angel, charcoal increased during periods in which pollen evidence based dry climatic conditions. This implicates there is no relationship between availability of fuel (optimally present under wet climate conditions) and fire intensity. So charcoal is related to dry climate found in the pollen record with an antiphase response towards climate in the northern hemisphere inferred from the Cariaco Basin.
Figure 12. Charcoal and pollen DCA 1 (explaining 9.4% of the variance) compared with precipitation data from the Cariaco Basin (offshore Venezuela), Cueva del Tigre Perdido (Northeast Peru) and the Santiago Cave (south Ecuador). The charcoal zones are shown in the background in tones of grey. Left hand three columns: right in the graphs represents wet climate, Left in the graph represents dry climate.
7. Conclusion
To conclude, the charcoal history of Laguna Angel is giving unexpected and interesting results. Absence of charcoal during the end of the Pleistocene, increasing and high amounts of charcoal during the early and middle Holocene, and decreasing amounts of charcoal during the late Holocene. These Holocene climatic conditions fit best with hypothesis two; high amounts of charcoal are related to dry periods in the pollen record, and no positive relation is seen between increasing Mauritiellaand charcoal. This non existing relation could be the effect of different land use by neotropical farmers, such as raised-field agriculture, but this seems unlikely. Even though increasing charcoal is often happening simultaneously with increasing precipitation and biomass, hypothesis two seems the most plausible.
Beside that, climate reconstructed from pollen records gives an opposite signal from the independent precipitation proxy from Cariaco Basin, 1000 km northeast of Laguna Angel. However, the pollen are showing similar climatic conditions as Cueva del Tigre Perdido and the Santiago Cave. Laguna Angel might therefore be showing southern hemisphere climate conditions. This could be the result of moisture arriving from the south instead of from the northern hemisphere and ITCZ.
To reconstruct the story of people in the Llanos Orientales, charcoal data fire reconstructions and environmental reconstructions from pollen countings from other lake sediments in the Llanos Orientales should be compared with Laguna Angel. These lakes could show different charcoal abundances with another link to climate. Also, phytolith counting can add to the counted pollen and might show other plant species reacting on climate changes or as an indicator of human impact. For example Musa trees or Zea mays . This could lead towards new insight in the paleoclimate of the savanna. Grain size distribution analysis and total organic carbon analysis could also add up to the climatic history of the Llanos Orientales.
8. Acknowledgements
I would like to thank prof. dr. Henry Hooghiemstra and dr. William Gosling for supervision, all the constructive feedback on my thesis, helping me with planning and giving me the possibility to do such an awesome project. Dr. Crystal McMichael for answering my questions and giving critical notes on the interpretation, even though she was not my supervisor. Dr. Kenneth Rijsdijk for bringing me in contact with prof. dr. Henry Hooghiemstra and giving me some extra time in the end. Annemarie Philip, analyst from the UvA pollen lab, for all explanation even though she seemed very busy. Britte Heijink Bsc and Steven Rolofes Bsc for explanation of microscopic analysis of charcoal particles, C2 and R studio. Wim Vessies for support and checking my thesis on spelling and grammar.
9. Appendices
Appendix A: Data
Supplementary data to this article can be found online at:
https://drive.google.com/drive/folders/1Cierkl0uFL9H1_NcDrUerU_ooK5R2HI6?usp=sharing
Appendix B: Pollen diagram
Pollen diagram showing all identified pollen and spores. Pollen zones follow the original paper by Behling & Hooghiemstra (1998), the charcoal zones (this thesis) are shown in the background in tones of grey. The full size figure can be downloaded from the link in Appendix A
Appendix C: Sample Preparation Protocol
For processing the charcoal samples, the following method is used (following Whitlock & Larsen, 2001):
Materials
- Erlenmeyers (size) for subsamples - Erlenmeyers for H2O2
- Resealable plastic bags (size) for samples - Resealable plastic bags (size) for fractions
- Two sieves; 160 μm and 75 μm, fitting on eachother - Electric griddle
- subsampling tube - H2O2
- Pen and paper (Lab journal) - Petri dishes
Steps
1. Take a sufficient amount of sample every x cm along the profile and put these into a small resealable plastic bag. x is determined by the amount of samples taken for Pollen analysis by Behling and Hooghiemstra (1998; 1999).
2. Subsample 1,3 cm3 of every sample with a subsampling tube and put these into a small Erlenmeyer.
3. Add enough H2O2in the Erlenmeyer to submerge the subsample and add a stirring rod. Do not pour the H2O2 straight from the bottle, but use an Erlenmeyer or so.
4. Put the Erlenmeyer with H2O2 and subsample on a griddle (150 ℃) and stir frequently to prevent burning. Write down the time you putted the Erlenmeyer on the griddle.
5. After 10-15 minutes, the subsamples have reacted with the H2O2 and are disintegrated. Take them of the griddle. Write down the time you took the Erlenmeyer of the griddle.
6. Sieve every subsample with two sieves to get two fractions: >160 μm and 75-160 μm and carefully put every fraction into a small resealable plastic bag or a petri dish. Write down the time you poured the subsample into the sieves.
Filled in example of record table
Sample Depth On Griddle
[hh:mm] Of griddle [hh:mm] Sieve [hh:mm] 1 4 14:04 14:18 14:25 2 8 14:04 14:18 14:28 3 12 14:05 14:19 14:34 4 16 14:05 14:19 14:37 5 20 14:05 14:19 14:41 6 24 14:05 14:19 14:45
10. Literature list
Andersen, A. N., Cook, G. D., & Williams, R. J. (2003). Fire in tropical savannas: the Kapalga experiment (Vol. 169). Springer Science & Business Media.
Armenteras, D., Romero, M., & Galindo, G. (2005). Vegetation fire in the savannas of the Llanos Orientales of Colombia. World Resource Review, 17(4), 531-543.
Behling, H., & Hooghiemstra, H. (1998). Late Quaternary palaeoecology and palaeoclimatology from pollen records of the savannas of the Llanos Orientales in Colombia. Palaeogeography Palaeoclimatology Palaeoecology, 139(3-4), 251-267.
Behling, H., & Hooghiemstra, H. (1999). Environmental history of the Colombian savannas of the Llanos Orientales since the Last Glacial Maximum from lake records El Pinal and Carimagua. Journal of Paleolimnology, 21(4), 461-476.
Bender, M. L. (2013). Paleoclimate. Princeton University Press.
Bernal, J. P., Cruz, F. W., Stríkis, N. M., Wang, X., Deininger, M., Catunda, M. C. A., Ortega-Obregón, C., Cheng, H., Edwards, H.L. & Auler, A. S. (2016). High-resolution Holocene South American monsoon history recorded by a speleothem from Botuverá Cave, Brazil. Earth and Planetary Science Letters, 450, 186-196..
Berrio, J.C., Hooghiemstra, H. (2007). Pollen Records, Late Pleistocene | South America. Amsterdam, Netherlands: Elsevier.
Blaauw, M., & Christen, J. A. (2011). Flexible paleoclimate age-depth models using an autoregressive gamma process. Bayesian analysis, 6(3), 457-474.
Blaauw, M., Wohlfarth, B., Christen, J. A., Ampel, L., Veres, D., Hughen, K. A., ... & Svensson, A. (2010). Were last glacial climate events simultaneous between Greenland and France? A quantitative comparison using non-tuned chronologies. Journal of Quaternary Science: Published for the Quaternary Research Association, 25(3), 387-394.
