Reconstruction of the mid-Holocene paleoclimate of the Ecuadorian Andean páramo at Tres Lagunas,Ecuador

Abstract

Previous paleoclimate studies have suggested a warm/dry event during the mid-Holocene in páramo vegetation of the tropical Andes of South America. However, the timing of the mid-Holocene warm/dry event in Tres Lagunas, Ecuador, remains uncertain, since a previously reported bog core record characterized the warm/dry event during a hiatus in sediment deposition. In order to understand the timing of the warm/dry event in Tres Lagunas in relation to regional records, a lake sediment core was collected. Subsamples from the core were analyzed for radiocarbon dates, pollen, magnetic susceptibility, and charcoal to reconstruct the depositional, vegetational, and fire history of the area. A near-constant sedimentation rate in the lake core indicated that the lake did not dry, in contrast to the adjacent previously reported bog core. Increases in lower elevation pollen types suggest a warm period between 2700 and 2200 cal. yr BP, with the peak of warming at 2200 cal. yr BP co-occurring with the end of hiatus in the bog core record. Statistically significant increases in charcoal influx and magnetic susceptibility from 3900 to 800 cal. yr BP also suggest a dry climate during this period. While the lake record corroborates the presence of a warm/dry period in Tres Lagunas during the mid-Holocene, this record suggests a shorter period of warm/dry climate where the intensity was not sufficient to cause the lake to dry. However, anthropogenic alteration of the landscape, either vegetation or burning, must also be considered when interpreting Holocene records from this region.

Keywords

charcoal Holocene warm/dry period magnetic susceptibility paleoenvironment pollen

Introduction

Of the paleoclimate reconstructions from the tropical Andes of South America, many have focused on the Andean páramo. The páramo represents alpine vegetation between treeline and permanent snowline between approximately 3600 and 4100 m in elevation (Di Pasquale et al., 2008; Ledru et al., 2013). The range of páramo vegetation is bounded by four biologically distinct subzones that change over steep elevational gradients, making the Andean páramo particularly sensitive to climatic change and an excellent paleothermometer (Bush et al., 2005; Paduano et al., 2003; Weng et al., 2006). The modern Andean páramo (especially at lower elevations) is subject to overgrazing of livestock and burning to enhance cultivation, both of which reduce biodiversity and potentially impact water retention capabilities (Crespo et al., 2011; Ochoa-Tocachi et al., 2016; Podwojewski et al., 2002). As the human population in the Andes grows and climate changes, understanding the resilience of the páramo is critical. In general, warm climate events cause the páramo vegetation to move upward in elevation as the lower forests shift to higher elevations. While during cool climate events, the páramo and treeline expand downward in elevation because of a redistribution of temperature-sensitive species. In addition to being a sensitive paleothermometer, the presence of naturally induced fires across the páramo landscape since the Pleistocene–Holocene transition 13,000 cal. yr BP allows for the reconstruction of paleomoisture.

Taking advantage of the climate-sensitive landscape, previously collected sediment cores suggest a warm/dry event in the early to mid-Holocene from ~10,000 to 5000 cal. yr BP consistent across the eastern and western cordillera of the tropical Andes, although the timing and extent of such an event still needs to be determined (Hansen and Rodbell, 1995; Hansen et al., 2003; Jantz and Behling, 2012; Paduano et al., 2003; Polissar, 2005; Velez et al., 2006; Weng et al., 2006). It is important to note that this warm/dry period does not necessarily represent a single synchronous event across the region, rather it could represent multiple small-scale events (Bush et al., 2005). One such record collected from a bog core on the eastern ridge of the western cordillera of the Ecuadorian Andes at Tres Lagunas (Figure 1; Jantz and Behling, 2012) attributes a large hiatus in sedimentation occurring from 7600 to 2200 cal. yr BP to the mid-Holocene warm/dry event.

Figure 1.

Map of coring locations in Ecuador, Peru, and Bolivia in South America. Tres Lagunas, Ecuador, is represented by the yellow circles. Sites mentioned in the text are represented by blue squares, where A is Abbott et al. (2003), B is Paduano et al. (2003), C is Weng et al. (2006), and D is Hansen et al. (2003). The inset map shows Tres Lagunas on the volcanic plateau of Quimsacocha. The location of the lake core (QU1-14B) at 3.05145°S, 79.24822°W is shown along with the location of the bog core collected in 2012 by Jantz and Behling at 3.047472°S, 79.24164°W.

The objective of this article was to reconstruct the paleoclimate record at Tres Lagunas on the eastern ridge of the western cordillera of the Ecuadorian Andes to understand the vegetation and fire history, with particular attention to the period of the depositional hiatus identified in the Jantz and Behling (2012) bog core record. Additionally, the timing of the mid-Holocene warm/dry event in Tres Lagunas was compared with other lake cores from high-elevation lakes in the tropical Andes of South America (Figure 1).

Modern climate

The Andean páramo has a humid tropical alpine climate with low temperatures and high precipitation. Average annual temperatures shift diurnally ranging between 12°C and 19°C during the day to 2°C and −3°C at night (Di Pasquale et al., 2008; Hansen et al., 2003; Jantz and Behling, 2012). Annual precipitation in the Quimsacocha and Cajas regions has been reported between 1280 and 2000 mm/yr (Hansen et al., 2003 Jantz and Behling, 2012). However, precipitation varies seasonally with the wetter months of November, March, and April averaging 250–280 mm of rain and the drier months of August and December averaging only 20 mm of rain (Jantz and Behling, 2012). A longitudinal climate gradient is present across the western and eastern ridge of the tropical Andes. Prevailing winds driving precipitation on the western cordillera of the Andes are sourced from the Pacific Ocean, while prevailing winds driving precipitation on the eastern cordillera of the Andes are sourced from the Amazon Basin (Hansen et al., 2003). Therefore, locations in the inter-Andean valley or Cordillera Real are influenced by a combination of the Pacific and Amazonian climates (Campozano et al., 2016), with local wind pattern forced by local topography.

Factors that affect the water balance include the strength of the summer monsoon circulation, changes in seasonal location of the Intertropical Convergence Zone (ITCZ) and El Niño-Southern Oscillation (ENSO; Hansen and Rodbell, 1995; Ledru et al., 2013; Moy et al., 2002; Paduano et al., 2003). Strong El Niño years correlate with wetter years on the western páramo (Moy et al., 2002) and drought in the central eastern Andes (Paduano et al., 2003). Previous records from the tropical Andes found that the position of the ITCZ and the ENSO phenomenon are the most important forces driving precipitation along the mountain range (Flantua et al., 2016). Increased solar insulation results in stronger easterly winds in the tropical Pacific, cooler land temperatures, and increased evapotranspiration (Clement et al., 2000). Additionally, increased insulation can cause northern movement of the ITCZ resulting in stronger southeasterly trade winds, increased upwelling, cooler sea surface temperatures, and reduced precipitation (Conroy et al., 2008; Haug et al., 2001; Mitchell and Wallace, 1991).

Vegetation and altitudinal zones

The Andean environment has been separated into five distinct vegetational zones associated with altitude: superpáramo, páramo, subpáramo, upper montane forest (UMF), and lower montane forest (LMF; Hansen et al., 2003; Sklenar and Ramsay, 2001; Figure 2). The superpáramo is sparsely vegetated and is usually found at approximately 4100–4500 m (Di Pasquale et al., 2008). The superpáramo has low-growing cushion plants from families such as Apiaceae and Asteraceae, forbs such as Chrysactinium, and Gentiana, and important pteridophytes such as Lycopodium and Huperzia (Hansen et al., 2003). Below the superpáramo is the páramo, at approximately 3600–4100 m. The páramo has denser vegetation coverage with tussock grasses from genera Festuca sp. and Calamagrostis sp. interspersed with cushion plants, terrestrial Puya hamata, with tall stem plants such as Asteraceae and small shrubs from Ericaceae, Melastomataceae, and Clusiaceae (Di Pasquale et al., 2008; Hansen et al., 2003). The subpáramo (2600–3600 m) is dominated by shrubs and is bound on the lower elevation by the treeline. Subpáramo vegetation includes Borarea, Clethera, and Gunnera (Jantz and Behling, 2012). The UMF occurs below the treeline (2100–3600 m) and is characterized by low stature trees with sclerophyllous leaves and abundant mosses on trees and the forest floor. The UMF is species poor compared with tropical rainforests in lower altitudes but is rich compared with montane forests in temperate regions (Di Pasquale et al., 2008). Common tree genera found in the UMF are Alnus, Weinmannia, and Ilex. Additionally, shrubs of Ericaceae and Melastomataceae can be found (Di Pasquale et al., 2008; Hansen et al., 2003; Jantz and Behling, 2012). Below the UMF is the LMF, which extends from 1300 to 2100 m, and has trees from the genera Cecropia, Piper, and Celtis as well as the families of Moraceae and Urticaceae (Hansen et al., 2003).

Figure 2.

Zonation of modern-day ecological regime in the páramo of Ecuador at a given elevation. In the event of a warm climate, the vegetation types will shift upward in elevation. In the event of a cool climate, the vegetation type will shift downward in elevation. As previous research has suggested, the vegetation does not move up or down slope as belts rather the centers of distribution change with changing temperature (Colinvaux et al., 1997).

Long-term shifts in the vegetation can be tracked through the pollen record. Up-elevation shifts in vegetation type, for example, LMF at higher-than-current elevation, indicate a warm climate. Conversely, down-elevation shifts in vegetation type, for example, superpáramo at lower-than-current elevation, indicate a cool climate (Hansen et al., 2003). Changes in the altitudinal position of the treeline (UMF boundary) throughout the Andes have been and continue to be instrumental in reconstructing temperature changes (Flantua et al., 2016).

Fire regime

The modern Andean páramo vegetation is burned approximately every 3–6 years by human intervention. The regular burning of the páramo has been attributed to clearing for farming or for livestock grazing (Di Pasquale et al., 2008; Jantz and Behling, 2012). There are no documented records of recent naturally induced fires, which are unlikely to occur because precipitation from regular thunderstorms maintains high-fuel moisture (Di Pasquale et al., 2008). While the modern climate is too wet for natural ignition of fire, fires (possibly natural) have occurred in the Andean páramo since the Pleistocene–Holocene transition 13,000 cal. yr BP as indicated by charcoal fragments preserved in the sediment record (Hansen and Rodbell, 1995). While the preserved fire record could indicate human activity, both natural and anthropogenic-sourced fires are suppressed during wet climate (Bush et al., 2015).

Potential human influence

Previous work suggests that humans lived in the high elevations of the tropical Andes during the mid-Holocene (Bush et al., 2015, 2016; Hansen and Rodbell, 1995; Hansen et al., 2003; Valdez, 2008; Valencia et al., 2010; Weng et al., 2006). Deforestation for fuel and farmland, burning of grasslands for soil regeneration, livestock grazing, and terracing each alter the natural vegetation, thereby modifying the preserved vegetation and fire history (Bush et al., 2017; Sarmiento, 2013; Sylvester et al., 2014). For example, unaltered pristine landscapes in the Peruvian Andes had different paleovegetation than anthropogenically altered land (Sylvester et al., 2017). Anthropogenically sourced pollen such as maize and quinoa have been used in paleoclimate records to indicate human inhabitation on the landscape. Zea mays (maize) pollen was found in the pollen record of Lagunas Chorreras, Ecuador, and Laguna Comperta, Peru, around 4000 and 5500 cal. yr BP, respectively, indicating human activity (Hansen et al., 2003; Weng et al., 2006). Chenopodiaceae (quinoa) pollen was recovered from Lake Pacucha, Peru, suggesting human presence 5500 cal. yr BP, 2500 years before maize pollen was found on the site (Valencia et al., 2010). The combined presence of Chenopodiaceae pollen and lithic remains in Laguna Baja, Peru, indicates human activity around the lake for the last 4000 cal. yr BP. The fire record of the lake 4000 cal. yr BP has, therefore, been interpreted as a record of anthropogenic activity, rather than climate (Hansen and Rodbell, 1995).

Site description

Tres Lagunas is a series of three lakes located on the eastern ridge of the western cordillera of the Ecuadorian Andes (3.05145°S, 79.24822°W) approximately 30 km southwest of the city of Cuenca (inset Figure 1). The lakes sit on a volcanic plateau at an elevation of 3800 m a.s.l. The region is underlain by late Miocene to Pleistocene ash-flow tuffs, breccias, pyroclastic flows, and ignimbrites (Hungerbühler et al., 2002). Glaciation of this volcanic plateau during the late Pleistocene (~126,000–11,800 BP) left geomorphic features such as U-shaped valleys, lakes, and moraines on this now-dormant volcanic terrain (Goodman, 1996). Tres Lagunas is situated on a high plateau of mild relief surrounded by low undulating valleys and plains.

Previous work

The sedimentation hiatus in the Tres Lagunas peat bog core record spanned most of the mid-Holocene from 7600 to 2200 cal. yr BP (Jantz and Behling, 2012). Explanations for the low sedimentation rate range from a decreased rate of peat formation, complete cessation of peat formation, to cessation of peat formation and peat decomposition caused by a combination of drying and warming of the bog. In order to refine the paleoclimate record at Tres Lagunas, Ecuador, a lake core was collected in the summer of 2014. The much greater water depth of the lake (8.7 m) relative to the bog (approximately 0.86 km northeast, Figure 1) increases the likelihood of a complete sediment record in the lake. Furthermore, off-shore locations de-emphasize local terrestrial inputs and integrate fine particles (e.g. pollen and charcoal fragments) from a larger catchment relative to terrestrial locations. If the hiatus is present in the lake sediment record of Tres Lagunas, then the warm/dry event that likely caused the bog to dry was severe enough to cause the lake to dry and cease deposition. In this way, the lake core, through analysis of pollen, charcoal, magnetic susceptibility, and deposition, helps to determine both the presence and severity of a warm/dry event in the mid-Holocene at Tres Lagunas, Ecuador.

Methods

On 7 August 2014, 1.40 m of sediment (QU1-14B) was collected from 8.7 m below the lake surface of lake B of Tres Lagunas, Ecuador (Figure 1). Because of time constraints of field work, the morphometry of the site’s basin was not mapped. The core was collected from a coring platform made of two rafts tethered to shore. Sampling occurred in the southern bay of lake B, Tres Lagunas, because the coring equipment (15 m of coring rods), which was hand carried to the remote sampling location, was not sufficient for coring the center of the lake (>20 m depth; Figure 1). The top 0.25 m was collected via gravity corer (QU1-14B-GC). The subsequent 1.20 m (from 0.20 to 1.40 m depth) was collected via two drives of a Livingstone corer (QU1-14B-LC). During sampling, the Livingstone corer was unable to penetrate past 1.40 m because of the presence of a stiff layer of sediment. The core was extruded and placed on plastic wrap overlying aluminum foil on which it was subsectioned (1-cm increments) and placed in Whirl-Pak bags that were stored in coolers with ice until refrigerated (2°C) at The University of Utah Records of Environmental and Disturbance (RED) laboratory prior to analysis. The sediment was homogeneous, dark brown, organic-rich material with no visual banding or changes in composition.

Chronology and lithology

An age–depth relationship for Tres Lagunas was determined using six AMS ¹⁴C radiocarbon dates on pollen concentrates from the lake core (Table 1). Two radiocarbon dates were measured at 21–22 cm from both QU1-14B-LC and QU1-14B-GC to ensure agreement between the two records. Dates were measured at the Center for Applied Isotope Studies, University of Georgia. The top of the core was assumed to represent modern day (2014 CE). Radiocarbon dates were converted to calibrated years before present (cal. yr BP) using Calib Rev7.1.0 (Stuiver et al., 2016). The age–depth relationship between calibrated years was modeled using the Southern Hemisphere (SH) terrestrial curve in the Bayesian age–depth modeling approach BACON (Blaauw and Christen, 2011). Deposition rates were calculated from the difference in depth over the difference in modeled age.

Table 1.

Radiocarbon dates, calibrated age, and 2σ age range^a, δ¹³C, and sediment accumulation rate are given.

Depth (cm)	UGAMS number	Core	¹⁴C age (yr BP)	Calibrated age^a (cal. yr BP)	δ¹³C
21–22	A22898	QU1-14B-GC	920 ± 32	901 ± 28	−23.5
21–22	A22589	QU1-14B-LC	1063 ± 30	967 ± 40	−23.2
59–60	A25283	QU1-14B-LC	2570 ± 25	2733 ± 55	−24.7
94–95	A22899	QU1-14B-LC	3786 ± 31	4172 ± 30	−25.4
104–105	A25284	QU1-14B-LC	4428 ± 32	5041 ± 30	−26.3
136–137	A22590	QU1-14B-LC	5717 ± 26	6501 ± 66	−26.1

Calib Rev7.1.0 (Stuiver et al., 2016).

Pollen analysis

Pollen analysis was conducted at 6- to 8-cm intervals (representing 140–500 years, as described in the ‘Results’ section). Within each interval, a 1 cm³ sample was processed with standard acid–base reduction methods (Faegri et al., 1989). A total of 0.55 mL of microsphere capsules (at concentration of 5 × 10⁴ microspheres/mL) were added to each sample as an exotic tracer during processing for calculation of pollen influx rates. A Lycopodium tracer could not be used as Lycopodium is known to exist in high elevations of Ecuador (Niemann and Behling, 2008). A minimum of 300 terrestrial grains were counted per sample using light microscopy at 500× magnification. Terrestrial pollen counts were converted into pollen percentages associated with páramo, subpáramo, upper montane, and lower montane regions. Grouping of pollen taxa into altitudinal zones (páramo, subpáramo, UMF, and LMF) described by Jantz and Behling (2012) was used to allow direct comparison of altitudinal zone shifts. The only exception is that Oreopanax was grouped with UMF and not subpáramo (Borchsenius, 1997). Pollen identification was completed via combination of The University of Utah RED laboratory reference collection as well as the PalDat Palynological online database (PalDat Palynological Database, 2016). Spores and aquatic grains were also counted but not included in the pollen sums or percentages.

Pollen influx (grains/cm²/yr) was calculated as the product of extrapolated number of grains counted in the core interval and the corresponding deposition rate for that interval. The number of grains counted was extrapolated via division of 100 by the percentage of sample counted. Pollen influx was deposition rate normalized to allow comparison of pollen taxa flux across different times in the core (Faegri et al., 1989). Preservation ratios were determined by the ratio of the difference to the sum of number of grains counted and microspheres:

P r e s e r v a t i o n r a t i o = \frac{N o . o f i d e n t i f i e d g r a i n s - N o . o f m i c r o s p h e r e s}{N o . o f i d e n t i f i e d g r a i n s + N o . o f m i c r o s p h e r e s}

(1)

Large preservation ratios indicate well-preserved pollen grains (Brunelle et al., in revision). Pollen zones were identified using statistical methods in CONISS (Grimm, 1987) to distinguish shifts in altitudinal zones. CONISS performs stratigraphically constrained incremental sum of squares cluster analysis on altitudinal pollen percentages allowing for the grouping of similar pollen percentages. The trimming of the dendrogram output by CONISS into pollen zones was performed at inflection points with a total sum of squares less than 0.4.

Charcoal and burned grass

Charcoal analysis was conducted at contiguous 1-cm intervals (representing approximately 23.5 years, as described in the ‘Results’ section). Within each interval, 5 cm³ of sample was suspended in sodium hexametaphosphate for 24 h in Whirl-Pak bags. The suspension was sieved (125- and 250-µm wire mesh) onto petri dishes for counting via microscopy. Charcoal fragments >125 µm were taken to indicate local fires (within the local catchment; Clark, 1988; Gardner and Whitlock, 2001). Total charcoal fragment counts were converted to accumulation rates (influx) via deposition rate-normalized concentrations or influx (particles/cm²/yr). For each charcoal sample, partially burnt grass was also identified based on morphology and counted.

Magnetic susceptibility

Magnetic susceptibility was performed at contiguous 1-cm intervals (representing approximately 23.5 years, as described in the ‘Results’ section). Blank measurements were run before and after samples. Within each interval, 8 cm³ of sample was packed into a plastic cube and analyzed using a magnetic susceptibility sensor (Bartington, MS2B 36-mm I.D. cup sample). All samples were analyzed in a single run to avoid influences from changes in temperature and barometric pressure. CONISS sum of squares cluster analysis was stratigraphically run to group time periods of similar charcoal influx, burned grass influx, and magnetic susceptibility measurements. CONISS sum of squares cluster analysis was performed separately for pollen to detect changes in temperature, while cluster analysis of charcoal, burned grass, and magnetic susceptibility was used to detect changes in moisture.

Results

Chronology

Six calibrated radiocarbon dates were used to create a chronology for the Tres Lagunas lake core (QU1-14B; Table 1, Figure 3). The 140-cm core represents the past 6733 cal. yr BP, and this age–date relationship is well described by a Bayesian curve (BACON version 2.2) using the SH post-bomb curve SH zone 3 (Reimer et al., 2004) and a SH calibration curve SHCal13 (Blaauw and Christen, 2011; Figure 3, dashed red lines). While QU1-14B represents the last 6500 cal. yr BP, this date almost certainly does not represent the age of the lake. The two radiocarbon dates measured at 21–22 cm from both QU1-14B-LC and QU1-14B-GC showed agreement of 901 ± 28 and 967 ± 40 cal. yr BP, for QU1-14B-LC and QU1-14B-GC, respectively. Using the BACON model, a near-constant sedimentation rate of ~2 cm/100 years was determined (average of 2.18 cm/100 yr with standard deviation of 0.45 cm/100 yr). The hiatus determined in the bog record (Figure 3, black line; Jantz and Behling, 2012) was absent in the lake record. The bulk density was consistent throughout the core (average 0.198 g/cm³, standard deviation of 0.017 g/cm³ from samples collected from top, middle, and bottom of the core).

Figure 3.

Age–depth relationship comparison between Jantz and Behling (2012) bog core (black points) and QU1-14B lake core (red points). Jantz and Behling (2012) age–depth relationship was determined from six radiocarbon dates. Linear interpolation was performed by Jantz and Behling (2012) and recreated in CLAM version 2.2 (Blaauw, 2010; Reimer et al., 2013). The age–depth relationship for QU1-14 was determined from five radiocarbon dates. The red dashed line represents Bayesian interpolation performed in BACON. The red circles represent the calibrated ages for QU1-14B (Table 1). The overlain gray box represents a proposed hiatus (Jantz and Behling, 2012) occurring between 2200 ± 66 and 7600 ± 66 cal. yr BP.

Pollen zones

A total of five pollen zones were identified from CONISS-determined shifts in altitudinal zone (Figure 4) indicated by páramo, subpáramo, UMF, and LMF vegetation types. No superpáramo vegetation was present at Tres Lagunas. Páramo pollen taxa dominated the entire record with a minimum of 60% of terrestrial grains classified as páramo taxa throughout the record (mean 70% páramo taxa). Of the páramo taxa, Poaceae dominates the entire record accounting for 23–50% of the total terrestrial grains counted in each section. Despite the lake being situated near the subpáramo vegetation belt, subpáramo taxa represented less than 5% of the pollen taxa found throughout the record. The low frequency of subpáramo taxa also occurred in the bog core (Jantz and Behling, 2012).

Figure 4.

Pollen diagrams for QU1-14B (Quimsacocha volcanic basin, Tres Lagunas, southern Ecuador, 140 cm). (a) Pollen influx was calculated for each altitudinal zone. Pollen zones were generated in CONISS using altitudinal zone pollen sums. The age scale is based on calibrated years before present (cal. yr BP). (b) Individual pollen taxa distributions were also graphed with CONISS total sum of squares.

Pollen zone 1 (136–101 cm: 6500–4200 cal. yr BP)

Pollen zone 1, approximately 6500–4200 cal. yr BP, began with a high influx of páramo, subpáramo, UMF, and LMF pollen around 6500 cal. yr BP (6370 páramo, 200 subpáramo, 1530 UMF, and 800 LMF grains/cm²/yr; Figure 4a). Around 5600 cal. yr BP, the lowest influxes of páramo, subpáramo, LMF, and UMF pollen (3480, 0, 506, and 101 grains/cm²/yr, respectively) in pollen zone 1 were associated with decreased Poaceae (páramo), Gunnera (subpáramo), Weinmannia (UMF), and Moraceae/Urticaceae (LMF; Figure 4). The period from 5600 to 4700 cal. yr BP was associated with consistently increasing UMF (primarily Alnus and Podocarpaceae) and LMF (primarily Solanum) pollen taxa. In general, pollen zone 1 was characterized by relatively low pollen influx for páramo, subpáramo, UMF, and LMF vegetation types (mean 5368, 82, 701, and 336 grains/cm²/yr, respectively), but a high influx of Isoëtes (aquatic, mean 12,009 grains/cm²/yr; Figure 4a).

Pollen zone 2 (101–61 cm: 4200–2700 cal. yr BP)

Pollen zone 2 ranges from approximately 4200 to 2700 cal. yr BP. At approximately 4200 cal. yr BP, there is an influx of subpáramo vegetation (primarily Gunnera) and LMF pollen taxa (primarily Acalypha and Solanum; Figure 4). However, the influxes of subpáramo and LMF vegetation only represent 5% and 9% of the total pollen sum. Pollen zone 2 can be characterized as a period of low páramo pollen influx because of the decrease in Apiaceae, Asteraceae, and Ericaceae (Figure 4). Botryococcus influx was greatest during this zone with 32,800 and 46,000 grains/cm²/yr deposited in 3400 and 2800 cal. yr BP.

Pollen zone 3 (61–44 cm: 2700–2000 cal. yr BP)

Pollen zone 3, ranging 2700–2000 cal. yr BP, represents the end of the hiatus proposed by Jantz and Behling (2012). The highest UMF and LMF influx occurs at approximately 2250 cal. yr BP due mostly to co-occurring peaks of Alnus (UMF), Myrtaceae (UMF), and Moraceae/Urticaceae (LMF) and presence of Solanum (LMF) and Celtis (LMF; Figure 4). The highest páramo influx occurred 2100 cal. yr BP (primarily Gentianaceae and Geranium). In general, pollen zone 3 is marked by the high pollen influx for páramo, UMF, and LMF (peaks of 17,896, 2854, and 4281 grains/cm²/yr, respectively; Figure 4a).

Pollen zone 4 (44–36 cm: 2000–1000 cal. yr BP)

Pollen zone 4 initiates with decreased páramo, UMF, and LMF pollen taxa influx (8650, 553, and 0 grains/cm²/yr, respectively). No subpáramo pollen was counted during pollen zone 2. The influx of LMF pollen taxa increased from 2000 to 1000 cal. yr BP from 0 to 915 grains/cm²/yr (primarily Solanum; Figure 4). The influxes of páramo and UMF pollen taxa were elevated around 1300 cal. yr BP because of increased Poaceae (páramo) and Alnus (UMF; Figure 4).

Pollen zone 5 (23–7 cm: 1000–250 cal. yr BP)

Pollen zone 5, representing 1000–250 cal. yr BP, began with low influxes of all pollen taxa (5300, 0, 830, and 0 grains/cm²/yr for páramo, subpáramo, UMF, and LMF pollen taxa). Influxes of páramo and UMF pollen taxa increased at approximately 690 cal. yr BP because of Plantago (páramo), Gentianaceae (páramo), and Alnus (UMF), while influxes of subpáramo and LMF taxa remained absent. The most modern sample, representing 315 cal. yr BP, indicated a large influx of subpáramo pollen (primarily Gunnera; Figure 4a), likely reflecting the presence of the lake in the modern subpáramo vegetation belt.

Charcoal influxes, burned grass influxes, and magnetic susceptibility

In total, three zones were identified from CONISS-determined shifts in charcoal fragment influx, burned grass influx, and magnetic susceptibility (Figure 5). Charcoal fragment influx, burned grass, and magnetic susceptibility each showed statistically significant changes (p-value < 0.05, Student’s t-test) around 800 and 3900 cal. yr BP corresponding to the CONISS-determined zones (Figure 5). Zone 1, representing 6500–3900 cal. yr BP, recorded a period of low charcoal and burned grass influx (mean of 0.496 and 0.022 particles/cm²/yr, respectively). The greatest peak in magnetic susceptibility (45.8 SI) occurred around 5800 cal. yr BP during zone 1 (6500–3900 cal. yr BP), with the highest mean magnetic susceptibility (13.21 SI; Figure 5). Magnetic susceptibility showed a general decrease in peak height from 6500 cal. yr BP to the present.

Figure 5.

Charcoal (particles/cm²/yr), burned grass (particles/cm²/yr), and magnetic susceptibility (SI) were measured every 1 cm in QU1-14B lake core. Results indicating a wetter period (vertical blue line) and drier periods (vertical red lines) are shown. CONISS total sum of squares analysis was completed to detect three zones of similarity. Statistically significant changes in the mean concentration of each constituent, determined with Student’s t-test where p-value < 0.005, occurred between the three CONISS-delineated zones.

The mean charcoal and burned grass influxes were elevated during zone 2, 3900–800 cal. yr BP (0.645 particles/cm²/yr and 0.052 particles/cm²/yr, respectively), indicating a period of increased fire with the primary fuel being grass (Figure 5). While fire activity was present throughout zone 2, indicated by peaks in charcoal fragments, the height of the peaks decreased from 3900 to 800 cal. yr BP (Figure 5). This trend is not apparent for burned grass. Zone 3, 800 cal. yr BP to the present, represents a period of low charcoal influx and magnetic susceptibility. The largest influx of burned grass in the record (0.169 particles/cm²/yr) occurred during zone 3.

Discussion

Owing to the hiatus in sedimentation present in the bog core record during the mid-Holocene warm/dry event (Jantz and Behling, 2012), a lake core was collected and analyzed from lake B of Tres Lagunas, Ecuador. The vegetation and fire records were reconstructed using pollen and charcoal analysis in conjunction with magnetic susceptibility to extrapolate the moisture and temperature present during the last 6500 cal. yr BP. While the primary objective was to reconstruct the Holocene warm/dry event in Tres Lagunas, Ecuador, the secondary objective was to compare the record from QU1-14B with high-elevation lake cores from the tropical Andes in order to refine the timing of the mid-Holocene warm/dry event in the region.

Paleoclimate record present during the mid-Holocene warm/dry event

The mid-Holocene warm/dry event characterized by Jantz and Behling (2012) spanned a hiatus in the sedimentation record, where 11 cm represented ~5430 cal. yr BP. While the presence of a hiatus in sediment accumulation suggests the bog dried, additional evidence for a mid-Holocene warm/dry event in the bog record includes (1) increases in pollen taxa associated with a warm climate and (2) increased charcoal fragments in the 11 cm, which together indicate a dry period with frequent fires (Jantz and Behling, 2012). In the bog core vegetation record, high values of Poaceae and presence of Senecio type, as well as Baccharis and Alnus taxa (dry indicators), coincide with low levels of Plantago (wet indicator; Jantz and Behling, 2012). Because the data indicate drying of the bog, the resulting hiatus likely overestimates the duration of the warm/dry period since charcoal and pollen concentrations may be inflated by the low sedimentation rate. A continuous record from lacustrine sediments better constrains the warm/dry event for this area. In contrast to the bog core (Jantz and Behling, 2012), the lake core record reported here indicates that Tres Lagunas lake B did not dry in the last 6500 years, as shown by (1) its near-constant sedimentation rate (Figure 3), (2) consistent pollen preservation ratio (ranging 0.8–0.95; Supplemental Information, available online), and (3) the significant contribution of aquatic spores in the total pollen sum for the entire length of the core (mean Isoëtes and Botryococcus contribute 27% and 38% of total pollen sum, respectively; Figure 4a).

In the absence of human influence, charcoal fragments and pollen reflect paleoclimate. The fire record (charcoal fragments) can best describe moisture (dry/wet), whereas the vegetation record (pollen) can be used as a proxy for temperature shifts (warm/cold). The fire record can only indicate dry/wet climate when sufficient fuel is present, with increased fires indicating drier climates (Bowman et al., 2009). Sufficient fuel, in the form of vegetation, was present throughout the record (total pollen influx ranging from 4400 to 25,000 grains/cm²/yr). Not only was vegetation present throughout the record, but the vegetation of the páramo is sensitive to temperature change, allowing shifts in vegetation-defined altitudinal zones to act as a paleothermometer (Colinvaux et al., 1997; Paduano et al., 2003). As long-term temperature changes (climate), these temperature-sensitive species reform and change their centers of distributions, allowing for migration of the plant type over time (Colinvaux et al., 1997). The vegetation record QU1-14B suggests a warm period from 2700 to 2200 cal. yr BP with a less-pronounced warm period between 1700 to 1000 cal. yr BP (Figure 6). The charcoal, burned grass, and magnetic susceptibility suggest a dry period from 3900 to 800 cal. yr BP (Figure 6).

Figure 6.

Shifts in temperature (warm/cold) in the lake record (QU1-14B) were inferred from the vegetation record where low-elevation pollen types were associated with warm climates. Shifts in moisture (wet/dry) in the lake record were determined from the fire record, where increased charcoal fragments indicated increased fire activity. The timing of the warm/dry periods in the bog core (Jantz and Behling, 2012) is also shown. The lake record was ~1500 cal. yr shorter than the bog core record (shown in black).

The dry period from 3900 to 800 cal. yr BP denoted by elevated charcoal influx, burned grass influx, and magnetic susceptibility suggests increased topsoil burning and erosion (Gedye et al., 2000). Notably, the number of El Niño events/100 years indicated by inorganic clastic laminae deposition in Laguna Pallcacocha, Ecuador, was statistically less from 6500 to 2200 cal. yr BP and then from 2200 cal. yr BP to present day (mean 4.81 events/100 years and mean 10.04 events, respectively), corroborating the dry period with less El Niño-sourced rain (Moy et al., 2002; Supplemental Information, available online). The decrease in charcoal influx from 3900 to 800 cal. yr BP (Figure 5) could be caused by active El Niño years. Numerous El Niño events were recorded around 1600, 1300, and 800 cal. yr BP (26, 31, and 25 events/100 years, respectively; Moy et al., 2002; Supplemental Information, available online). Additionally, the period of high fire variability between 3900 and 800 cal. yr BP where numerous fire episodes were recorded on the landscape might have impacted the ability of the vegetation record to represent climate.

It has been suggested that fires in the páramo likely prevent extension of the forest into the páramo. This is because tussock grasses of the páramo-type vegetation respond faster after burning than do the trees of the UMF (Di Pasquale et al., 2008). If UMF/LMF taxa were prevented from extending up-elevation, then the 3900 to 800 cal. yr BP dry/fire record might not be accompanied by a warm vegetation signature. While peaks in LMF pollen percentage around 4000, 2200, and 1000 cal. yr BP are present, the signature may be muted from co-occurring fire activity on the landscape (Figure 4).

While increased fire during the dry period from 3900 to 800 cal. yr BP likely retarded up-elevation growth of warm climate vegetation (UMF and LMF vegetation), distinct warm periods appear in the vegetation record from 2700 to 2200 cal. yr BP and from 1700 to 1000 cal. yr BP (pollen zone 3 and pollen zone 2, Figures 4 and 6). The peak in LMF and UMF pollen type around 2200 cal. yr BP represents the period with the highest pollen influx for páramo, UMF, and LMF pollen (19,741, 3512, and 423 grains/cm²/yr, respectively) and co-occurs with the end of the hiatus reported in the bog record (Jantz and Behling, 2012; Figure 4). While the altitudinal zone shift supports a transition period around 2200 cal. yr BP, individual pollen taxa are not as supportive. Unlike the bog core record, Baccharis- and Senecio-type pollen are absent from the lake record from 2700 to 2000 cal. yr BP The peak in LMF pollen type at 2200 cal. yr BP. is dominated by Moraceae/Urticaceae (21 grains) and Solanum (10 grains; Figure 4). The high influx of Isoëtes around 2200 cal. yr BP (Figure 4; 11,658 grain/cm²/yr) could further indicate a dry period with a low lake level. Previous studies cite an inverse relationship between Isoëtes spores and rain, wherein lower levels of Isoëtes spores are associated with increased precipitation and higher lake levels (Flantua et al., 2016).

Around 2000 cal. yr BP, the influxes of LMF and UMF decrease with an increase in subpáramo pollen (primarily Oreopanax). The shift to a colder and moister environment was also suggested by Jantz and Behling (2012). The less-pronounced warm period from 1700 to 1000 cal. yr BP was represented by a peak in páramo and UMF pollen type around 1330 cal. yr BP. This period was followed by a peak in LMF pollen type around 1060 cal. yr BP primarily due to increases in Cecropia, Acalypha, and Solanum; Figure 4). A peak in charcoal and burned grass around 1100 cal. yr BP suggests a fire where Poaceae was burned potentially muting the páramo signature. The bog core also suggested a drier climate from 1600 to 800 cal. yr BP, where Eryngium-, Alnus-, and Huperzia-type pollen increased (Jantz and Behling, 2012). This period was bracketed by considerably moister and colder environment where high values of Poaceae and Isoëtes spores where measured (Jantz and Behling, 2012).

The lake core record and the bog core record from the Tres Lagunas region agree with each other on timing of warm and dry events. Both records suggest a shift occurring around 2200 cal. yr BP where the vegetation shifted from LMF vegetation type to páramo vegetation type indicating a cool climate. They also agree on a warm/dry period from ~1600 to 1000 cal. yr BP. Unfortunately, the lake core record (QU1-14B) only preserved the last 6500 cal. yr BP and did not cover the start of the event that caused the hiatus in the bog core record. However, because the deposition in the lake was near constant over the last 6500 cal. yr BP, and because the period from 6500 to 4000 cal. yr BP was characterized as temperate and wet, we propose that the hiatus-causing event in the Jantz and Behling (2012) bog record occurred prior to 7600 cal. yr BP.

Human influence

While increases in charcoal concentration can indicate a drier climate, human intervention on the landscape can also lead to increases in charcoal concentrations in the lake core record, where the landscape might have been anthropogenically burned to promote crop growth. Jantz and Behling (2012) suggest that humans have lived in Tres Lagunas for the last 8000 years based on human artifacts found in nearby caves (Valdez, 2008). Pollen can be used to denote the presence of humans and agriculture, where nonnative crop–pollen types such as Zea mays (maize) are found. Zea mays was found in the pollen record of Lagunas Chorreras dating back 4000 yr BP, again only 35 km from Tres Lagunas (Hansen et al., 2003). In Laguna Baja, the authors suggest that fire activity for the last 4000 yr BP was predominantly controlled by anthropogenic activity, rather than climate (Hansen and Rodbell, 1995). While Zea mays has not been found in the lake or bog core record from Tres Lagunas, the possibility that humans burned the landscape for agriculture or pasturing cannot be overlooked. Because of potential human influence on the fire record, time periods where both the vegetation and fire record corroborate a warm and a dry period (from 2700 to 1000 cal. yr BP) more definitively support climate shifts rather than human activity.

Regional comparison of paleoclimate records

The paleoclimate of the Holocene has been reconstructed from several high-elevation cores from both the eastern and western ridges of the tropical South American Andes (Abbott et al., 2003; Bush et al., 2005; Colinvaux et al., 1997; Hansen and Rodbell, 1995; Hansen et al., 2003; Hillyer et al., 2009; Jantz and Behling, 2012; Ledru et al., 2013; Niemann and Behling, 2007; Niemann et al., 2009; Paduano et al., 2003; Polissar, 2005; Rull et al., 2010; Weng et al., 2006; Wille et al., 2000). These records suggest that the mid-Holocene warm/dry event had an earlier onset and was more intense in the southern tropical Andes. The southern-most lakes analyzed in a north–south trending transect of seven lake cores from the eastern cordillera of southern Peru and Bolivia recorded the greatest duration and intensity of an extreme arid event from 7000 to 6000 cal. yr BP based on magnetic susceptibility and ratios of carbon to nitrogen (Abbott et al., 2003; Figure 1). Paleoclimate records from the eastern cordillera of the Andes in Lake Titicaca, Peru/Bolivia (3810 m a.s.l., Figure 1) suggest a dry event from 9000 to 3100 cal. yr BP, inferred from a shift in pollen type (Cyperaceae replaced by Poaceae, Polylepis, and Pediastrum) and an increase in charcoal concentration (Paduano et al., 2003). The dry event recorded in Lake Titicaca ended concurrently with the onset of a dry event in Tres Lagunas noted by an increase in charcoal concentration (Figure 5). A warm/dry period from 9500 to 5000 cal. yr BP at Laguna La Comperta, Peru (4600 m a.s.l., Figure 1), loosely agrees with the timing in Lake Titicaca but reports the warmest time occurring around 8500 cal. yr BP based on an increase in magnetic susceptibility (Weng et al., 2006).

In addition to a latitudinal trend, records from north of Tres Lagunas suggest an onset of the mid-Holocene warm/dry event that predates the record from Tres Lagunas. A warm period at Lagunas Chorreras, Ecuador (3700 m a.s.l., approximately 35 km north of the Tres Lagunas site, Figure 1), was interpreted between 10,000 and 7500 cal. yr BP, with shifts in moist montane UMF pollen suggesting warm temperatures and moderate seasonality (Hansen et al., 2003). An increase in fire activity (increased charcoal concentrations >100 µm and 30–100 µm) at Lagunas Chorreras from 9000 to 6000 cal. yr BP indicates a dry climate (Hansen et al., 2003).

Because the record from Tres Lagunas was 6500 cal. yr BP, it is not possible to determine whether there was a prior, longer, more intense warm/dry period that caused the bog at Tres Lagunas to dry. However, both records from Tres Lagunas, Ecuador, suggest a shorter warm/dry event more recently than other cores from the region. At Laguna Chorreras, Ecuador, the presence of Apiaceae and other taxa preferring moist habitats were attributed to a wet climate between 7500 and 4000 cal. yr BP (Hansen et al., 2003), agreeing with the record in QU1-14B and the bog core record (Jantz and Behling, 2012). However, low charcoal values and increases in moist UMF pollen taxa after 4000 cal. yr BP were interpreted as a wet/cold period at Lagunas Chorreras (Hansen et al., 2003) during a time of apparent warm/dry climate in Tres Lagunas. In both QU1-14B and the bog core record (Jantz and Behling, 2012), 4000 cal. yr BP recorded an increase in charcoal influx (Figure 5).

Conclusion

In order to reconstruct the paleoclimate of Tres Lagunas, Ecuador, a lake sediment core was collected from Tres Lagunas lake B. The charcoal and burned grass records suggest a dry landscape subject to frequent burning of grass from 3900 to 800 cal. yr BP. The vegetation record, likely impacted by frequent burning of the landscape, suggests that the warm periods occurred from 2700 to 2000 cal. yr BP denoted by up-elevations shifts in pollen type and again from 1700 to 1000 cal. yr BP. The vegetation record and previously collected bog core suggest an important shift occurred at 2200 cal. yr BP where vegetation type shifted down-elevation, indicating a potential end to the warm/dry climate of Tres Lagunas. However, since the potential for human influence on the landscape cannot be overlooked, time periods when the vegetation record agrees with the fire record for a warm/dry period more definitively support a climatic shift. Following this line of logic, a warm/dry period from 2700 to 1000 cal. yr BP occurred at Tres Lagunas, Ecuador. As anthropogenic alteration is likely at Tres Lagunas, a future core from a known un-anthropogenically altered landscape would be worthwhile.

Supplemental Material

Fredericketal_2017_SupportingInformation_20180103_LF – Supplemental material for Reconstruction of the mid-Holocene paleoclimate of the Ecuadorian Andean páramo at Tres Lagunas, Ecuador

Supplemental material, Fredericketal_2017_SupportingInformation_20180103_LF for Reconstruction of the mid-Holocene paleoclimate of the Ecuadorian Andean páramo at Tres Lagunas, Ecuador by Logan Frederick, Andrea Brunelle, Mathew Morrison, Patricio Crespo and William Johnson in The Holocene

Footnotes

Acknowledgements

We are grateful to our Ecuadorian collaborators from the University of Cuenca and Polytechnic Institute, especially Dr Ximena Diaz, for their help in collecting the lake core. A special thanks to the 2014 Learning Abroad class in helping collect the lake core.

Funding

Funding for this analysis was provided by The University of Utah Department of Geology and Geophysics Student Research Grant, Don Currey Research Grant, The University of Utah Global Change and Sustainability Center Research Grant, The University of Utah RED laboratory and The University of Utah Undergraduate Research Opportunities Program.

References

Abbott

Wolfe

et al . (2003) Holocene paleohydrology and glacial history of the central Andes using multiproxy lake sediment studies. Palaeogeography, Palaeoclimatology, Palaeoecology 194: 123–138.

Blaauw

(2010) Methods and code for ‘classical’ age-modeling of radiocarbon sequences. Quaternary Geochronology 5: 512–518.

Blaauw

Christen

(2011) Flexible paleoclimate age-depth models using autoregressive gamma process. Bayesian Analysis 6: 457–474.

Borchsenius

(1997) Oreopanax (Araliaceae) in Ecuador. Nordic Journal of Botany 17(4): 373–396.

Bowman

Balch

Artaxo

et al . (2009) Fire in the Earth System. Science 324: 481–484.

Brunelle

Eckerle

Petersen

et al . (in revision) A high-resolution Holocene paleoenvironmental history of Lake Bonneville/Great Salt Lake. Geological Society of America.

Bush

Sublette Mosblech

Church

(2015) Climate change and the agricultural history of a mid-elevation Andean montane forest. The Holocene 25(9): 1522–1532.

Bush

Correa-Metrio

McMichael

et al . (2016) A 6900-year history of landscape modification by humans in lowland Amazonia. Quaternary Science Reviews 141: 52–64.

Bush

Correa-Metrio

Van Woesik

et al . (2017) Human disturbance amplifies Amazonian El Niño-Southern Oscillation signal. Global Change Biology 23: 3181–3192.

10.

Bush

Hansen

Rodbell

et al . (2005) A 17000-year history of Andean climate and vegetation change from Laguna de Chochos, Peru. Journal of Quaternary Science 20: 703–714.

11.

Campozano

Célleri

Trachte

et al . (2016) Rainfall and cloud dynamics in the Andes: A Southern Ecuador case study. Advances in Meteorology 2016: 3192765.

12.

Clark

(1988) Particle motion and the theory of stratigraphic charcoal analysis: Source area, transportation, deposition, and sampling. Quaternary Research 30: 81–91.

13.

Clement

Seager

Cane

(2000) Suppression of El Nino during the mid-Holocene by changes in the Earth’s orbit. Paleoceanography 15: 731.

14.

Colinvaux

Bush

Steinitz-Kannan

et al . (1997) Glacial and postglacial pollen records from the Ecuadorian Andes and Amazon. Quaternary Research 48: 69–78.

15.

Conroy

Overpeck

Cole

et al . (2008) Holocene changes in eastern tropical Pacific climate inferred from a Galapagos lake sediment record. Quaternary Science Reviews 27: 1166–1180.

16.

Crespo

Feyen

Buytaert

(2011) Identifying controls of the hydrological response of small catchments in the tropical Andes (Ecuador). Journal of Hydrology 407: 164–174.

17.

Di Pasquale

Marziano

Impagliazzon

et al . (2008) The Holocene treeline in the northern Andes (Ecuador): First evidence from soil charcoal. Palaeogeography, Palaeoclimatology, Palaeoecology 259: 17–34.

18.

Faegri

Kaland

Kzywinski

(1989) Textbook of Pollen Analysis. New York: Wiley.

19.

Flantua

SGA

Hooghiemstra

Vuille

et al . (2016) Climate variability and human impact in South America during the last 2000 years: Synthesis and perspectives from pollen records. Climate of the Past 12: 483–523.

20.

Gardner

Whitlock

(2001) Charcoal accumulation following a recent fire in the Cascade Range, northwestern USA and its relevance for fire-history studies. The Holocene 11: 541–549.

21.

Gedye

Jones

Tinner

et al . (2000) The use of mineral magnetism in the reconstruction of fire history: A case study from Lago di Origlio, Swiss Alps. Palaeogeography, Palaeoclimatology, Palaeoecology 164: 101–110.

22.

Goodman

(1996) Glacial geology and soil catena development on Morains in Las Cajas National Park, Ecuador. Unpublished Bachelor Thesis, Union College.

23.

Grimm

(1987) CONISS: A FORTRAN 77 program for stratigraphically constrained cluster analysis by method of incremental sum of squares. Computers & Geosciences 13: 13–35.

24.

Hansen

Rodbell

(1995) A Late-Glacial/Holocene pollen record from the Eastern Andes of Northern Peru. Quaternary Research 44: 216–227.

25.

Hansen

Rodbell

Seltzer

et al . (2003) Late-Glacial and Holocene Vegetational history from two sites in the Western Cordillera of Southwestern Ecuador. Palaeogeography, Palaeoclimatology, Palaeoecology 194: 79–108.

26.

Hillyer

Valencia

Bush

et al . (2009) A 24,700-yr paleolimnological history from the Peruvian Andes. Quaternary Research 71: 71–82.

27.

Haug

Hughen

Sigman

et al . (2001) Southward migration of the intertropical convergence zone through the Holocene. Science 293(1): 1304–1308.

28.

Hungerbühler

Steinmann

Winkler

et al . (2002) Neogene stratigraphy and Andean geodynamics of southern Ecuador. Earth Science Reviews 57: 75–124.

29.

Jantz

Behling

(2012) A Holocene environmental record reflecting vegetation, climate, and fire variability at the Páramo of Quimsacocha, Southwestern Ecuadorian Andes. Vegetation History and Archaeobotany 21(3): 169–185.

30.

Ledru

Jomelli

Samaniego

et al . (2013) The Medieval Climate Anomaly and the Little Ice Age in the eastern Ecuadorian Andes. Climate of the Past 9: 307–321.

31.

Mitchell

Wallace

(1991) The annual cycle in equatorial convection and sea surface temperature. Journal of Climatology 5: 1140–1156.

32.

Moy

Seltzer

Rodbell

et al . (2002) Variation of El Niño/Southern Oscillation activity at millennial timescales during the Holocene epoch. Nature 420: 162–165.

33.

Niemann

Behling

(2007) Late Quaternary vegetation, climate and fire dynamics inferred from the El Tiro record in the southeastern Ecuadorian Andes. Journal of Quaternary Science 23(3): 203–212.

34.

Niemann

Behling

(2008) Past Vegetation and Fire Dynamics: Gradients in a Tropical Mountain Ecosystem of Ecuador. Berlin

35.

Niemann

Haberzettl

Behling

(2009) Holocene climate variability and vegetation dynamics inferred from the (11700 cal. yr BP) Laguna Rabadilla de Vaca sediment record, southeastern Ecuadorian Andes. The Holocene 19(2): 307–316.

36.

Ochoa-Tocachi

Buytaert

De Bièvre

et al . (2016) Impacts of land use on the hydrological response of tropical Andean catchments. Hydrological Process 30: 4074–4089.

37.

Paduano

Bush

Baker

et al . (2003) Vegetation and Fire History of Lake Titicaca since the Last Glacial Maximum. Palaeogeography, Palaeoclimatology, Palaeoecology 194: 259–279.

38.

PalDat Palynological Database (2016) PalDat Palynological Database – An Online Publication on Recent Pollen. Society for the Promotion of Palynological Research in Austria. Available at: www.paldat.org (accessed 10 January 2016).

39.

Podwojewski

Poulenard

Zambrana

et al . (2002) Overgrazing effects of vegetation cover and properties of volcanic ash soil in the páramo of Llangahua and La Esperanza (Tungurahua, Ecuador). Soil Use Management 18: 45–55.

40.

Polissar

(2005) Lake records of Holocene climate change, Cordillera de Merida, Venezuela. PhD Thesis, University of Massachusetts.

41.

Reimer

Bard

Bayliss

et al . (2013) IntCal13 and Marine13 radiocarbon age calibration curves, 0–50,000 years cal BP. Radiocarbon 55: 1869–1887.

42.

Reimer

Brown

Reimer

(2004) Discussion: reporting and calibration of post-bomb 14C data. Radiocarbon 46(3): 1299–1304.

43.

Rull

Abbott

Vegas-Vilarrubia

et al . (2010) Paleoenvironmental trends in Venezuela during the last glacial cycle. In: Sánchez-Villagra M, Aguilera O and Carlini A (eds) Urumaco and Venezuelan Paleontology – The Fossil Record of the Northern Neotropics. Bloomington, IN: Indiana University Press, pp. 52–83.

44.

Sarmiento

(2013) Contesting Páramo: Critical Biogeography of the Northern Andean Highlands. Charlotte, NC: Kona Publishers.

45.

Sklenar

Ramsay

(2001) Diversity of zonal páramo plant communities in Ecuador. Journal of Conservation Biogeography 7(3): 113–124.

46.

Stuiver

Reimer

(2016) CALIB 7.1. Available at: http://calib.org (accessed 20 December 2016).

47.

Sylvester

MDPV

Kessler

(2014) Inaccessible ledges as refuges for the natural vegetation of the high Andes. Journal of Vegetation Science 25: 1225–1234.

48.

Sylvester

Heitkamp

Sylvester

MDPV

et al . (2017) Relict high-Andean ecosystems challenge our concept of naturalness and human impact. Scientific Reports 7: 3334.

49.

Valencia

Urrego

Silman

et al . (2010) From Ice Age to Modern: A record of landscape change in the Andean cloud forest. Journal of Biogeography 37: 1637–1647.

50.

Valdez

(2008) Inter-zonal relationships in Ecuador. In: Silverman

Isbell

(eds) Handbook of South American Archeology. New York: Springer, pp. 865–888.

51.

Velez

Hooghiemstra

Metcalfe

et al . (2006) Late Glacial and Holocene environmental and climatic changes from a limnological transect through Colombia, northern South America. Palaeogeography, Palaeoclimatology, Palaeoecology 234: 81–96.

52.

Weng

Bush

Curtis

et al . (2006) Deglaciation and Holocene climate change in the western Peruvian Andes. Quaternary Research 66: 87–96.

53.

Wille

Negret

Hooghiemstra

(2000) Paleoenvironmental history of the Popayan area since 27000 yr BP at Timbio, Southern Colombia. Review of Palaeobotany and Palynology 109: 45–63.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.65 MB