A 5100-year fire reconstruction from western
Amazonia
Author: Niels Heining Supervisor: dr. C.N.H. McMichael Date: July 4, 2016 Institution: Institute of Biodiversity and Ecosystem Dynamics, University of Amsterdam Bachelor: Earth ScienceTable of Contents 1. Introduction 3 2. Materials and methods 4 2.1 Site description 4 2.2 Field and analytical methods 6 3. Results 8 4. Discussion/conclusion 11 4.1 Drought events and fire 11 4.2 Previous Amazonian charcoal records 14 4.3 Rainforest response to a drier climate 17 5. References 18 Abstract It is expected that the Western Amazon will experience less precipitation and higher temperatures in the near future caused by global climate change. The response of tropical rainforest to drought events are poorly understood. During the mid- Holocene (9000-4000 cal BP), climate was drier than it was in the late- Holocene (4000 cal BP – present). A lake sediment core taken from lake Kumpak, Ecuador, is used to reconstruct fire frequency in western Amazonia over the last 5,100 years. Charcoal analysis was used to reconstruct fire frequency. The Kumpak charcoal record is related to climate and other charcoal records from western Amazonia. Kumpak’s fire reconstruction showed an undisturbed site with no significant fire peaks over the mid- and late- Holocene. Nearby lake Ayauchi showed a pattern with a high fire frequency in the late- Holocene, caused by human disturbance. It is concluded that the wettest part of the Amazon was not susceptible for natural fire to occur, but based on other sites, this conclusion cannot be extrapolated for the whole western Amazon basin.
1. Introduction
The Amazon is a huge area with almost 5.5 million km2 tropical rainforest and is the world’s most biodiverse area, containing approximately 16,000 estimated tree species (ter Steege et al., 2013), with one out of ten known species in the world found in the Amazon (Wright, 2002). Besides that, the rainforest is of great importance for global climate patterns. The Amazon is seen as the ‘lungs of the world’ because of the supply of large amounts of oxygen to the atmosphere (Moran, 1993). The Amazon rainforest regulates atmospheric and terrestrial energy flows and has an influence on global carbon, water and nutrient dynamics (Poore, 1976). Global climate models predict a decrease in precipitation and increased temperatures for tropical regions in the future due to anthropogenic climate change, which can cause severe drought events (Allen et al., 2010). These changed climatic conditions can severely alter the climatic functions of the Amazon, but the implications of a drier climate for the Amazon basin, and the response of tropical tree species, are not fully understood yet. One of the major risks concerning a warmer and drier climate is an increase of forest fires. Nowadays, some of the driest parts of the Amazon are vulnerable for natural fire by lightning (Zeng et al., 2008), but most fires burn themselves out in a few meters, because the surrounding vegetation and soils are too wet (Ray et al., 2005). Thus, most fires in Amazonia under the modern climate regime have an anthropogenic origin (Bush et al., 2008; Bush et al., 2015). Trees in tropical rainforests are poorly adopted to fire stress and even low-intensity fires could be a risk for tropical trees (Barlow and Peres, 2008).
One hypothesis is that the upcoming reduction in precipitation will lead to a widespread dieback of forest vegetation, which can in turn lead to a conversion from a rainforest ecosystem to a savannah (Betts et al., 2004; Cox et al., 2004; Huntingford et al., 2013). An increased frequency of extreme seasonal drought events can further reinforce the intensity of forest fires by increasing the desiccated fuel load (Barlow and Peres, 2008). Other studies suggested that the Amazonian rainforest is more resilient than previously thought, and will be able to maintain its high biodiversity (Mayle and Power, 2008; Nepstad et al., 2007). Understanding the extent to which Amazonian forests have experienced past fire and the associated vegetation responses is crucial for examining long-term resilience of tropical rainforests. One analog for studying how forests will respond to upcoming drought is assessing how they have responded to past droughts, such as during the mid- Holocene. Climatic records showed that climate in the Amazon was drier in the mid- Holocene (c. 9000 to 4000 calibrated years before present (cal BP)) compared to the late- Holocene (c. 4000 cal BP to present) (e.g. Mayle and Power, 2008; Mayle et al., 2000). The onset of the wet Amazonian climate as we know it today was spatially heterogeneous. Studies on Andean lakes showed a rise in lake levels ranging from 5500 to 3000 years ago (Baker et al., 2001; Hillyer et al., 2009; Thompson et al., 1998). Van Breukelen et al. (2008) studied a speleothem calcite record from the Tigre Perdido Cave in lowland Peru. They found that climate started to get wetter from c. 4000 cal BP, but that the mid-Holocene dry period lasted till c. 3000 years ago, as is confirmed by a comparable stalagmite study by Mosblech et al. (2012). The mid- to late- Holocene difference in climate was caused by three factors (Bush et al., 2007c). First of all, millennial-scale variability in hydrologic balance of lake levels is predicted to be a function of a changing relative distribution of insolation over the lower latitude of South America and the adjacent oceans (Rowe et al., 2002). The effect of
insolation was increased seasonality with dry summers/springs around 5000 years ago. Secondly, millennial-scale Holocene climate variability is caused by changing frequency and intensity patterns of El Nino/Southern Oscillation (ENSO). The 2-8 year periodicity of ENSO has been significant since 5000 cal BP and increased in frequency till 1200 cal BP (Moy et al., 2002). Thirdly, migration of the the Intertropical Convergence Zone (ITCZ) over the Holocene, caused climatic variability during this time period (Haug et al., 2001). The direction of change forced by migration of the ITCZ depended on the geographic location compared to the equator. Amazonia south of the equator experienced drier climatic conditions over the mid- Holocene, forced by a southward movement of the ITCZ over this time period (Haug et al., 2001).
Because naturally ignited fires did not occur in the Amazon basin, and human disturbance is seen as the main driver of Holocene fires in the wet tropical Amazon (e.g. McMichael et al., 2011; Bush et al., 2015), I hypothesized that the drier climate of the mid-Holocene had no significant effect on fire occurrence in Western Amazonia. Several fire reconstructions from western Amazonia contained an increase in fire occurrence after 3000 years (McMichael et al., 2012). However, other sites showed signs of increased fire frequency in the mid- Holocene, such as Lake Gentry in Peru, which contained a significantly higher fire frequency between 6000 and 4000 years ago (McMichael et al., 2012). Because of the influence of human disturbance on fire occurrence, the effect of climatic factors is often difficult to interpret. Here we present a fire history from western Amazonia and address the following question: To what extent is fire frequency in Western Amazonia related to drier climatic conditions during the mid- to late- Holocene?
2. Materials and methods
Charcoal analysis was conducted on a sediment core from Lake Kumpak, Ecuador in Western Amazonia, to reconstruct fire history over the mid- to late- Holocene. The sediment core archived information about the last 5500 years and could be analysed in high resolution. The reconstruction gave information about potential fire events and human disturbance around Lake Kumpak. The relationship between climatic conditions and fire frequency in Western Amazonia was studied by using the new Lake Kumpak data and by analyzing previous charcoal datasets from lake sediment cores in the same region.
2.1 Site description
Lake Kumpak (3°2’S, 77°49’W) is located in the Western Amazon basin in the province of Morono-Santiago, Ecuador (fig.1, 2). The lake is an old volcanic crater situated on the foothills of the Andean mountains at an elevation of 350 meters with a surface area of 1.29 km2. The lake is roughly circular and has steep slopes at the edges, though the bottom shelves gently to the center (Liu and Colinvaux, 1988). The steep slopes of the lake caused high erosion, and consequently high sedimentation rates. Lake Kumpak is a closed lake, which means that there is no outflow, suggesting that the water level is mainly controlled by precipitation. It has a small watershed and is mainly influenced by the local environment (Colinvaux et al., 1985) Inflow comes from narrow streams that deliver sediment-laden water to the lake, mostly after the occurrence of a tropical storm or a landslide (Liu and Colinvaux, 1988). High sedimentation rates, up to 1 cm year-1, are observed during these
events, creating thick sediment layers. Consequently, a high resolution charcoal record could be constructed.
Climatic conditions in the Kumpak area support dense tropical rainforest, with a mean annual temperature of 24 degrees Celsius and about 2000-3000 mm precipitation year-1. There is no dry season, although a small decrease in precipitation occurs in December and February (Liu and Colinvaux, 1988).
Figure 1: Location of Lake Kumpak and Lake Ayauchi in Amazonia. Retrieved from: Bush and McMichael, unpublished paper Figure 2: Aerial photo of Lake Kumpak. Nowadays some forest around the lake is cleared and a small village is situated at the shore. Retrieved from: Bush and McMichael, unpublished paper
Lake Kumpak has been studied earlier in the 1980’s by Colinvaux and colleagues (Colinvaux et al., 1985; Liu and Colinvaux, 1988). By then, this kind of research was
pioneering and Lake Kumpak provided one of the first Amazonian sediment cores archiving information over a multi-millennia timescale. Colinvaux’s research was conducted on a lower temporal resolution (subsample taken every 30-40 cm in first 9.7 m, 60-100 cm in bottom 9 m) than was done in this research (subsample for every 5 cm for 5500 - 3000 cal BP). They focused on lake properties and made a forest and climate reconstruction based on sedimentation rates and pollen analysis. Liu and Colinvaux (1988) did not find any charcoal or other human disturbance indicators on a large scale, besides a single Mays Zea pollen grain dated 1150 cal BP (Liu and Colinvaux, 1988).
The original cores are no longer available for analysis. In 2014, McMichael and Bush led a field expedition to collect a high-resolution sediment core from Lake Kumpak using a Colinvaux-Vohnout piston corer (Bush and McMichael, unpublished paper). The new core sample is eighteen meters long and contains laminated sediments, which makes it applicable for analysis at decadal to sub-decadal time scales. An age-depth model was constructed based on 15 accelerator mass spectrometry (AMS) dates, giving a chronology spanning 5100 cal BP (fig.3). The age-depth relationship was modeled using Bacon source code for R (Blaauw and Christen, 2011).
Figure 3: Age-depth model of the Lake Kumpak record. Values are interpolated based on 14C techniques
(Blaauw and Christen, 2011). Retrieved from: Bush and McMichael, unpublished paper
Another study has been conducted on Lake Ayauchi (Bush et al., 1989), a lake 25 kilometers away from Lake Kumpak with comparable characteristics, although it is situated closer to a river (Colinvaux et al., 1985). Lake Ayauchi was first cored in 1983 and re-cored in 2007 (McMichael et al., 2012) and 2016 to obtain a higher resolution sediment core. Bush et al. (1989) used pollen and phytolith analysis and found evidence for maize cultivation 6000 years ago. The maize pollen was accompanied by an increase in herb taxa, characteristic for disturbed land. No evidence was found for year-round occupation and a lake-level rise 800 years BP, was assumed to have led to abandonment of maize agriculture (Bush et al., 1989). McMichael et al. (2012) used the new core and additional soil samples in the lake area to
analyze Lake Ayauchi on a higher resolution. They focused on charcoal abundance and found a temporally discontinuous fire history with distinct peaks throughout the record. Fire occurrence increased after 3000 years ago. McMichael et al. (2011) infer that the charcoal found after 4400 years BP are caused by anthropogenic activities. Before this time frame, the charcoal found might be related to natural fire. The record from Lake Kumpak can provide evidence for the natural fire hypothesis, in case charcoal peaks are found at a time period not directly linked to human disturbance.
2.2 Field and analytical methods
The abundance of macroscopic charcoal is a useful proxy for local fire reconstructions (Whitlock and Larsen, 2002). Charcoal is an inorganic carbon compound that forms when organic material is exposed to temperatures ranging from 290-500 degrees Celsius and an incomplete combustion took place. Due to this incomplete combustion and a lack of oxygen, charring and carbonization produce charcoal (Patterson et al., 1987). Charcoal is persistent on geological time scales due to its resistance for oxidation and microbial activity and can therefore be used as a fire proxy over millennial timescales (Mooney and Tinner, 2011). While the presence of macroscopic charcoal is a solid indicator of local fire influence (Whitlock and Larsen, 2001), a lack of charcoal abundance does not directly mean an absence of fire over this period. Ohlson and Tryterud (2000) conducted an experiment in which they did not find charcoal in 14% of the plots they burnt. Charcoal presence in the lake sediment core comes from an actual fire event, but there is also a certain probability that charcoal will be washed into the lake sediments for years after a fire event (Higuera et al., 2009).
The Lake Kumpak core was subsampled in Florida by Christine Akesson and each sampled contained 0.5 cc3 of sediment. Samples were collected at five centimeters, approximately spanning 10-20 years, throughout the core. At this temporal resolution the risk of missing important events is limited. All samples were bleached for one night with 10% H2O2 to remove darker, non-charcoal organic materials. Subsequently, the sample with H2O2 was boiled at 200 degrees Celsius to break the organic material down. Finally, the sample was sieved with a 160 𝜇m sieve. This was done to only allow charcoal from a relevant spatial distance of origin to be found under the microscope (Whitlock and Larsen, 2002; Ohlson and Tryterud, 2000). Charcoal particles that can be transported by wind over large distances should not be taken into account as we are interested in local fire events surrounding Lake Kumpak. Because the small watershed and the exclusion of microscopic charcoal, only local charcoal was measured (fig.4). This was done to exclude influence from human settlements at Lake Ayauchi for example, to be able to distinguish the fire records of the two lakes. It has to be taken into account that charcoal can be deposited in the lake, several years after a fire event (Patterson et al., 1987). Charcoal can be abundant in the soil surrounding the lake and be deposited years later after a storm or landslide event. This charcoal in the sediment is called secondary charcoal, contrary to primary charcoal that is deposited in the lake sediment immediately after a fire event (Whitlock and Larsen, 2002; fig.4).
Figure 4: Schematic overview of different ways charcoal can reach the lake sediment. Regional and extralocal
charcoal will be excluded from the Kumpak fire reconstruction. Retrieved from: Whitlock and Larsen, 2002.
Charcoal can visually be observed under the microscope by its black colour and opaque, angular and usually planar fragments (Whitlock and Larsen, 2002). Charcoal fragments were observed using an Olympus SZH10 microscope (magnification x7 – 70) and photographed with a Fujifilm X-M1 camera. Surface area of the charcoal is measured using Image-J software (Schneider et al., 2012) into mm2 /cm3. Charcoal surface area was converted into volume using the formula used in Weng (2005). Peak detection was used with the program CharAnalysis to decompose the record into low- and high-frequency components and to separate signal from noise (Higuera, 2009). Charcoal found in the record was interpolated using a mean sampling resolution of 15 years. Charcoal accumulation rates (CHAR) were calculated based on the interpolated dataset. This record was decomposed by subtracting 100 year trends to smooth background CHAR, estimated using a locally weighted regression. The charcoal peaks were detected with CharAnalysis by subtracting background from interpolated CHAR. A Gaussian mixture model was used to define noise-related variations in the peak series. The value 0.05 was used as the cut-off probability for minimum count analysis and peak frequencies were smoothened over 200 years (Huisman, unpublished paper). Seringe Huisman provided the charcoal data for the first 3000 years of the Lake Kumpak record. 3. Results The Lake Kumpak sediment core showed no signs of charcoal over most of the record. One small charcoal piece over the 3000-5100 cal BP timeframe was observed around 3375 cal BP (fig. 5). The charcoal piece fell into the drier mid- Holocene period (Van Breukelen et al., 2008), but is not significant enough to state that it corresponded with a significant fire event. No other traces of charcoal were found over this timeframe in the record. It should be noted that the period between 3900 and 4100 cal BP were not sampled for charcoal
analysis because of disturbance in the sediment core. This is a period fitting well into the mid- Holocene dry period and could have given interesting results.
Figure 5. Raw and interpolated charcoal accumulation rates from Lake Kumpak (5500 - 3000 cal BP), with
charcoal presence found only around 3378 cal BP. The orange line indicates the transition between the mid- and late- Holocene, around 4000 cal BP.
Looking at the Lake Kumpak charcoal record from 5100 cal BP till present, one significant fire peak appeared (fig.6b). Other pieces of charcoal found in the record are insignificant. A modern fire in the twentieth century was the only evidence for fire around Lake Kumpak. The forest clearance around Lake Kumpak visible on the aerial photograph taken in 2014 (fig.2), likely corresponded with the charcoal peak around 1960 A.D. The charcoal particle found in the transition phase of the mid- to late- Holocene is probably linked to a forest clearance in this time period. The Mays Zea pollen found around 1150 cal BP by Liu and Colinvaux (1989) could not be linked to charcoal presence in the Lake Kumpak record.
Charcoal peaks at Lake Ayauchi were found over the whole timespan, and fire frequency at lake Ayauchi increased around 3000 cal BP, at the onset of the wetter late- Holocene (fig.7). The first piece of fire, however, was found in the mid-Holocene around 6000 cal BP, with maize pollen also found in this same time period (Bush et al., 1989). The increase in charcoal abundance in the late-holocene also corresponded with increased agricultural evidence found in the Ayauchi pollen record (McMichael et al., 2012). Figure 9 shows a combined figure with the Kumpak and Ayauchi charcoal data. It became clear that most Kumpak charcoal abundance did not correspond with a peak at Ayauchi, except for a small peak c. 2100 cal BP.
Figure 6. a) Charcoal record from Lake Kumpak, showing charcoal abundance (raw CHAR) c. 3375, 2100, 710,
500 and 0 cal BP. b) Shows a peak detection figure with the use of the program CharAnalysis. Background CHAR is separated from actual fire peaks using interpolated data. The only significant fire peak at the Kumpak site is detected in modern time. Figure 7: a) Charcoal record (raw CHAR) from Lake Ayauchi, 20 km away from Lake Kumpak, showing increased
charcoal abundance after 3000 cal BP. The increase is related to human disturbance around this lake. Fire events before 3000 cal BP, can be related to the mid- Holocene dry period. b) Peak analysis with interpolated data shows that most charcoal pieces are identified as a fire peak.
Figure 8: Graph Showing charcoal abundance of Ayauchi and Kumpak in one figure. It is shown that values of the Kumpak peaks are significantly lower and do not correspond with a fire event at Ayauchi. The green line indicates the transition between the mid- and late- Holocene, around 4000 cal BP. 4. Discussion/Conclusion Lake Kumpak provided one of the first undisturbed sediment cores from western Amazonia. Sedimentation rates up to 1 cm year-1 provided the opportunity to analyze paleoecological changes around the lake at high temporal resolution (i.e. 5 years between samples) over the last 5100 years. Lake Kumpak did not show signs of human activity in ancient times (Liu and Colinvaux, 1988), which is a finding confirmed by this charcoal record. Thus Lake Kumpak has provided one of the first sediment cores with a marked absence of human influence on a millennia-timescales. In contrast, nearby Lake Ayauchi, which is analyzed in previous research, showed a high fire frequency over the mid- and late- Holocene and can therefore be classified as a highly disturbed sediment record, located in the same climatological area as Lake Kumpak. These two lakes provided an excellent opportunity to evaluate the effect of drought events with and without the inference of human settlements and therefore to distinguish between natural- and anthropogenic-induced fire.
4.1 Drought events and fire
There is substantial evidence for a mid- Holocene dry period in the Amazon (Mayle et al., 2000; Mayle and Power, 2008). Most of this research is based on lake levels in high-elevation Andean lakes, but comes from multiple proxies: Peak dust concentrations and snow accumulation minima in Andean ice cores (Thompson et al, 1998), oxygen isotope ratios in lacustrine calcite (Seltzer et al., 2000) and diatom, geochemical and seismic evidence for lake-level low-stands in Lake Titicaca (Baker et al., 2001). Van Breukelen et al. (2008) distinguished between temperature and precipitation. The data suggested a drier period ended around 4000 years ago. It can be assumed that regional climate conditions, i.e.
at Kumpak and Ayauchi are comparable, so drought events causing a widespread fire should have been visible both in the Kumpak and Ayauchi records.
Figure 9a, b showed the charcoal record of lake Kumpak and Ayauchi in the same figure as isotopic data from Caverna Tigre Perdido. It became clear that no significant relationship exists between the mid-Holocene dry period and charcoal abundance in lowland Ecuadorian Amazonia. At Lake Kumpak, the two small charcoal peaks did not correspond to a peak in the δ18O data. The amount of charcoal found is not sufficient to conduct a statistical analysis with these two datasets.
A higher charcoal abundance is found in the late- Holocene wetter period at Lake Ayauchi. No clear match between the charcoal peaks and high δ18O values was visible here either. These data confirmed the hypothesis that the drier climate of the mid- Holocene had no effect on fire frequency in western Amazonia. It was shown that even periods of known drought are not enough to cause natural fire in the western Amazon. This finding was in line with the fact that most Amazonian plants are not evolutionarily adapted to dealing with fire (Mayle et al., 2004). It is anticipated that with a higher fire frequency, Amazonian fire-sensitive species will be replaced by fire-adapted species (Cochrane and Laurance, 2008).
The Kumpak and Ayauchi charcoal record provided evidence for the hypothesis that the mid- Holocene dry period did not directly relate to a higher fire frequency. However, a combination of a drier climate and human disturbance showed an increased fire frequency, as was seen in the Ayauchi record. It can be concluded that in the wettest part of the Amazon, human influence was necessary to ignite fire, what is in line with previous research (e.g. McMichael et al., 2012; Bush et al., 2007c). Fires were used by pre-European settlers on a daily basis for cooking or land clearance activities for agricultural purposes (Bush et al., 2016). Small human-induced fires could have escaped to form larger wildfires in periods of strong droughts (Cochrane and Laurance, 2008; Nepstad et al., 2001). Most fires found in western Amazonia could be related to pollen data that provided evidence for agricultural activity (Bush et al., 2007b). Predicting the effect of people in combination with fire and drought events in the landscape is an interesting field for further research and environmental management.
Figure 9. Comparison of the Cueva del Tigre Perdido δ18O fluid inclusion (orange line) with the charcoal records from a) Lake Kumpak and b) Lake Ayauchi. Higher δ18O values indicate a drier climate. The figures show no clear relationship between precipitation rates in the Peruvian Amazon and charcoal abundance in Ecuadorian
Amazonia. The charcoal abundance in modern times found in the Kumpak record is not shown, because of the lack of isotope data from the Tigre Perdido. Figure 10: Precipitation map of South-America, with the sites of the several lakes from table 1 and Cueva del Tigre Perdido (Van Breukelen et al., 2008) included. Lake Kumpak and Ayauchi are situated in a wetter part of the Amazon basin. Made with ArcGIS. 4.2 Previous Amazonian charcoal records
Neither Kumpak nor Ayauchi contained evidence that mid- Holocene drought was strong enough to induce naturally formed fires to spread. These lakes are situated in Ecuador (northwest Amazonia) and showed different results compared to less wet areas in Peru (southwest Amazonia) (Bush et al., 2007a). A comparison between Kumpak/Ayauchi and other western Amazonian lake records, provided evidence for the prediction that mid- Holocene drought was heterogeneous in its timing and its intensity across space (Mayle and Power, 2008). The results of Lake Kumpak and Ayauchi could therefore not be extrapolated for whole western Amazonia. Climatic variability within the Amazon caused significant differences in fire response to drought. Kumpak and Ayauchi are situated in the wettest part of the Amazon (fig.10), an area that might have been too wet for natural fire to occur, even in the dry mid-Holocene. Areas that are not vulnerable for natural fire nowadays, but receive less precipitation compared to the Kumpak area, might become susceptible for natural fire in the near future due to anthropogenic climate change. The widespread dieback of tropical rain forest (Cox et al., 2004; Betts et al., 2004) caused by reduced precipitation might be of risk for those tropical rainforest ecosystems, not in the wettest part of the Amazon.
As stated before, Kumpak and Ayauchi are situated in the wettest part of the Amazon, what has a consequence for the relationship between climate and fire. Precipitation rates in the Kumpak/Ayauchi area are 2000-3000 mm yr-1 (Liu and Colinvaux, 1988; McMichael et al., 2012). To get a more thorough picture of the effect of drought on
fire in Amazonian tropical rainforests, other research sites with different precipitation regimes can be taken into account. The location of the research site and the corresponding precipitation pattern was of great importance for the relation between drought and fire. Figure 10 shows a current precipitation map with the different charcoal record sites, table 1 provides an overview of the abundances of charcoal and human disturbance in those records. Human disturbance is important, because it was likely the main cause of fire events and here we were interested in distinguishing natural-ignited fire.
4.2.1 Northwestern Amazonian lakes
Similar conclusions to the Kumpak/Ayauchi area were drawn based from a lake sediment core close to the Yasuni river, lowland Ecuador (Weng et al., 2002). Another northwestern Lake in Peru showed contrary results (Bush et al., 2016). Yasuni lake Maxus 4 encompasses the last 9500 years and contained therefore information about fire frequency in the mid- and late- Holocene. Mean annual precipitation exceeds 3000 mm yr-1, with rainfall occurring in all months and natural fires do not occur in this wet area nowadays (Weng et al., 2002). The core provided evidence that this was neither the case over the last 9500 years. No charcoal peaks were found in the entire core. The pollen record of Maxus 4 did not show any signs of human disturbance over the core as well (Weng et al., 2002). A relatively dry period between 4900-3700 cal BP was detected, but the forest remained wet enough for natural fire to not occur. Lake Sauce receives c. 1475 mm mean annual precipitation with a dry season between November and December. Bush et al. (2016) stated that most of the charcoal found throughout the record (the last 6900 years) was related to anthropogenic activity, but that two profound peaks at 6700 cal BP and 4270 cal BP might have correspond to megadroughts in the dry mid- Holocene.
4.2.2 Southwestern Amazonian lakes
Lake Gentry, Parker, Vargas and Werth are all located in the same watershed in the Madre de Dios region in southeastern Peru. The Gentry-Parker area is drier than Kumpak with annual precipitation being 1700-2000 mm and a dry season covering 2-4 months (Bush et al., 2007a). Still, under current climatic conditions forest fires are supposed not to occur naturally (Bush et al., 2007a). Lake Werth only formed 3300 years ago and archives therefore no information about the mid- Holocene, but it contained zero evidence for fire over the the late- Holocene. Lake Vargas showed a big gap in the record (7200-2500 cal BP) because it dried out during this timeframe. Therefore, it did not contain information about fire frequency over the mid-Holocene either, but provide evidence for a mid- Holocene dry period. The charcoal peaks at Gentry in the mid-Holocene are expected to be related to times of regional drying and not to the abundance of maize and cassava pollen, because fire frequency was higher in the mid- Holocene than when people and maize came in during the late- Holocene (McMichael et al., 2012). At lake Parker, charcoal was abundant through all of the core from 7200 cal BP until 250 cal BP, but no evidence for early human disturbance was detected (Bush et al., 2007a). The fires in all of the four lakes after 3000 cal BP were assumed to be caused by human interference because of the present inability of the forest to burn naturally (Bush et al., 2007a).
At the boundary between rainforest and savannah lies the Noel Kempff Mercado National Park (NKMNP) in lowland Bolivia. The two lakes, Laguna Chaplin and Bella Vista, used in research by Burbridge et al. (2004) are partly located in riverine Amazonian and humid evergreen forests. The park receives c. 1500 mm of annual precipitation with a dry
season that lasts for six months. The cores covered 50000 years of sediment in low temporal resolution. They found that savannah vegetation expanded during the last glacial period till climate turns wetter at the beginning of the late- Holocene. Charcoal abundance showed a peak around 9000 cal BP in both lakes, with decreasing abundance till present (Burbridge et al., 2004). This means there was a sharp relation between fire and climate in this region on the southern limit of the Amazon basin.
Table 1. Previous research on sediment cores. If charcoal was found, the related time period is given, as is
evidence found for human disturbance. It is indicated if there is coherent relationship between the mid- Holocene and fire frequency.
Location Charcoal Human
disturbance
Dry
mid-Holocene/fire relationship
Source
Lake Kumpak - 1500-2200 BP No Liu and Colinvaux,
1988 Lake Ayauchi Yes, found from
6,000 yrs ago with an increased frequency around 3,000 yrs ago Maize pollen found around 6,000 yrs ago, peak disturbance from 3,000-450 yrs BP
No Bush et al., 1989; McMichael et al., 2011
Lake Gentry Charcoal found from 6,000 years ago. Decrease after 4,000 years ago. Cultivation of maize 6,000 to 500 yrs bp
Yes McMichael et al., 2011; McMichael et al., 2012a; Bush et al., 2007a Lake Parker Relative consistent
charcoal record from 7,200 years ago till 500 years ago
Charcoal peaks are related to human disturbance. Occuring all over the record
Possible McMichael et al., 2011; McMichael et al., 2012a; Bush et al., 2007a Lake Maxus 4 No charcoal found
over the last 95n 00 years
- No Weng et al., 2002
Lake Sauce 6700 and 4270 BP Maize pollen from 6320 BP. 3380-700 BP peak maize pollen
Yes Bush et al., 2016
Laguna Chaplin and
Bella Vista Abundant throughout the whole record, peak around 9000 BP
No clear evidence of agricultural activity
Yes Burbridge et al., 2004
From these sites around western Amazonia, summarized in table 1, it is concluded that the results from Lake Kumpak and Ayauchi cannot easily be extrapolated over a large area. In the wettest parts of the Amazonian rainforest without a profound dry season, it may be concluded that natural fires do not occur nowadays and presumably also did not in the drier mid-Holocene. In slightly drier areas, the drier mid-Holocene might have positively affected fire occurrence, although fire is not a natural phenomenon in this areas today. At
the limits of the evergreen tropical rain forest, relatively dry climatic conditions not only increased fire frequency, but also shifted the ecosystem in the direction of a savannah. 4.3 Rainforest response to a drier climate From proxies that helped us to reconstruct past climates, it was inferred that a drier mid-Holocene period existed in western Amazonia (e.g. Van Breukelen et al., 2008; Baker et al., 2001). For future implications, a similar dry period is anticipated caused by increased anthropogenic CO2-emissions over the past centuries (Malhi et al., 2008). However, mid-Holocene climatic circumstances are not one to one relatable to future climate predictions for two reasons. First, by 2050, CO2 concentrations are projected to be at least twice as high as mid- Holocene CO2 concentrations (Mayle et al., 2004). Secondly, the amplitude and frequency of ENSO was an important factor causing increased aridity during El Nino years, but was significantly smaller during the mid-Holocene compared to today (Moy et al., 2002). Moreover, this research showed that people are the main cause of fire in western Amazonia. The large modern influence of people on the rainforest will be the main reason the drought response will be different in the future than it was in the mid- Holocene (Cochrane and Laurance, 2008; Nepstad et al., 2001).
Besides these differences, the mid- Holocene might still be a good analogue for the future. Precipitation in the Amazon basin is expected to reduce by c. 20% over the twenty-first century by climate models of the IPCC (Houghton et al., 2001). A reduction in precipitation would have significant influences on the basin, especially because the rates of change are much higher than what we have seen since the Last Glacial Maximum (LGM) (Malhi et al., 2008). However, climatic responses to higher CO2 levels are not spatially uniform along the Amazon basin. It was expected that the largest effect will be at the eastern part of the basin with reduced precipitation, especially during the dry season, while the western part in contrast may experience a wetter climate. The lesser impact of increased CO2 levels in western Amazonia is because precipitation patterns in this region are mostly controlled by moisture convergence forced by the Andes and not by Atlantic sea surface temperatures (Li and Dickinson, 2006). Although this does not mean western Amazonia is not vulnerable for climatic change. Duration and severity of a dry season are critical factors for evergreen tropical rainforest and Malhi et al. (2008) stated that based on 23 climate models, the chance of intensification of the dry season in western Amazonia is 30%. In this study, it was shown that drier climatic conditions did not directly mean that fire frequency increased in western Amazonia. Human disturbance is seen as the main driver of fire in the tropical rainforest, an effect that can be accelerated by drought events. However, the effects on vegetation are not fully understood yet (Huntingford et al., 2013). A shift to more drought-tolerant species may occur as a response to a drier climate, such as drought-tolerant lianas and semi-deciduous trees, instead of evergreen broadleaf trees (Mayle et al., 2004). Experiments with artificial drought experiments (Fisher et al., 2007; Nepstad et al., 2007), flux towers and satellite remote sensing of forest greenness (Huete et al., 2006) all pointed in the direction that the Amazonian rainforest is more drought-resilient than predictions by global climate models. A main reason for this resilience came from the ability of vegetation for inclusion of deeper rooting (Fisher et al., 2007). Besides that, Cowling et al. (2001) conducted an experiment in which they found that low CO2 concentrations were the main cause of reduced canopy cover, instead of reduced precipitation. Resilience is an important concept for forest conservation practices, as it
provides information about the capacity of the forest to respond to human or climatologically disturbances (Berkes et al., 2000). This research showed that no conclusions based on resilience to past fires can be made for the whole Amazon, as they did not occur over the entire basin. More research on charcoal, pollen and phytolith data is needed to get a better insight in the response of tropical rainforest to changing climatic circumstances and human disturbances. 5. References
Allen, C.D., Macalady, A.K., Chenchouni, H., Bachelet, D., McDowell, N., Vennetier, M., Kitzberger, T., Rigling, A., Breshears, D.D., Hogg, E.T. and Gonzalez, P., 2010. A global overview of drought and heat-induced tree mortality reveals emerging climate change risks for forests. Forest ecology and management, 259(4), pp.660-684.
Baker, P.A., Seltzer, G.O., Fritz, S.C., Dunbar, R.B., Grove, M.J., Tapia, P.M., Cross, S.L., Rowe, H.D. and Broda, J.P., 2001. The history of South American tropical precipitation for the past 25,000 years. Science,291(5504), pp.640-643.
Barlow, J. and Peres, C.A., 2008. Fire-mediated dieback and compositional cascade in an Amazonian forest. Philosophical Transactions of the Royal Society of London B:
Biological Sciences, 363(1498), pp.1787-1794.
Berkes, F., Folke, C., & Colding, J., 2000. Linking social and ecological systems: management practices and social mechanisms for building resilience. Cambridge University Press.
Betts, R. A., Cox, P. M., Collins, M., Harris, P. P., Huntingford, C., & Jones, C. D., 2004. The role of ecosystem-atmosphere interactions in simulated Amazonian precipitation decrease and forest dieback under global climate warming. Theoretical and applied climatology, 78(1-3), pp.157-175.
Blaauw, M., & Christen, J. A., 2011. Flexible paleoclimate age-depth models using an autoregressive gamma process. Bayesian Analysis, 6(3), pp.457-474.
Bush, M. B., Dolores, R. P., & Colinvaux, P. A., 1989. A 6,000 year history of Amazonian maize cultivation. Nature, 340, pp. 303-305.
Bush, M. B., & Silman, M. R., 2004. Observations on Late Pleistocene cooling and precipitation in the lowland Neotropics. Journal of Quaternary Science, 19(7), pp.677-684.
Bush, M.B., Silman, M.R., De Toledo, M.B., Listopad, C., Gosling, W.D., Williams, C., De Oliveira, P.E. and Krisel, C., 2007a. Holocene fire and occupation in Amazonia: records from two lake districts. Philosophical Transactions of the Royal Society of
London B: Biological Sciences,362(1478), pp.209-218.
Bush, M. B., Silman, M. R., McMichael, C., & Saatchi, S., 2008. Fire, climate change and biodiversity in Amazonia: a Late-Holocene perspective. Philosophical Transactions of the Royal Society B: Biological Sciences, 363(1498), pp:1795-1802.
Bush, M. B., McMichael, C. H., Piperno, D. R., Silman, M. R., Barlow, J., Peres, C. A., ... & Palace, M. W., 2015. Anthropogenic influence on Amazonian forests in pre-history: An ecological perspective. Journal of Biogeography, 42(12), pp.2277-2288.
Bush, M.B., Gosling, W.D. and Colinvaux, P.A., 2007b. Climate change in the lowlands of the Amazon basin. In Tropical rainforest responses to climatic change (pp. 55-76). Springer Berlin Heidelberg.
Bush, M. B., Silman, M. R., & Listopad, C. M. C. S., 2007c. A regional study of Holocene climate change and human occupation in Peruvian Amazonia. Journal of
Biogeography, 34(8), pp.1342-1356.
Bush, M. B., Correa-Metrio, A., McMichael, C. H., Sully, S., Shadik, C. R., Valencia, B. G., ... & Overpeck, J. T., 2016. A 6900-year history of landscape modification by humans in lowland Amazonia. Quaternary Science Reviews, 141, pp.52-64.
Cochrane, M. A., & Laurance, W. F., 2008. Synergisms among fire, land use, and climate change in the Amazon. AMBIO: A Journal of the Human Environment, 37(7), pp.522-527.
Colinvaux, P. A., Miller, M. C., Liu, K. B., Steinitz-Kannan, M., & Frost, I., 1985. Discovery of permanent Amazon lakes and hydraulic disturbance in the upper Amazon basin.
Nature, 313(5997), pp.42-45.
Cowling, S. A., Maslin, M. A. & Sykes, M. T., 2001 Paleovegetation simulations of lowland Amazonia and implications for neotropical allopatry and speciation. Quatern. Res.
55, pp:140–149.
Cox, P. M., Betts, R. A., Jones, C. D., Spall, S. A., & Totterdell, I. J., 2000. Acceleration of global warming due to carbon-cycle feedbacks in a coupled climate model. Nature, 408(6809), pp:184-187.
Cox, P. M., Betts, R. A., Collins, M., Harris, P. P., Huntingford, C., & Jones, C. D., 2004. Amazonian forest dieback under climate-carbon cycle projections for the 21st century. Theoretical and applied climatology, 78(1-3), pp.137-156.
Fisher, R. A., Williams, M., Costa, D., Lola, A., Malhi, Y., da Costa, R. F., ... & Meir, P., 2007. The response of an Eastern Amazonian rain forest to drought stress: results and modelling analyses from a throughfall exclusion experiment. Global Change Biology, 13(11), pp.2361-2378.
Haug, G.H., Hughen, K.A., Sigman, D.M., Peterson, L.C. and Röhl, U., 2001. Southward migration of the intertropical convergence zone through the Holocene. Science,
293(5533), pp.1304-1308.
Higuera, P. (2009). CharAnalysis 0.9: Diagnostic and analytical tools for sediment charcoal analysis. User’s Guide, Montana State University, Bozeman, MT.
Hillyer, R., Valencia, B.G., Bush, M.B., Silman, M.R. and Steinitz-Kannan, M., 2009. A 24,700-yr paleolimnological history from the Peruvian Andes.Quaternary Research, 71(1), pp.71-82.
Houghton, J. T., Ding, Y., Griggs, D. J., Noguer, M., Van der Linden, P. J., Dai, X., Maskell, K. & Johnson, C. A. (eds) 2001. Climate Change 2001: the scientific basis. Contribution of
Working Group 1 to the Third Assessment Report of the IPCC. Cambridge University
Press.
Huete, A. R., Didan, K., Shimabukuro, Y. E., Ratana, P., Saleska, S. R., Hutyra, L. R., ... & Myneni, R., 2006. Amazon rainforests green-up with sunlight in dry season. Geophysical research letters, 33(6).
Huntingford, C., Zelazowski, P., Galbraith, D., Mercado, L. M., Sitch, S., Fisher, R., ... & Malhi, Y., 2013. Simulated resilience of tropical rainforests to CO2-induced climate change. Nature Geoscience, 6(4), pp.268-273.
Li, W., Fu, R., & Dickinson, R. E., 2006. Rainfall and its seasonality over the Amazon in the 21st century as assessed by the coupled models for the IPCC AR4. Journal of Geophysical Research: Atmospheres, 111(D2).
Liu, K. B., & Colinvaux, P. A., 1988. A 5200-year history of Amazon rain forest. Journal of
Biogeography, 15, pp.231-248.
Mayle, F.E., Burbridge, R. and Killeen, T.J., 2000. Millennial-scale dynamics of southern Amazonian rain forests. Science, 290(5500), pp.2291-2294.
Mayle, F. E., Beerling, D. J., Gosling, W. D., & Bush, M. B., 2004. Responses of Amazonian ecosystems to climatic and atmospheric carbon dioxide changes since the last glacial maximum. Philosophical Transactions of the Royal Society B: Biological Sciences, 359(1443), pp:499-514.
Mayle, F. E. & Power, M. J., 2008. Impact of a drier Early–Mid-Holocene climate upon Amazonian forests. Philosophical Transactions of the Royal Society B: Biological
Sciences, 363, pp.1829-1838. McMichael, C.H., Bush, M.B., Piperno, D.R., Silman, M.R., Zimmerman, A.R. and Anderson, C., 2011. Spatial and temporal scales of pre-Columbian disturbance associated with western Amazonian lakes. The Holocene, 22(2), pp. 131-140. McMichael, C. H., Bush, M. B., Silman, M. R., Piperno, D. R., Raczka, M., Lobato, L. C., ... & Palace, M., 2013. Historical fire and bamboo dynamics in western Amazonia. Journal of Biogeography, 40(2), pp.299-309.
McMichael, C.H., Correa-Metrio, A. and Bush, M.B., 2012a. Pre-Columbian fire regimes in lowland tropical rainforests of southeastern Peru.Palaeogeography,
Palaeoclimatology, Palaeoecology, 342, pp.73-83.
McMichael, C.H., Piperno, D.R., Bush, M.B., Silman, M.R., Zimmerman, A.R., Raczka, M.F. and Lobato, L.C., 2012b. Sparse pre-Columbian human habitation in western Amazonia. Science, 336(6087), pp.1429-1431.
McMichael, C., Piperno, D., Neves, E., Bush, M., Almeida, F., Mongeló, G., & Eyjolfsdottir, M. B., 2015. Phytolith assemblages along a gradient of ancient human disturbance in western Amazonia. Frontiers in Ecology and Evolution,3, pp. 141.
Mooney, S. D., & Tinner, W., 2011. The analysis of charcoal in peat and organic sediments.
Mires and Peat, 7(09), pp.1-18.
Moran, E.F., 1993. Deforestation and land use in the Brazilian Amazon.Human Ecology,
21(1), pp.1-21.
Mosblech, N. A., Bush, M. B., Gosling, W. D., Hodell, D., Thomas, L., van Calsteren, P., ... & van Woesik, R., 2012. North Atlantic forcing of Amazonian precipitation during the last ice age. Nature Geoscience, 5(11), pp.817-820.
Moy, C. M., Seltzer, G. O., Rodbell, D. T., & Anderson, D. M., 2002. Variability of El Niño/Southern Oscillation activity at millennial timescales during the Holocene epoch. Nature, 420(6912), pp.162-165.
Nepstad, D., Carvalho, G., Barros, A. C., Alencar, A., Capobianco, J. P., Bishop, J., ... & Prins, E., 2001. Road paving, fire regime feedbacks, and the future of Amazon forests. Forest ecology and management, 154(3), pp.395-407.
Nepstad, D.C., Tohver, I.M., Ray, D., Moutinho, P. and Cardinot, G., 2007. Mortality of large trees and lianas following experimental drought in an Amazon forest. Ecology, 88(9), pp.2259-2269.
Ohlson, M., & Tryterud, E., 2000. Interpretation of the charcoal record in forest soils: forest fires and their production and deposition of macroscopic charcoal. The Holocene,
10(4), pp.519-525.
Patterson, W. A., Edwards, K. J., & Maguire, D. J., 1987. Microscopic charcoal as a fossil indicator of fire. Quaternary Science Reviews, 6(1), pp.3-23.
Poore, D., 1976. The values of the tropical moist forest ecosystems and the environmental consequences of their removal. FAO.
Ray, D., Nepstad, D., & Moutinho, P., 2005. Micrometeorological and canopy controls of fire susceptibility in a forested Amazon landscape. Ecological Applications, 15(5), pp:1664-1678.
Rowe, H. D., Dunbar, R. B., Mucciarone, D. A., Seltzer, G. O., Baker, P. A., & Fritz, S., 2002. Insolation, moisture balance and climate change on the South American Altiplano since the Last Glacial Maximum. Climatic Change, 52(1-2), pp.175-199.
Schneider, C. A., Rasband, W. S., & Eliceiri, K. W., 2012. NIH Image to ImageJ: 25 years of image analysis. Nat methods, 9(7), pp:671-675.
Seltzer, G., Rodbell, D. and Burns, S., 2000. Isotopic evidence for late Quaternary climatic change in tropical South America. Geology, 28(1), pp.35-38.
Sternberg L.S.L., 2001. Savanna-forest hysteresis in the tropics. Global Ecology and
Biogeography 10, pp.369–378.
Ter Steege, H., Pitman, N. C., Sabatier, D., Baraloto, C., Salomão, R. P., Guevara, J. E., ... & Monteagudo, A., 2013. Hyperdominance in the Amazonian tree flora. Science, 342(6156), 1243092.
Thompson, L.G., Davis, M.E., Mosley-Thompson, E., Sowers, T.A., Henderson, K.A., Zagorodnov, V.S., Lin, P.N., Mikhalenko, V.N., Campen, R.K., Bolzan, J.F. and Cole-Dai, J., 1998. A 25,000-year tropical climate history from Bolivian ice cores. Science,
282(5395), pp.1858-1864.
Van Breukelen, M.R., Vonhof, H.B., Hellstrom, J.C., Wester, W.C.G. and Kroon, D., 2008. Fossil dripwater in stalagmites reveals Holocene temperature and rainfall variation in Amazonia. Earth and Planetary Science Letters, 275(1), pp.54-60.
Vizy, E.K. and Cook, K.H., 2007. Relationship between Amazon and high Andes rainfall.
Journal of Geophysical Research: Atmospheres, 112, pp.1-14.
Weng, C., Bush, M.B. and Athens, J.S., 2002. Holocene climate change and hydrarch succession in lowland Amazonian Ecuador. Review of Palaeobotany and Palynology,
120(1), pp.73-90.
Weng, C., 2005. An improved method for quantifying sedimentary charcoal via a volume proxy. The Holocene 15, pp.298.
Whitlock, C., & Larsen, C., 2002. Charcoal as a fire proxy. In Smol, J.P., Birks, H.J.B. & Last, W.M. (eds.) Tracking environmental change using lake sediments (pp. 75-97). Kluwer Academic Publishers, Dordrecht, The Netherlands.
Wright, J.S., 2002. Plant diversity in tropical forests: a review of mechanisms of species coexistence. Oecologia, 130(1), pp.1-14. Zeng, N., Yoon, J. H., Marengo, J. A., Subramaniam, A., Nobre, C. A., Mariotti, A., & Neelin, J. D., 2008. Causes and impacts of the 2005 Amazon drought. Environmental Research Letters, 3(1), pp.1-9.
