Holocene climate variability and environmental history at the Patagonian forest/steppe ecotone: Lago Mosquito (42°29'37.89''S,71°24'14.57''W) and Laguna del Cóndor (42°20'47.22''S,71°17'07.62''W)

Abstract

Along the eastern Andes, a sharp ecotone separates steppe from North Patagonian forest dominated by Nothofagus spp. and Austrocedrus chilensis. The longitudinal position of the ecotone is largely determined by effective moisture, which in turn is partly governed by the strength and latitudinal position of the Southern Westerlies. As a result, changes in the ecotone provide an opportunity to examine past climate variations. Holocene environmental history at two sites in close proximity is inferred from pollen and high-resolution charcoal data. Prior to 9000 cal. yr BP, vegetation resembled a steppe, in accordance with widespread aridity. Fires were infrequent, likely as a consequence of fuel discontinuity associated with low vegetation cover. At 9000 cal. yr BP, forest taxa expanded into steppe and fires became frequent, indicating that summers were arid enough to support fires but winter moisture was sufficient for Nothofagus spp. to expand. A two-step increase in effective moisture is inferred for the middle Holocene. The first step occurred at 8500 cal. yr BP, as interpreted from the increase in A. chilensis in the region, probably as a consequence of an eastward migration from glacial refugia. The second step at 5500 cal. yr BP is based on a Nothofagus spp. expansion into the steppe. Steppe readvances into the forest between 5250 and 3000 cal. yr BP indicate decreased temperatures and/or effective moisture. The last 3000 years are characterized by expansions of A. chilensis and an eastward shift of the ecotone, suggesting more humid conditions. European settlement is reflected in the establishment of non-native species and disturbance-adapted taxa.

Keywords

charcoal analysis fire history Northern Patagonia climate Southern Westerlies vegetation history

Introduction

Along the eastern Andes, a sharp ecotone separates low-elevation steppe from North Patagonian forest dominated by Nothofagus dombeyi and Austrocedrus chilensis. The geographic position of the ecotone is largely determined by effective moisture, which in turn is broadly governed by the strength and latitudinal position of the Southern Westerlies (SW), and the orographic effects of the Andes (Villalba et al., 2003). As a result, changes in the position, sharpness, and composition of the Patagonian forest-steppe ecotone provide an opportunity to examine past climate variations.

Previous work in the Andes (Bianchi and Ariztegui, 2011, this issue; Huber and Markgraf, 2003; Markgraf, 1984; Whitlock et al., 2006) has shown that climate-induced vegetation change occurs through the direct effects of climate on plant fitness, and indirectly through altered disturbance regimes, such as insect breakouts and fires. Fire is an intrinsic element of the forest-steppe border with large effects on vegetation composition and landscape structure (Veblen et al., 1999). In recent decades, a human-induced change in the fire regime has led to an increase in re-sprouting shrubs (i.e. Nothofagus antarctica, Discaria spp.) at expense of trees, and this shift has contributed to the observed shrinkage of the Nothofagus dombeyi forest (Mermoz et al., 2005). Evaluations of climate–fire–vegetation linkages are therefore critical for understanding ecosystem dynamics and local responses to regional climate change, such as those predicted to occur in the future (Intergovernmental Panel on Climate Change (IPCC), 2007).

In order to improve our understanding of forest-steppe ecotone dynamics, we reconstruct the vegetation, fire, and climate history based on pollen, charcoal, and lithologic records contained in sediment cores from Lago Mosquito (also known as Lago Pellegrini, lat. 42°29′37.89″S, long. 71°24'14.57''W, 556 m elevation, 8 m coring water depth, 461 ha surface area, surface runoff from Arroyo Mosquito) and Laguna del Cóndor (also known as Laguna Escondida, lat. 42°20'47.22''S, long. 71°17'07.62''W, 818 m elevation, 8.5 m coring water depth, c. 175 ha surface area, no incoming streams). These lakes are 18 km apart along a west–east precipitation transect and located within the transition from open Austrocedrus chilensis woodland to steppe, near the town of Cholila (Chubut Province) in western Patagonia (Figure 1). Their sediments preserve evidence of explosive Holocene eruptions of large stratovolcanoes from the southern portion of the Andean Southern Volcanic Zone (lat. 41.5°–46°S; Stern, 2004). Both lakes lie close to terminal Pleistocene moraines (Caldenius, 1932), but the origin of L. Mosquito is related to Holocene alluvial fans that dammed westward-flowing streams and created the lake upvalley (Whitlock et al., 2006). By analyzing two proximal sites, we expect to override the effects of chaotic (Bennett, 1993; Ives et al., 2008; May, 1976) and ecology-independent random processes (Hubbell, 2001) operating on the vegetation history at each site, and obtain a reliable reconstruction of past environmental changes in the area (Blaauw et al., 2010; Briles et al., 2008), increasing our understanding of the ecotonal history as well as its relationship to past climate changes.

Figure 1.

Location of L. Mosquito and L. Cóndor.

Modern setting

A latitudinal and altitudinal gradient dominates Patagonian temperatures and precipitation. In northern Patagonia, mean annual temperatures range from 12°C in the intermountain valleys, to 6°C in the subalpine deciduous forest near the treeline (Villalba et al., 2003). Precipitation in the area is related to frontal systems associated with migratory surface cyclones. Pacific cyclones migrate eastwards along the storm tracks, whose main position follows the jet stream (Garreaud et al., 2003). Prevailing strong southwesterlies are the main characteristic that delimits Patagonia as a uniform climatic region (Garreaud et al., 2008). A west wind component occurs at least 75% of the time along the entire Chilean coast (Miller, 1976), and 50–70% of the time in the eastern plains (Prohaska, 1976). The Andes constitute an effective barrier to tropospheric flow. The uplift of low-level winds over the western slope of the Andes produces continental orographic precipitation (Villalba et al., 2003), and forced subsidence over the Argentine flanks of the Andes causes an adiabatic warming of the air masses, resulting in dry conditions in eastern Patagonia (Paruelo et al., 1998). As a consequence, precipitation decreases from 4000–6000 mm/yr in Chile, to 700 mm/yr at San Carlos de Bariloche airport, located approximately 60 km east of the crest of the Andes (Villalba et al., 2003).

Seasonal and annual precipitation variability is influenced by changes in intensity and latitudinal position of the SW, which result in shifts in the storm tracks. In northern Patagonia, storm frequency is highest in winter, when 40% of the precipitation occurs (Miller, 1976). Above 1000 m elevation, winter precipitation is usually in the form of snow. In the foothills and the steppe, distance from the Andes explains over 90% of the spatial variability of mean annual precipitation, and interannual precipitation variability is greatest at the dry end of the humidity gradient (Jobbágy et al., 1995).

The latitudinal position of the SW is governed by the strength and position of the southeastern Pacific high-pressure cell and the subpolar low-pressure trough centered along the Antarctic Circle (Mayr et al., 2005). These circulation systems show latitudinal shifts related to seasonal changes in the temperature gradient between the equator and the poles. The pronounced summer pole-to-equator pressure gradient leads to strongest southwesterlies, whose core focuses at lat. 45°–50°S. In winter, the intensification of the subpolar low and the equatorward displacement of the southeastern Pacific high-pressure cell result in jet stream migration to lower latitudes (lat. 40°S; Paruelo et al., 1998), whereas the low-level component expands equatorward but weakens, particularly at lat. 50°S (Garreaud et al., 2008).

The steep west-to-east precipitation gradient in the eastern Andes is reflected in a vegetation transition from rainforests to xerophytic forests to steppe (Hajek and di Castri, 1975; Jobbágy et al., 1996) that occurs in <50 km. The montane slopes of the Andes are dominated by tall forests of the evergreen Nothofagus dombeyi. Further east, where annual precipitation declines to 1500 mm, Austrocedrus chilensis and N. dombeyi form extensive co-dominant stands. Under the more xeric conditions to the east, A. chilensis forms pure dense stands at its western limit, and open woodlands with abundant sclerophyllous shrubs and small trees (e.g. Lomatia hirsuta, Schinus patagonicus, Embothrium coccineum, Maytenus sp., and the deciduous Nothofagus antarctica) towards the east. Further east where precipitation is <500 mm, A. chilensis is replaced by steppe characterized by spiny shrubs (Adesmia spp., Mulinum sp., Berberis sp.) and bunchgrasses (Stipa spp. and Festuca spp.) (Seibert, 1982).

Fire is an intrinsic element of Patagonian ecosystems with large effects on vegetation composition and landscape structure (Veblen et al., 1999). Nothofagus forests, dominated by evergreen and broadleaved species that tend to act as biologic fire breaks, experience infrequent drought-induced stand-replacing fires (Kitzberger et al., 1997). In contrast, the fine fuels of the steppe desiccate quickly and are commonly dry enough to support frequent surface fires. Vegetation discontinuity limits fires in these dry areas, where dry years of fire occurrence are often preceded by wet years of fuel accumulation (Huber et al., 2004). Shrubs and bunch grasses are the dominant fuels in the ecotonal A. chilensis forests. Their fire regime is characterized by relatively frequent low-severity surface fires (Kitzberger et al., 1997; Veblen et al., 1999). Many decades of fire exclusion, livestock grazing, escarpment from non-native forestry plantations, however, have allowed woody fuels to accumulate sufficiently to increase the potential of crown fires (Veblen et al., 2008).

Methods

Sediment cores were collected from the center of the L. Mosquito and L. Cóndor basins with a modified Livingston piston sampler from a floating, anchored platform. Cores were extruded in the field and wrapped in cellophane and aluminum foil to protect them from contamination and oxidation, and shipped to Montana State University, where they were stored and refrigerated. In the laboratory, cores were split longitudinally into a working half and an archive. The working half was lithologically described, photographed, and analyzed for magnetic susceptibility (MS), sequential loss on ignition (LOI) and charcoal and pollen content. Inductively coupled plasma mass spectrometry (ICP-MS) analyses were performed to five tephra samples for characterization and stratigraphic correlation purposes. Their major oxide, trace and rare earth elements (REE) were determined.

Description of the lithology was based on identification of sedimentary structures, mineralogy, and biological components, following Schnurrenberger et al. (2003). MS was measured at 0.5 cm intervals with a spot-reading sensor (MS2E) directly on the split-surface of the core. The data were used as an approximation of sediment magnetic mineral concentration (Gedye et al., 2000), which provides information on allochthonous clastic sediment inputs from erosion and volcanic eruptions. LOI analysis was carried out on 1 cm³ samples taken at 2 cm intervals. Samples were dried at 90°C for 24 h, and then ignited at 550°C and 900°C. The weight loss between each step measures water, organic matter and carbonate content of the sediment, respectively (Dean, 1974).

Nine samples from the L. Cóndor LC06A, and 19 samples from the L. Mosquito Mos03A and Mos03C cores were submitted for AMS dating (Table 1). Chronologies were developed from modeling sediment age as a function of sediment depth. Two-sigma calibrated ages and probability distributions were determined for each radiocarbon date using CALIB 6.0.1 (Stuiver et al., 1993). Calibration was performed with the Southern Hemisphere radiocarbon calibration data set for samples <11,000 cal. yr BP (McCormac et al., 2004) and the IntCal09 Calibration Curve for samples >11,000 cal. yr BP (Reimer et al., 2009). Core depth was corrected for sediment compaction and adjusted by excluding volcanic ash layers >1.5 cm thick, which are assumed to have been rapidly deposited. Uncertainties in age determinations were considered in the age–depth model by using Monte Carlo sampling (1000 iterations) to generate cubic splines through the calibrated probability distributions of all the dates (Higuera et al., 2009). The importance of each age in the fitted spline was weighted based on its standard deviation, such that ages with larger associated errors had less influence in the models (Telford et al., 2005). The final age–depth models were based on the median of all the runs (Figure 2).

Table 1.

Radiocarbon and calibrated radiocarbon dates from Lago Mosquito and Laguna del Cóndor.

Core	Depth (cm)	Adjusted midpoint depth (cm)^a	Lab no.	Material	¹⁴C yr BP	Corrected ¹⁴C yr error	Median probability age (cal. yr BP)^b
Lago Mosquito
Mos03A	0–0.1	0.5	n/a	inferred	0		−53
Mos03A*	45–46	45.5	AA58860	sediment	645	33	606
Mos03A	105–106	105.5	AA58861	sediment	577	32	540
Mos03A	245–246	245.5	AA58862	sediment	1600	33	1444
Mos03A*	308–309	308.5	AA58874	charcoal	1610	150	1471
Mos03A	308–309	308.5	AA58863	sediment	1818	34	1670
Mos03A*	348.5–349.5	349	AA58875	charcoal	2300	160	2245
Mos03A	349–350	349.5	AA58864	sediment	2066	34	1961
Mos03A	420–421	420.5	AA58865	sediment	2421	34	2405
Mos03A	491–492	491.5	AA58876	charcoal	2699	76	2762
Mos03A*	491–492	491.5	AA58866	sediment	2859	35	2906
Mos03A	615–616	615.5	AA58867	sediment	3711	37	3978
Mos03A	740–741	740.5	AA59431	sediment	4495	40	5051
Mos03A	825–826	825.5	AA58868	sediment	4629	39	5237
Mos03A	925–926	925.5	AA58869	sediment	5038	39	5710
Mos03A	1025–1026	1025.5	AA58870	sediment	5714	47	6441
Mos03A	1125–1126	1125.5	AA58871	sediment	6499	43	7362
Mos03A	1208–1209	1208.5	AA58872	sediment	6892	45	7668
Mos03A	1308–1309	1308.5	AA58873	sediment	7218	51	7983
Mos03C	1482	1482	AA58903	wood	8200	47	9091
Laguna del Cóndor
LC06A	0–0.5	0.25	n/a	inferred	0		−56
LC06A*	24.5–25	24.75	AA81902	charcoal	6195	49	7040
LC06A	40–40.5	40.25	AA84187	sediment	3057	38	3198
LC06A	102–102.5	102.25	AA81900	charcoal	4819	55	5501
LC06A	153.5–154	153.75	AA81904	charcoal	5993	55	6767
LC06A*	203.5–204	203.75	AA81901	charcoal	6818	64	7618
LC06A	243–243.5	224.75	AA81905	charcoal	8601	55	9524
LC06A	265.5–266	247.25	AA81467	sediment	8988	52	10,046
LC06A*	278.5–279	255.25	AA70029	sediment	14,669	79	17,845

Adjusted depths were used to calculate the age–depth model. Only true depths are referred to in text.

Calibrated ages were based on CALIB 6.0 (Stuiver et al., 2005; http://radiocarbon.pa.qub.ac.uk/calib/calib.html).

Not included in the chronologies.

Figure 2.

Age–depth model for (a) L. Mosquito, and (b) L. Cóndor. 95% confidence intervals are shown in gray. Black squares are calibrated dates used to develop the models and gray squares are dates that were not included in the models. See Table 1 for more information.

We developed seven alternative age–depth models for L. Mosquito by including all dates, and then excluding individual dates from three levels that had been dated twice. All the models passed through the 95% confidence interval of the ¹⁴C dates, and in no case did the overall nature of the age–depth relationship change. Given that the lithology of the Mos03A core did not suggest changes in sedimentation, for this study we consider the model that yielded least abrupt changes in sedimentation rates.

Three ¹⁴C dates from the LC06A core were not included in the age–depth model. The oldest date (17,500 cal. yr BP) corresponded to a peat layer at 255.5 cm depth, and overlying gyttja yielded an age of 10,046 cal. yr BP at 247.25 cm depth. This discrepancy suggests a hiatus of approximately 7200 years in sedimentation between the deposition of the peat layer and the creation of the lake. For this reason, the date obtained from the peat layer was not used in the chronology and the overlying date was extrapolated to the gyttja-peat contact. A date at 24.5 cm depth (7040 cal. yr BP) was older than ages defined by other three ¹⁴C dates obtained from deeper levels and excluded from the model, and so was one at 203.75 cm (7620 cal. yr BP) whose inclusion in the model would have resulted in a dramatic change in sedimentation rate that was not reflected in the lithology. The resulting chronology was well constrained.

Charcoal analysis was performed on 2 cm³ volume samples, taken at contiguous 0.5 cm intervals, following the methodology outlined by Whitlock and Larsen (2001) to reconstruct local fire histories. The material was wet-screened through a 125 mm mesh sieve and charcoal particles were tallied under a stereomicroscope. Grass and wood charcoal were tallied separately. Charcoal counts were converted to charcoal concentration (particles/cm³), and then to charcoal accumulation rates (CHAR; particles/cm² per yr). In order to account for variable sampling intensity with time and override the effects of dynamic sedimentation rates, the CHAR time series was interpolated to its median resolution. A low frequency component of the series (background CHAR) was defined by smoothing the record with a locally weighted regression (Higuera et al., 2009). The L. Mosquito record was smoothed with a moving average. Because of the presence of outliers, a moving median was chosen for the L. Cóndor CHAR series. Background CHAR integrates charcoal contributions from across the landscape (Long et al., 1998) and reflects area burned (Higuera et al., 2010). The positive residuals of the model include fire episodes (one or more local fires during the time span of the charcoal peak) (Whitlock and Anderson, 2003) as well as noise. Local thresholds were used to identify significant fire-related peaks from noise (Higuera et al., 2005). Fire frequency was calculated as fire events/time and smoothed over a time window of 1000 years. Grass-to-total-charcoal ratios were used to infer the relative contribution of grass and woody fuels, allowing provisional classification of fires into surface fires (grass-to-total-charcoal ratio>0.5) and crown fires (grass-to-total-charcoal ratio <0.5). Chusquea spp. seem unlikely or minor contributors to the grass charcoal, given that they do not grow in the Mosquito-Cóndor area at present.

For pollen analysis, 1 cm³ of sediment was taken at 8 cm intervals and prepared with standard techniques (Faegri and Iversen, 1989). A known amount of Lycopodium tracer spores was added to each sample to allow calculation of pollen accumulation rate (PAR; grains/cm² per yr), which was interpreted as a crude measure of plant abundance. Pollen grains were identified at 400× magnification, using a reference collection and published atlases (Heusser, 1971; Markgraf and D’Antoni, 1978). In all cases, counts exceeded 300 terrestrial grains, excluding Cyperaceae, aquatic taxa and spores. Terrestrial pollen percentages were based on the sum of trees, shrubs and herbs and plotted using C2 (Juggins, 2007).

The pollen diagram was grouped according to modern affinities of the probable pollen contributors. The resulting pollen groups were ‘Rainforest taxa’, ‘Xerophytic forest taxa’, and ‘Shrubland/steppe taxa’. Taxa whose pollen percentages remained low (<2%) throughout the core were assigned to two pollen categories: ‘Other rainforest taxa’ (i.e. Fuchsia sp., Drimys sp., Gevuina sp., Weinmannia sp.) and ‘Other shrubland/steppe taxa’ (i.e. Solanaceae, Caryophyllaceae, Galium sp.). Nothofagus dombeyi-type pollen includes the closed forest trees N. dombeyi and N. pumilio, and the disturbance-adapted open-forest tree N. antarctica. Cupressaceae pollen is attributed largely to Austrocedrus chilensis, although Fitzroya cupressoides and Pilgerodendron uviferum are potential long-distance contributors. Rhamnaceae pollen is likely to come from species of the genera Discaria sp. and/or Colletia sp.

A regional composite forest-to-steppe taxa ratio was constructed and interpreted as a measurement of the position of the forest-steppe ecotone relative to its mean Holocene position. Pollen percentages from L. Mosquito and L. Cóndor were normalized to stabilize the variance and standardized to prevent overrepresented taxa from dominating the signal, making the values comparable across the two sites. Forest-to-steppe taxa ratios were calculated with the transformed pollen data from each watershed. The composite record was constructed by combining the forest-to-steppe taxa ratios from both sites and extracting the low frequency component of the resulting time series with a smoothing cubic spline (smoothing parameter=0.675, lambda=3.457, equivalent degrees of freedom=10.644). Unless otherwise indicated, all the graphs and statistical analyses were performed with R (R Development Core Team, 2010).

Results

Lithology

The Lago Mosquito record – whose lithology was composed of a basal unit of laminated silty organic clay (1506–1421 cm), overlained by fine detritus gyttja (1421–0 cm) with several tephras – was published by Whitlock et al. (2006). Given that the authors provide a detailed description of the cores (Mos03A and Mos03C), we only present the lithology of the LC06A core retrieved from L. Cóndor. For consistency, the radiocarbon dates from Mos03A and Mos03C were re-calibrated and a new chronology was developed for L. Mosquito using the same techniques employed in the L. Cóndor age model (Table 1).

Three lithological units were identified in the 289 cm long LC06A core (Figure 3). The basal unit, 289–279 cm depth, was composed of laminated clay. It was overlain by a decomposed 1.5 cm peat layer (279–277.5 cm depth) and a unit of fine-detritus gyttja (277.5–0 cm depth). The gyttja unit contained a black scoria tephra at 273–268.5 cm depth (CT-1), a thick white pumiceous tephra at 230.5–213.5 cm depth (CT-2), and a thin ash layer at 209–208 cm depth (CT-3).

Figure 3.

Lithologic, MS and LOI data, and sedimentation rates for the LC06A core. Radocarbon dates are shown.

Tephra CT-2 was composed of white pumice lapilli of rhyolitic composition (74% SiO₂) and, according to the chronology proposed for L. Cóndor, was deposited at 9212 ± 100 cal. yr BP. This tephra is correlated with a white to yellow rhyolite pumice fall layer exposed in several road cuts in the surroundings of the Chilean town of Chaitén that overlies a 9370 yr old pyroclastic surge deposit (Figure 4a). That pumice fall layer has been the only reported deposit from Chaitén volcano prior to the 2008 eruption (Naranjo and Stern, 2004). The geochemistry of Tephra CT-2 matches the characteristic major, trace and REE composition of the lower white pumice tephra present in the L. Mosquito core (1506 cm depth; 9263 ± 101 cal. yr BP). It also matches other rhyolitic tephra units found in sediment records and surface exposures in Argentina, including the white to yellow pumice lapilli tephra unit exposed in several road cuts along the Los Alerces National Park (Figure 4a and b), which has previously been attributed to the Michinmahuida volcano (Naranjo and Stern, 2004). The major element composition of Tephra CT-2 is similar to that of the Chaitén 2008 ash (Alfano et al., 2010; Horwell et al., 2008), indicating that this volcano, the only known source of rhyolitic products in the area, has likely been more active than previously reported.

Figure 4.

(a) K₂O/Na₂O versus SiO₂ content of tephra layers (open symbols) in Lago Mosquito (MT) and L. Cóndor (CT-2) sediment cores, road-cuts in Los Alerces National Park (LAT), ashfall from the Chaitén 2008 eruption (CHA 08) and tephras from volcanoes from the study area (solid symbols CHA: Chaitén; COR: Corcovado; MEL: Melimoyu; MIC: Michinmahuida and YAN: Yanteles from Naranjo and Stern, 2004). Right: detail of the cluster of samples interpreted as derived from Chaitén volcano, including two tephras previously attributed to Michinmahuida volcano. (b) Chrondrite-normalized REE patterns for selected rhyolitic tephra samples from L. Mosquito and L. Cóndor cores and road-cuts along Los Alerces National Park.

Except for the tephra layers and a gyttja unit from 48 to 35 cm depth, which had low organic and carbonate content and high MS, the organic content of the sediment showed little variation, with values ranging from 0.4 to 4% (Figure 3). Sediment carbonate content was highest in the laminated-clay unit (0.6–0.9%), declined to 0.3% at the clay-to-peat transition, and remained low in the gyttja unit (0.1–0.5%). We suggest that the laminated clay unit was deposited in a periglacial environment, and the decomposed peat layer indicates the presence of a wetland soon after deglaciation.

The pollen and charcoal records

Pollen zones were defined by visually inspecting the L. Cóndor pollen diagram (Figure 5). The main features of the L. Mosquito vegetation and fire history are described in association with contemporaneous L. Cóndor pollen zones in order to infer regional patterns of vegetation and fire. Only the dominant pollen types are shown. For further details on the L. Mosquito pollen record, see Whitlock et al. (2006).

Figure 5.

Lithology, selected pollen percentages, PAR, CHAR, background CHAR, local fire episodes and grass-to-total charcoal ratio for L. Mosquito, and L. Cóndor.

Zone Co-1 (249.5–210 cm depth; >9150 cal. yr BP) was characterized by non-arboreal taxa, primarily Poaceae (>30%). Steppe/shrubland taxa, including Asteraceae (>3%), Chenopodiaceae (>10%) and Apiaceae (>10%), were abundant, and Maytenus sp. reached values of 9%. The assemblage resembled modern pollen spectra from the steppe (Markgraf et al., 1981; Páez et al., 2001), and low PAR values (40–60 grains/cm² per yr) indicate sparse vegetation cover. Background CHAR levels were high (0.7 particles/cm² per yr), but charcoal peaks displayed a frequency of c. 1 fire episode per 200 years, suggesting that local fire was not an important element of the ecosystem.

The vegetation that grew near L. Mosquito was very similar to that at L. Cóndor in terms of community composition. Nothofagus dombeyi-type pollen percentages were lower at L. Mosquito (20%) than at L. Cóndor, where they reached 45%. N. dombeyi-type pollen influx (not shown), however, was lower at L. Cóndor (<124 grains/cm² per yr) than at L. Mosquito (<765 grains/cm² per yr), suggesting that Nothofagus spp. are unlikely to have been present at L. Cóndor and the observed pollen is from sources to the west. Higher PAR values at L. Mosquito (200–480 grains/cm² per yr) indicate that vegetation cover was greater than at L. Cóndor, implying a west-to-east productivity gradient. Background CHAR values of up to 48 particles/cm² per yr suggest maximum Holocene biomass burning, and forest-to-steppe taxa ratios indicate the steppe reached its westernmost position during this interval.

Zone Co-2 (210–91.5 cm depth; 9150–5250 cal. yr BP) featured a sharp increase in N. dombeyi-type (to 65%) at the bottom of the zone, and a rise of Cupressaceae (to 15%) at c. 8500 cal. yr BP. Steppe/shrubland taxa, Poaceae and Asteraceae were less abundant than before (<25%, <5% and <3%, respectively), and PAR values increased to 250 particles/cm² per yr. The assemblage compares well with modern samples from sclerophyllous shrubland and steppe (Markgraf et al., 1981; Páez et al., 2001). Background CHAR, which initially was 1.5 particles/cm² per yr, rose to 5 particles/cm² per yr at 6250 cal. yr BP and remained high for the following 2500 years, implying a trend towards increased biomass burning. A similar pattern was observed for fire-episode frequency, as charcoal peaks gradually increased in frequency from near-absence at the bottom of the zone to 1 episode per 100 years at 6250 cal. yr BP.

Following an initial increase in N. dombeyi-type pollen percentages at c. 8900 cal. yr BP, the L. Mosquito pollen assemblage suggests a shrubland with elements from the forest and significant areas of steppe. Background CHAR at L. Mosquito decreased from 48 particles/cm² per yr at the bottom of the zone, to 5 particles/cm² per yr at 6500 cal. yr BP, the same values registered at L. Cóndor at this time. These CHAR levels indicate decreased biomass burning, and total-to-grass-charcoal ratios, ranging between 0.01 and 0.39, suggest primarily woody fuels. An eastward expansion of the forest can be inferred from forest-to-steppe ratios.

Zone Co-3 (91.5–27 cm; 5250–2200 cal. yr BP) contained high values of open-forest taxa, such as Cuppresaceae (up to 17%) and Rhamnaceae (7%), and lower amounts of N. dombeyi-type pollen (<55%) than the previous zone. The highest percentages of steppe/shrubland taxa (>7%) in the record were observed in this zone. This assemblage is comparable with modern pollen samples from the A. chilensis forest/steppe ecotone (Markgraf et al., 1981; Páez et al., 2001). PAR values declined at the base of the zone (to 60 grains/cm² per yr) and remained low until the top of the zone, indicating open vegetation prevailed throughout the period. Background CHAR and fire-episode frequency were low, reaching zero values between 3500 and 2700 cal. yr BP, implying negligible fire activity for the period.

The L. Mosquito pollen assemblage featured a trend towards decreased N. dombeyi-type pollen and increased Poaceae as well. Increases in Cupressaceae pollen mark the beginning of local A. chilensis expansion into forest. Grass-to-total charcoal ratios suggest a rapid shift from canopy to surface fires at 4500 cal. yr BP. Forest-to-steppe taxa ratios suggest a westward steppe expansion into the forest.

Zone Co-4 (27–0 cm depth; 2200 cal. yr BP to present) had higher pollen values of N. dombeyi-type (>65%) and epiphytic Misodendrum sp. (to 6%), and lower values of Rhamnaceae (less than 1%), Poaceae (<35%) and other shrubland/steppe taxa (<1%) than the previous zone. PAR increased abruptly with respect to the previous zone, reaching the maximum values of the record (600 grains/cm² per yr) at the top. Comparison with modern pollen assemblages suggests an open xerophytic forest (Markgraf et al., 1981; Páez et al., 2001). Background CHAR remained very low (<1 particles/cm² per yr) and charcoal peaks were virtually absent, indicating local fires were rare.

Austrocedrus chilensis populations west of the L. Mosquito area are inferred from the increase of Cupressaceae pollen at the expense of N. dombeyi-type in the L. Mosquito record. During the last 200 years, Rumex sp. (R. acetosella-type), and Pinus spp. (not shown) were present in small amounts, reflecting the effects of European settlement on vegetation composition.

Discussion

Holocene vegetation, climate and fire history of the forest-steppe ecotone at lat. 42°S, based on the pollen and charcoal records at Lago Mosquito and Laguna del Cóndor, is summarized in Figure 4. A cubic spline was fit to the forest-to-steppe taxa records from both sites to develop a composite record, which was used to infer the west–east position of the ecotone relative to its mean Holocene location. Positive anomalies relative to the mean position for the last 10,000 years indicate eastward shifts of the ecotone; negative anomalies represent westward expansions of steppe at the expense of forest. The paleovegetation reconstruction was compared with changes in fire frequency and fuel composition at each site to understand the role of fire in promoting ecotonal shifts. In addition, the climate implications of the vegetation and fire regime changes in Mosquito-Cóndor region (Figure 6) are discussed in light of other paleoclimatic records from southern South America.

Figure 6.

Position of the forest-steppe ecotone as inferred from a composite forest-to-steppe taxa ratio record. Present-day position of L. Mosquito and L. Cóndor relative to the ecotone is based on modern forest-to-steppe taxa ratios. The proportion of Nothofagus and Austrocedrus pollen to total forest taxa pollen, and the fire episode frequency records from both sites are shown.

Early Holocene (>9000 cal. yr BP)

Prior to 9000 cal. yr BP, the forest-steppe ecotone was located west of its mean Holocene position, such that both L. Mosquito and L. Cóndor supported steppe vegetation (Figures 5 and 6) dominated by grasses, Chenopodiaceae, Asteraceae and Apiaceae. The pollen data suggest drought-tolerant shrubs and trees, such as Rhamnaceae, Maytenus sp. and possibly Nothofagus antarctica, grew in the area, probably on south-facing slopes and cool deep Andean valleys, where the water balance would have been more positive. These re-sprouting shrubs burned frequently at L. Mosquito, the westernmost site, and less often at L. Cóndor, likely reflecting lower fuel loads and/or fuel discontinuity associated with sparser vegetation in the east (Huber et al., 2004). Along with high fire activity, Nothofagus spp. began to expand at both sites at 9000 cal. yr BP, coupled with a reduction of steppe taxa, suggesting that summers were arid enough to dry fuels and support fires, while winter moisture was sufficient for forest expansion (Markgraf et al., 2007).

Evidence of dry summers in the early Holocene has been found throughout Patagonia. East of the Andes, vegetation was structurally more open than today, with steppe-tundra in the south (lat. 52°S; Fesq-Martin et al., 2004), parkland between lat. 49 and 54°S (Huber et al., 2004), xerophytic forest at lat. 41°S (Whitlock et al., 2006) and steppe in the north (lat. 40°S; Markgraf and Bianchi, 1999). Patagonian records between lat. 40 and 56°S on both sides of the Andes show high fire activity (Whitlock et al., 2007), and Chilean sites between lat. 35° (Jenny et al., 2003) and 40°S (Bertrand et al., 2008) indicate lower-than-present lake levels. Regional aridity has been attributed to the effects of higher-than-present winter and annual insolation (Berger and Loutre, 1991) on temperature and atmospheric circulation patterns. Increased annual insolation led to higher temperatures and evapotranspiration. Reduced seasonality weakened and/or latitudinally shifted the SW and storm tracks south of their present position (Moreno et al., 2010; Whitlock et al., 2007), resulting in decreased precipitation throughout Patagonia. In the Mosquito-Cóndor region, the early onset of the fire season combined with the high flammability and rapid regrowth of steppe shrubs (Mermoz et al., 2005) would have made the ecotone especially prone to fire.

Middle Holocene (9000–5250 cal. yr BP)

During the middle Holocene, the forest-steppe ecotone in the Mosquito-Cóndor region was located east of its early-Holocene position and dominated by scherophyllous shrubs, likely as a consequence of lower-than-before temperatures and/or higher effective moisture. PAR indicates that vegetation cover was greater at L. Mosquito than at L. Cóndor, suggesting a west-to-east vegetation gradient, with more productive communities towards the west. Higher charcoal peak frequency and background CHAR at L. Mosquito indicate that fire episodes (Figures 5 and 6) were more frequent and fuel biomass was greater west of the Mosquito-Cóndor area. Low grass-to-total-charcoal ratios suggest that such fires were fueled by woody vegetation.

Throughout the middle and southern latitudes of Chile and Argentina, the onset of cooler and/or effectively wetter-than-before conditions in the middle Holocene has been inferred from multiple climate proxies and attributed to changes in insolation. Decreasing annual insolation and amplification of the seasonal cycle of insolation (Berger and Loutre, 1991) increased seasonality and resulted in lower-than-before annual temperatures. As a consequence, sedimentological and/or geochemical evidence from lat. 35° (Jenny et al., 2003) and 40°S (Bertrand et al., 2008) suggests higher-than-before lake levels, and high-resolution pollen and charcoal records document forest expansions, coupled with reduced fire activity throughout Patagonia (Abarzúa and Moreno, 2008; Markgraf et al., 2007; Moreno, 2004; Whitlock et al., 2007; Wille and Schaebitz, 2009). Cooler and/or effectively wetter conditions explains the advance of mountain glaciers in the middle- and high-latitude Andes (Douglass et al., 2005; Porter, 2000), as well as sea-ice expansion and sea-surface cooling in the Atlantic sector of the Southern Ocean (Lamy et al., 2001; Liu et al., 2003).

Shifts in the relative position of the forest-steppe ecotone in the Mosquito-Cóndor region suggest a two-step increase in effective moisture during the middle Holocene (Figure 6). The first step occurred at 8500 cal. yr BP and is inferred from the eastward expansion of Nothofagus spp. from L. Mosquito and the higher levels of Austrocedrus chilensis at L. Cóndor, inferred from the rise in Cupressaceae pollen from previously negligible levels (<6%). The second step at 5500 cal. yr BP is based on the increase of Nothofagus spp. near L. Cóndor, suggesting an advance of forest into the steppe.

The history and ecology of Austrocedrus chilensis in this region warrant special attention. Cupressaceae pollen percentages rose to 15% at L. Cóndor at 8500 cal. yr BP, and comparison with modern pollen samples (Markgraf et al., 1981; Páez et al., 2001) indicates that these values came from small populations. Palynological data (Markgraf, 1980, 1983, 1984, 1987; Markgraf and Bianchi, 1999; Markgraf et al., 1986) and genetic evidence (Pastorino and Gallo, 2002) suggest that A. chilensis survived the last glaciation east of the Andes in multiple small refugia, possibly east of its present distribution. Its middle-Holocene presence at L. Cóndor and its delayed appearance at L. Mosquito at 3500 cal. yr BP are consistent with postglacial expansion westward from the steppe margin.

Present-day establishment patterns of Austrocedrus chilensis vary according to habitat type, associated species, and the influence of fuel cover on fire regimes. At mesic sites, where competition from Nothofagus sp. and fire-resistant shrubs is intense, high-severity fire effectively precludes establishment of A. chilensis (Veblen et al., 1995). At drier sites, low-severity fire limits A. chilensis seedling establishment, but the trees become more resistant to patchy surface fires as they mature. Dendroecological studies reveal that discontinuous age structures are characteristic of A. chilensis stands close to the forest-steppe ecotone (Kitzberger et al., 1997). The age gaps reflect intermittent opportunities for seedling establishment during periods of above-average moisture availability (Villalba and Veblen, 1997). Increased effective moisture during the growing season at 8500 cal. yr BP may have allowed a small A. chilensis population to establish, possibly in the rocky hills between the two sites or in the L. Cóndor watershed, whereas high severity fires may have limited expansion westward to L. Mosquito and the eastern Andes before 3500 cal. yr BP.

Late Holocene (5250 cal. yr BP–present)

Between 5250 and 3000 cal. yr BP, steppe vegetation was present in the Mosquito-Cóndor region and the forest-steppe ecotone was sharply defined by the pollen dominance of Nothofagus dombeyi-type and Rhamnaceae and low levels of Cuppresaceae. The westward expansion of steppe taxa (Figure 6) implies colder conditions and/or a loss of growing season moisture, especially from 3700 to 3000 cal. yr BP. An associated rise in magnetic susceptibility and decreased sediment organic-matter content at L. Cóndor (48 cm depth, 3658 cal. yr BP to 35 cm depth, 2844 cal. yr BP; Figure 3), suggest increased inorganic input to the lake, possibly as a consequence of greater windiness or more surface run-off from exposed ground.

Evidence of cooler conditions between 3700 and 3000 yr BP is found throughout the mid and high latitudes (lat. 30–53°S), including sedimentological data that indicate low lake paleoproductivity (Bertrand et al., 2008), and paleoecological records that suggest expansions of cold resistant taxa (Heusser, 1990, 1995; Villagrán and Varela, 1990). This period also marks extensive glaciation in the Andes (Douglass et al., 2005; Grosjean et al., 1997, 1998; Mercer, 1982; Moy et al., 2009). Cooler conditions and year-round precipitation may have been caused by a northwards migration of the winter polar front and SW from their middle Holocene position (van Geel et al., 2000).

The last 3000 years are characterized by the eastward expansion of the forest-steppe ecotone in the Mosquito-Cóndor region. The expansion was mainly caused by an increase of Austrocedrus chilensis population size at both sites, following a shift in the fire regime from intense woody fires to surface fires (Figures 5 and 6). This directional change in the position and composition of the ecotone is consistent with wetter springs and/or summers than before that would have promoted A. chilensis expansion, as well as surface fire regimes. Increased effective moisture during the growing season has also been inferred from changes in forest composition in the mid-latitudes west of the Andes (Abarzúa and Moreno, 2008; Haberle and Bennett, 2004; Villagrán and Varela, 1990), suggesting that the climate change at lat. 42°S could have resulted from intensified westerly wind activity (Moy et al., 2009).

During the period of European settlement (< c. ad 1800), a decline of Nothofagus dombeyi-type pollen, the appearance of non-native pollen taxa and increased CHAR and/or fire episode frequency have been registered in sediment records throughout Patagonia (Abarzúa and Moreno, 2008; Haberzettl et al., 2006; Huber and Markgraf, 2003; Wille et al., 2007). In contrast, neither L. Cóndor or L. Mosquito show pollen or charcoal changes that suggest a dramatic decline of closed-forest taxa or major changes in the fire regime during the last 200 years. Nonetheless, trace amounts of pollen from non-native species such as Rumex sp. (R. acetosella-type) and Pinus spp. were identified in the L. Mosquito record, and the large population of sweet briar rose (Rosa rubiginosa) currently growing at the site suggests that agriculture, forest clearance, and intensive grazing had significant effects in this region, just as they did across the forest-steppe ecotone of Patagonia (Huber and Markgraf, 2003).

Conclusions

During the early Holocene prior to 9000 cal. yr BP, the forest-steppe ecotone was located west of its Holocene mean position. Lower-than-present seasonality led to the early onset of the fire season. The high fuel flammability in the western region (west of L. Mosquito) resulted in frequent fires, but at the eastern limit (near L. Cóndor), low vegetation cover could not support frequent fires, even though the climate was suitable.

At c. 9000 cal. yr BP, Nothofagus spp. and other forest taxa expanded into steppe, resulting in an eastward shift of the ecotone. This change in vegetation is reflected in the fire regime: fire frequency and area burned increased in the study area and also regionally. The middle-Holocene fire and vegetation suggests that summers were arid enough to support fires at the forest-steppe ecotone, and winter moisture was sufficient for Nothofagus spp. to expand. About 500 years later, Austrocedrus chilensis established near L. Cóndor, probably as part of a westward migration from steppe margin glacial refugia. It is possible that increased effective moisture allowed a small A. chilensis population to persist in the rocky hills between the two sites or in the L. Cóndor watershed for 5000 years before its expansion westward to L. Mosquito.

The 5250 to 3000 cal. yr BP period was cold and/or dry, as evidenced by the contraction of forest to the west. Cooling is inferred from marine and terrestrial records in the southern middle and high latitudes and related to a northward shift in the polar front and southwesterlies.

The last 3000 years are characterized by an eastward expansion of the forest-steppe ecotone, with dominance of Austrocedrus chilensis. This shift indicates a recent trend towards wetter conditions. European settlement is reflected in the establishment of non-native and disturbance-adapted species.

Footnotes

Acknowledgements

We thank V. Markgraf and B. Gresswell for participation in fieldwork. Permission to core Lago Cóndor was provided by Estancia Leleque. V. Nagashima helped with lab analyses. The paper benefitted from the comments of two anonymous reviewers.

Funding

This work was supported by the National Science Foundation (grant numbers ATM-0714061, OISE-0966472).

References

Abarzúa

Moreno

(2008) Changing fire regimes in the temperate rainforest region of southern Chile over the last 16,000 yr. Quaternary Research 69: 62–71.

Alfano

Bonadonna

Volentik

ACM

. (2010) Tephra stratigraphy and eruptive volume of the May, 2008, Chaitén eruption, Chile. Bulletin of Volcanology 73: 613–630.

Bennett

(1993) Holocene forest dynamics with respect to southern Ontario. Review of Palaeobotany and Palynology 79: 69–71.

Berger

Loutre

(1991) Insolation values for the climate of the last 10 million years. Quaternary Science Reviews 10: 297–317.

Bertrand

Charlet

Charlier

. (2008) Climate variability of southern Chile since the Last Glacial Maximum: A continuous sedimentological record from Lago Puyehue (40°S). Journal of Paleolimnology 39: 179–195.

Bianchi

Ariztegui

(2011) Vegetation history of the Río Manso superior catchment area, Northern Patagonia (Argentina), since the last deglaciation. The Holocene (this issue). doi: 10.1177/0959683611405083.

Blaauw

Bennett

Christen

(2010) Random walk simulations of fossil proxy data. The Holocene 20(4): 645–649.

Briles

Whitlock

Bartlein

. (2008) Regional and local controls on postglacial vegetation and fire in the Siskiyou Mountains, northern California, USA. Palaeogeography Palaeoclimatology Palaeoecology 265: 159–169.

Caldenius

(1932) Las glaciaciones cuaternarias en la Patagonia y Tierra del Fuego. Buenos Aires: Dirección General de Minas y Geología, 152 pp.

10.

Dean

Jr (1974) Determination of carbonate and organic matter in calcareous sediments and sedimentary rocks by loss on ignition: Comparison with other methods. Journal of Sedimentology and Petrology 44: 242–248.

11.

Douglass

Singer

Kaplan

. (2005) Evidence of early Holocene glacial advances in southern South America from cosmogenic surface-exposure dating. Geology 33: 237–240.

12.

Faegri

Iversen

(1989) Textbook of Pollen Analysis. 4th Edition. Alden.

13.

Fesq-Martin

Friedmann

Behrmann

. (2004) Late-glacial and Holocene vegetation history of the Magellanic rain forest in southwestern Patagonia, Chile. Vegetation History and Archaeobotany 13(4): 249–255.

14.

Garreaud

Vuille

Compagnucci

. (2008) Present-day South American climate. Palaeogeography, Palaeoclimatology, Palaeoecology 281: 229–241.

15.

Gedye

Jones

Tinner

. (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.

16.

Grosjean

Gayh

Messerll

. (1998) A late-Holocene (<26,000 BP) glacial advance in the south-central Andes (29°S), northern Chile. The Holocene 8: 473–479.

17.

Grosjean

Valero-Garces

Gayh

. (1997) Mid- and late-Holocene limnogeology of Laguna del Negro Francisco, northern Chile, and its paleoclimatic implications. The Holocene 7: 151–159.

18.

Haberle

Bennett

(2004) Postglacial formation and dynamics of North Patagonian rainforest in the Chonos Archipelago, Southern Chile. Quaternary Science Reviews 23: 2433–2452.

19.

Haberzettl

Wille

Fey

. (2006) Environmental change and fire history of southern Patagonia (Argentina) during the last five centuries. Quaternary International 158: 72–82.

20.

Hajek

di Castri

(1975) Bioclimatoligía de Chile. Santiago: Universidad Católica de Chile, 160 pp.

21.

Heusser

(1971) Pollen and Spores of Chile. Modern Types of the Pteridophyta, Gymnospermae, and Angiospermae. Tucson: University of Arizona Press.

22.

Heusser

(1990) Ice age vegetation and climate of subtropical Chile. Palaeogeography, Palaeoclimatology, Palaeoecology 80: 107–127.

23.

Heusser

(1995) Three Late Quaternary pllen diagrams from southern Patagonia and their palaeoecological implications. Palaeogeography, Palaeoclimatology, Palaeoecology 118: 1–24.

24.

Higuera

Brubaker

Anderson

. (2009) Vegetation mediated the impacts of postglacial climatic change on fire regimes in the south-central Brooks Range, Alaska. Ecological Monographs 79: 201–219.

25.

Higuera

Sprugel

Brubaker

(2005) Reconstructing fire regimes with charcoal from small-hollow sediments: A calibration with tree-ring records of fire. The Holocene 15: 238–251.

26.

Higuera

Whitlock

Gage

(2010) Linking tree-ring and sediment-charcoal records to reconstruct fire occurrence and area burned in subalpine forests of Yellowstone National Park, USA. The Holocene doi: 10.1177/0959683610374882.

27.

Horwell

Michnowicz

Le Blond

(2008) Report on the mineralogical and geochemical characterisation of Chaitén ash for the assessment of respiratory health hazard. International Volcanic Health Hazard Network (IVHNN) report, Durham University, 36 pp. (available from http://www.ivhhn.org).

28.

Hubbell

(2001) The Unified Neutral Theory of Biodiversity and Biogeography. Princeton University Press.

29.

Huber

Markgraf

(2003) European impact on fire regimes and vegetation dynamics at the steppe-forest ecotone of southern Patagonia. The Holocene 13: 567–579.

30.

Huber

Markgraf

Schaebitz

(2004) Geographical and temporal trends in Late Quaternary fire histories of Fuego-Patagonia, South America. Quaternary Science Reviews 23: 1079–1097.

31.

Intergovernmental Panel on Climate Change (2007) Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, 2007. Solomon

Qin

Manning

(eds) Cambridge, New York: Cambridge University Press.

32.

Ives

Einarsson

Jansen

VAA

. (2008) High-amplitude fluctuations and alternative dynamical states of midges in Lake Myvatn. Nature 452: 84–87.

33.

Jenny

Valero-Garcés

Urrutia

. (2003) Moisture changes and fluctuations of the Westerlies in Mediterranean Central Chile during the last 2000 years: The Laguna Aculeo record (33°50′S). Quaternary International 87: 3–18.

34.

Jobbágy

Paruelo

León

RJC

(1995) Estimación de la precipitación y de su variabilidad interanual a partir de información geográfica en el NW de Patagonia, Argentina. Ecologia Austral 5: 47–53.

35.

Jobbágy

Paruelo

León

RJC

(1996) Vegetation heterogeneity and diversity in flat and mountain landscapes of Patagonia. Journal of Vegetation Science 7: 599–608.

36.

Juggins

(2007) C2 Version 1.5 User Guide. Software for Ecological and Palaeoecological Data Analysis and Visualization. Newcastle-upon-Tyne: Newcastle University.

37.

Kitzberger

Veblen

Villalba

(1997) Climatic influences on fire regimes along a rainforest-to-xeric woodland gradient in northern Patagonia, Argentina. Journal of Biogeography 24: 35–47.

38.

Lamy

Hebbeln

Röhl

. (2001) Holocene rainfall variability in southern Chile: A marine record of latitudinal shifts of the Southern Westerlies. Earth and Planetary Science Letters 185: 369–382.

39.

Liu

Brady

Lynch-Stieglitz

(2003) Global ocean response to orbital forcing in the Holocene, Paleoceanography 18: 1041–1060.

40.

Long

Whitlock

Bartlein

. (1998) A 9000-year fire history from the Oregon Coast Range, based on a high-resolution charcoal study. Canadian Journal of Forest Research 28: 774–787.

41.

McCormac

Hogg

Blackwell

. (2004) SHCal04 Southern Hemisphere calibration, 0–11.0 cal kyr BP. Radiocarbon 46: 1087–1092.

42.

Markgraf

(1980) Paleoclimatic reconstruction of the last 15,000 years in subantarctic and temperate regions of Argentina. Memoires Museum National Historie Naturelle 27: 87–97.

43.

Markgraf

(1983) Late and postglacial vegetational and paleoclimatic changes in subantarctic, temperate, and arid environments in Argentina. Palynology 7: 43–70.

44.

Markgraf

(1984) Late Pleistocene and Holocene vegetation history of temperate Argentina: Lago Morenito, Bariloche. Dissertation Botanique 72: 235–254.

45.

Markgraf

(1987) Paleoenvironmental changes at the northern limit of the subantarctic Nothofagus forest, lat. 37°S, Argentina. Quaternary Research 28: 119–129.

46.

Markgraf

Bianchi

(1999) Paleoenvironmental changes during the last 17,000 years in western Patagonia: Mallín Aguado, Province of Neuquén, Argentina. Bamberger Geographische Schriften 19: 175–193.

47.

Markgraf

D’Antoni

(1978) Pollen flora of Argentina. Tucson: University of Arizona Press.

48.

Markgraf

Bradbury

Fernandez

(1986) Bajada de Rahue, Province of Neuqu_en, Argentina: An interstadial deposit in northern Patagonia. Palaeogeography, Palaeoclimatology, Palaeoecology 56: 251–258.

49.

Markgraf

D’Antoni

Ager

(1981) Modern pollen dispersal in Argentina. Palynology 5: 43–63.

50.

Markgraf

Whitlock

Haberle

(2007) Vegetation and fire history during the last 18,000 cal yr BP in Southern Patagonia: Mallín Pollux, Coyhaique, Province Aisén (45°41′30″S, 71°50′30″W, 840 m elevation). Palaeogeography, Palaeoclimatology, Palaeoecology 257: 492–507.

51.

Mayr

Fey

Haberzettl

. (2005) Paleoenvironmental changes in southern Patagonia during the last millennium recorded in lake sediments from Laguna Azul (Argentina). Palaeogeography, Palaeoclimatology, Palaeoecology 228: 203–227.

52.

Mercer

(1982) Holocene glacier variations in southern South America. Striae 18: 35–40.

53.

Mermoz

Kitzberger

Veblen

(2005) Landscape influences on occurrence and spread of wildfires in Patagonian forests and shrublands. Ecology 86: 2705–2715.

54.

Miller

(1976) The climate of Chile. In: Schwerdtfeger

(ed.) Climates of Central and South America. Amsterdam: Elsevier, 113–145.

55.

Moreno

(2004) Millennial-scale climate variability in northwest Patagonia over the las 15,000 yr. Journal of Quaternary Science 19: 35–47.

56.

Moreno

Francois

Villa-Martínez

. (2010) Covariability of the Southern Westerlies and atmospheric CO₂ during the Holocene. Geology 38: 727–730.

57.

Moy

Moreno

Dunbar

. (2009) Climate change in Southern South America during the last two millennia. In: Vimeux

Sylverstre

Khodri

(eds) Past Climate Variability in South America and Surrounding Regions. Springer, 353–395.

58.

Naranjo

Stern

(2004) Holocene tephrochronology of the southernmost part (42°30′–45°S) of the Andean Southern Volcanic Zone. Revista Geológica de Chile 31: 225–240.

59.

Páez

Schäbitz

Stutz

(2001) Modern pollen–vegetation and isopoll maps in southern Argentina. Journal of Biogeography 28: 997–1021.

60.

Paruelo

Beltrán

Jobbágy

. (1998) The climate of Patagonia: General patterns and controls on biotic processes. Ecologia Australia 8: 85–101.

61.

Pastorino

Gallo

(2002) Quaternary evolutionary history of Austrocedrus chilensis, a cypress native to the Andean–Patagonian forest. Journal of Biogeography 29: 1167–1178.

62.

Porter

(2000) Onset of Neoglaciation in the Southern Hemisphere. Journal of Quaternary Science 15(4): 395–408.

63.

Prohaska

(1976) The Climate of Argentina, Paraguay and Uruguay. In: Schwerdtfeger

(ed.) Climates of Central and South America. Amsterdam: Elsevier, 13–111.

64.

R Development Core Team (2010) R: A Language and Environment for Statistical Computing. Vienna: R Foundation for Statistical Computing, ISBN 3-900051-07-0, URL http://www.R-project.org.

65.

Reimer

Baillie

MLG

Bard

. (2009) IntCal09 and Marine09 radiocarbon age calibration curves, 0–50,000 years cal BP. Radiocarbon 51: 1111–1150.

66.

Schnurrenberger

Russell

Kelts

(2003) Classification of lacustrine sediments based on sedimentary components. Journal of Paleolimnology 29: 141–153.

67.

Seibert

(1982) Carta de la vegetación de la región de El Bolsón, Río Negro y su aplicación a la planificación del uso de la tierra. Documenta Phytosociologica 2: 1–120.

68.

Stern

(2004) Active Andean volcanism: Its geologic and tectonic setting. Revista Geológica de Chile 31: 161–206.

69.

Stuiver

Reimer

(1993) CALIB 6.0.1. www program and documentation. http://radiocarbon.pa.qub.ac.uk/calib/calib.html

70.

Telford

Heegaard

Birks

HJB

(2005) All age–depth models are wrong: But how badly? Quaternary Science Reviews 23: 1–5.

71.

van Geel

Heusser

Renssen

. (2000) Climatic change in Chile at around 2700 BP and global evidence for solar forcing: A hypothesis. The Holocene 10: 659–664.

72.

Veblen

Burns

Kitzberger

. (1995) The ecology of the conifers of southern South America. In: Enright

Hill

(eds) The Ecology of Southern Conifers. Melbourne University Press.

73.

Veblen

Kitzberger

Raffaele

. (2008)The historical range of variability of fires in the Andean-Patagonian Nothofagus forest region. International Journal of Wildland Fire 17: 724–741.

74.

Veblen

Kitzberger

Villalba

. (1999) Fire history in northern Patagonia: The roles of humans and climatic variation. Ecological Monographs 69: 47–67.

75.

Villalba

Veblen

(1997) Spatial and temporal variation in Austrocedrus growth along the forest-steppe ecotone in northern Patagonia. Canadian Journal of Forestry Research 27: 580–597.

76.

Villalba

Lara

Boninsegna

. (2003) Large scale temperature changes across the southern Andes: 20th-century variations in the context of the past 400 years. Climate Change 59: 177–232.

77.

Villagrán

Varela

(1990) Palynological evidence for increased aridity on the Central Chilean coast during the Holocene. Quaternary Research 34: 198–207.

78.

Whitlock

Anderson

(2003) Fire history reconstructions based on sediment records from lakes and wetlands. In: Veblen

Baker

Montenegro

( eds) Fire and Climatic Change in Temperate Ecosystems of the Western Americas. New York: Springer, 265–295.

79.

Whitlock

Larsen

CPS

(2001) Charcoal as a fire proxy. In: Smol

Birks

HJP

Last

(eds) Tracking Environmental Change Using Lake Sediments: Volume 3 Terrestrial, Algal, and Siliceous Indicators. Dordrecht: Kluwer Academic Publishers, 75–97.

80.

Whitlock

Bianchi

Bartlein

. (2006) Postglacial vegetation, climate, and fire history along the east side of the Andes (lat 41–42.5°S), Argentina. Quaternary Research 66: 187–201.

81.

Whitlock

Moreno

Bartlein

(2007) Climatic controls of Holocene fire patterns in southern South America. Quaternary Research 68: 28–36.

82.

Wille

Schaebitz

(2009) Late-glacial and Holocene climate dynamics at the steppe/forest ecotone in southernmost Patagonia, Argentina: The pollen record from a fen near Brazo Sur, Lago Argentino. Vegetation History and Archaeobotany 18: 225–234.

83.

Wille

Maidana

Schaebitz

. (2007) Vegetation and climate dynamics in southern South America: The microfossil record of Laguna Potrok Aike, Santa Cruz, Argentina. Review of Paleobotany and Palynology 146: 234–246.