Letters to Nature

Nature 409, 804-808 (15 February 2001) | doi:10.1038/35057252; Received 12 February 2000; Accepted 28 November 2000

Interhemispheric climate links revealed by a late-glacial cooling episode in southern Chile

Patricio I. Moreno1,2, George L. Jacobson, Jr1,3, Thomas V. Lowell4 & George H. Denton1,5

  1. Institute for Quaternary and Climate Studies,
  2. Department of Biological Sciences,
  3. Department of Geological Sciences, University of Maine, Orono, Maine 04469, USA
  4. Department of Geology, University of Cincinnati, Cincinnati, Ohio 45221, USA
  5. Present address: Departamento de Biología, Facultad de Ciencias, Universidad de Chile, Casilla 653, Santiago, Chile.

Correspondence to: Patricio I. Moreno1,2 Correspondence and requests for materials should be addressed to P.I.M. (e-mail: Email: pimoreno@uchile.cl).


Understanding the relative timings of climate events in the Northern and Southern hemispheres is a prerequisite for determining the causes of abrupt climate changes. But climate records from the Patagonian Andes1, 2, 3, 4 and New Zealand5, 6, 7, 8 for the period of transition from glacial to interglacial conditions—about 14.6–10kyr before present, as determined by radiocarbon dating—show varying degrees of correlation with similar records from the Northern Hemisphere. It is necessary to resolve these apparent discrepancies in order to be able to assess the relative roles of Northern Hemisphere ice sheets and oceanic, atmospheric and astronomical influences in initiating climate change in the late-glacial period. Here we report pollen records from three sites in the Lake District of southern Chile (41°S) from which we infer conditions similar to modern climate between about 13 and 12.214Ckyrbefore present (bp), followed by cooling events at about 12.2 and 11.414Ckyrbp, and then by a warming at about 9.814Ckyrbp. These events were nearly synchronous with important palaeoclimate changes recorded in the North Atlantic region9, supporting the idea that interhemispheric linkage through the atmosphere was the primary control on climate during the last deglaciation. In other regions of the Southern Hemisphere, where climate events are not in phase with those in the Northern Hemisphere, local oceanic influences may have counteracted the effects that propagated through the atmosphere.

Southern Chile is ideal for the study of interhemispheric linkages throughout the Quaternary period for three reasons: (1) it has insolation regimes that are out-of-phase with northern mid-latitudes, (2) it is located on the windward side of southern South America, making it highly sensitive to variations in the southern westerlies, and (3) it is far removed from the direct influence of Northern Hemisphere ice sheets and sites of North Atlantic Ocean thermohaline convection. The southern westerlies carry the precipitation that nourishes alpine glaciers and temperate rainforests in southern Chile. Modern vegetation in the Chilean Lake District (38–42°S) includes—from low to high altitudes—the Valdivian, North Patagonian (NPRF) and subantarctic rainforest communities10. The chief components of the low- to mid-elevation Valdivian rainforest community are Eucryphia, Nothofagus dombeyi, Nothofagus obliqua, several species of Myrtaceae and Proteaceae, epiphytic ferns, and woody lianas. Cupressaceae and Podocarpaceae are characteristic of the NPRF community above 500m elevation11, 12. At 41°S, the deciduous Nothofagus pumilio and Nothofagus antarctica dominate the subantarctic rainforests near to the tree line at ~1,100m elevation, along with composite and ericaceous shrubs, and herbs.

We developed pollen records from three sites in the lowlands of the Lake District (Fig. 1). Glacial deposits and landforms corresponding to the last ice age have been examined in detail in this region1, 2, 12, 13, 14; these studies show an abrupt withdrawal of Andean piedmont glacier lobes at about 14.614Ckyr bp, marking the onset of the last termination.

Figure 1: Location map, showing the pollen sites in the Chilean Lake District and the Lago Mascardi site in Argentina.
Figure 1 : Location map, showing the pollen sites in the Chilean Lake District and the Lago Mascardi site in Argentina. Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, or to obtain a text description, please contact npg@nature.com

All sediment cores were obtained using a 5-cm-diameter Wright piston corer. We selected small, closed-basin sites to obtain our sediment cores in order to establish comparisons among sites having similar depositional environments. Sites differ in hydrology as only one of them has persisted as a lake (the Lago Condorito site), while the other sites have undergone terrestrialization processes since the mid-Holocene (Huelmo), or have completely dried out during the early Holocene (the Canal de la Puntilla site). These contrasts in hydrology, along with local differences in microclimate and topography, account for the differences in present vegetation and past pollen assemblages. Despite these differences, all records show the same timing, character, and direction of vegetation change (Figs 2, 3). There is a gradient of decreasing summer and total annual precipitation towards the north, due to a lower frequency of storm tracks, intensification of the rain-shadow effect caused by the Coastal Range, which increases in altitude toward the north, and the maritime influence of the Seno Reloncaví seaway in the south. Precipitation occurs evenly throughout the year in the Seno Reloncaví sector of the Lake District, with decreased frontal activity during the summer months.

High resolution image and legend (42K)

Close-interval sampling and dating of sediment cores (Table 1) from the Canal de la Puntilla (40°56′S, 72°54′W, 120m above sea level), Huelmo (41°31′S, 73°00′W, 25m.a.s.l.) and Lago Condorito sites (41°45′S, 73°07′W, 100m.a.s.l.) (Fig. 1) yielded pollen records with an average time resolution of ~5514Cyears between samples. These sites show a coherent pollen assemblage between ~13 and 12.214Ckyr bp with abundant trees and vines, along with forest ferns (Figs 2, 3). This assemblage indicates the predominance of closed-canopy NPRF under a temperate and humid climate, suggesting conditions approaching modern climate in the foothills of the mountain ranges at ~41°S. The expansion of Podocarpus at ~12.214Ckyr bp and the persistence of rainforest vegetation (Figs 2, 3) are consistent with an onset of cooler conditions, as suggested by the latitudinal and altitudinal distribution of conifers in modern NPRF10, 11. This vegetation change coincided with a prominent transition from basal peat to gyttja in the Lago Condorito record, suggesting that the cooling event was accompanied by an increase in precipitation/effective moisture.

Figure 2: Percentage diagrams of selected taxa from the Canal de la Puntilla, Huelmo and Lago Condorito sites.
Figure 2 : Percentage diagrams of selected taxa from the Canal de la Puntilla, Huelmo and Lago Condorito sites. Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, or to obtain a text description, please contact npg@nature.com

Percentage scales among species vary for visual depiction only. The horizontal dashed lines mark the onset of vegetation and climate changes at ~12.2, ~11.4 and 9.814Ckyrbp. The bars on the right-hand side of the diagrams indicate the mean depth of the radiocarbon dates listed in Table 1. We used polynomial equations to develop age models, in order to assign interpolated radiocarbon ages for the pollen, charcoal, and loss-on-ignition results. All age models yielded R2 > 0.95 and P < 0.001. The abundance of Nothofagus shown in the pollen diagrams refers to the palynomorph Nothofagus dombeyi-type, which includes several species that occur in several forest formations in the temperate region of Chile and Argentina. Although these species occur along a broad latitudinal and altitudinal range, their relative contribution to the local pollen rain is diminished in the low-altitude, thermophilous forest communities. Thus, in the context of the pollen assemblages of Canal de la Puntilla, Huelmo and Lago Condorito between 13 and 914Ckyrbp, the expansion of Nothofagus along with Podocarpus nubigena at 12.2 and ~11.214Ckyr strongly suggests the onset of cooler conditions. The modest increase in the percentages of P. nubigena, and the absence of Pseudopanax laetevirens and Tepualia stipularis in Canal de la Puntilla, suggests that a precipitation gradient between the northern and southern portions of the Lake District was in effect during late-glacial time. All sites were small lakes or ponds between ~13 and 9.814Ckyrbp, as suggested by their notably similar and homogeneous lithologies (silty gyttja, organic silt, and a horizontally lain volcanic ash layer deposited at ~9.814Ckyrbp). The only exception is Lago Condorito, which shows a transition from basal peat to gyttja at ~12.814Ckyrbp. Undisturbed, continuous, in situ sediment deposition at all sites during late-glacial time is shown by multiple cores obtained close to each other at all sites, and is further supported by several lines of evidence (sediment and X-ray stratigraphy, magnetic susceptibility logs, loss-on-ignition analyses, presence of volcanic ash horizons, age–depth curves, and so on; data not shown).

High resolution image and legend (103K)

Figure 3: Influx diagrams of pollen, spores and charcoal particles from the Canal de la Puntilla, Huelmo and Lago Condorito sites.
Figure 3 : Influx diagrams of pollen, spores and charcoal particles from the Canal de la Puntilla, Huelmo and Lago Condorito sites. Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, or to obtain a text description, please contact npg@nature.com

All sites show the expansion and persistence of Podocarpus between ~12.2 and 9.814Ckyrbp. The Canal de la Puntilla and Huelmo records show a sharp increase of cold-resistant trees and declines in other trees and forest ferns at ~11.414Ckyrbp. Unlike the Lago Condorito record, the onset of disturbance by fire in the Huelmo site (760cm) lagged behind the cooling event at ~11.414Ckyrbp (784cm). The abrupt increase of charcoal in Canal de la Puntilla occurred ~1,00014Cyears after the cooling event at ~11.414Ckyrbp. An explanation involving an accentuation of rainfall seasonality and/or continentality is inadequate to account for the occurrence of fires during late-glacial time, because the stratigraphic and geographical distribution of charcoal particles is inverse to the inferred precipitation gradient. Ignition by volcanic activity can be ruled out as a causal agent for the charcoal record, because of the lack of tephra layers between ~13 and 9.814Ckyrbp at all sites. The influx scale for microscopic charcoal particles from Lago Condorito is truncated to emphasize the fire history from 13 to 9.514Ckyrbp. Units of influx: grains or particles per cm2 per yr.

High resolution image and legend (72K)

The Canal de la Puntilla and Huelmo sites show a decline in thermophilous taxa and expansion of Podocarpus and Nothofagus starting at ~11.414Ckyrbp (Figs 2, 3). These cold-resistant trees persisted until ~9.814Ckyrbp in all sites. We interpret these results as an intensification of the vegetation changes that started at ~12.214Ckyrbp, thus suggesting another cooling event at ~11.414Ckyrbp. Hygrophilous NPRF taxa persisted between ~11.4 and 9.814Ckyrbp (arboreal pollen >90%), suggesting conditions similar to the climate that now exists at mid-to-high elevations in the mountain ranges at ~41°S. Subsequent warming at 9.814Ckyrbp led to the expansion of thermophilous taxa and the decline/disappearance of Podocarpus in all sites.

The Huelmo and Lago Condorito sites show the spread of Weinmannia and Tepualia starting at 10.9 and 11.214Ckyrbp, respectively (Figs 2, 3). Weinmannia is now found in several rainforest communities along a broad latitudinal and altitudinal range in the temperate region of southern Chile. It is a shade-intolerant emergent tree, described as a long-lived pioneer species tolerant of infertile and poorly drained soils15, 16. In addition, Tepualia thickets are known to expand following disturbance in temperate forest communities17. Considering the autoecology of these species, the pollen assemblages between 11.4 and 9.814Ckyrbp indicate a shift from a closed-canopy rainforest community to a woodland that was dominated by species favoured by local disturbance. This vegetation change coincides with a prominent increase in charcoal particles at ~1114Ckyrbp (Fig. 3), implying local disturbance by fire. The pollen sites showing the earliest and most severe effects of fire are located near Monte Verde, thought to be the oldest archaeological site in the Americas18. Remains interpreted as partially burned plant and animal artefacts in association with hearths have been cited as evidence for the presence of humans and their use of fire between ~12.5 and 1214Ckyrbp. Possibly, the high climate variability between ~11.2 and 9.914Ckyrbp afforded the conditions that allowed humans to set small-scale fires that altered the vegetation near the Huelmo and Lago Condorito sites.

Our data suggest that climate approaching modern conditions prevailed in the Chilean Lake District between ~13 and 12.214Ckyrbp. This was followed by a general reversal in trend with cooling events at ~12.2 and ~11.414Ckyrbp, and then by subsequent warming at 9.814Ckyrbp. The total temperature depression between ~11.4 and 9.814Ckyrbp was relatively minor (≤3°C), as indicated by the persistence of rainforest vegetation. The timing, direction, and relative magnitude of these events matches the late-glacial record from nearby Lago Mascardi19 (Fig. 1), which indicates retreat of the Monte Tronador ice cap between 13 and 12.414Ckyrbp, a reversal starting at ~12.414Ckyrbp and culminating with glacial readvance between 11.4 and 10.214Ckyrbp. These records from the Andean region of mid-latitude South America show a notable resemblance in timing and structure to palaeoclimate fluctuations recorded in Europe and Greenland. In contrast, Bennett et al.4 found no palynological evidence for climate change during late-glacial time in the Chilean channels (45°–47°S). If this interpretation is correct, their results would imply that: (1) a major climate boundary existed between 42° and 45°S during late-glacial time; or (2) late-glacial climate changes did not reach a critical threshold to trigger discernible vegetation changes in palynological records, and thus the impoverished late-glacial flora of the Chilean channels was insensitive to the magnitude of late-glacial cooling; and/or (3) plant succession, soil development, and migration from glacial refugia were the dominant factors controlling vegetation change in this newly deglaciated region during the critical time period.

Our results suggest that mid-latitude climate in the Southern Hemisphere changed in unison with the North Atlantic region between ~13 and 1014Ckyrbp. This is in contrast with the palaeoclimate signal derived from sediment cores in the South Atlantic Ocean20 and ice cores from interior Antarctica21, where the pattern of climate change is opposite to that found in the Chilean Lake District between ~13 and 1014Ckyrbp. Determining the representativeness of these opposing results on a hemispheric scale has important implications for understanding the climate mechanisms operative during ice ages, because one set of data supports an in-phase interhemispheric linkage in the atmosphere2, 22, 23, whereas the other favours an out-of-phase relationship via a bipolar see-saw in deep ocean circulation24. The solution could well be that synchronous climate changes, propagated in the atmosphere over much of the planet, were counteracted in Antarctica by a bipolar see-saw of thermohaline circulation, whose effects in the Southern Hemisphere were confined to the high southern latitudes.



  1. Denton, G. H. et al. Interhemispheric linkage of paleoclimate during the last glaciation. Geogr. Ann. A 81, 107–155 (1999). | Article |
  2. Lowell, T. V. et al. Interhemispheric correlation of Late Pleistocene glacial events. Science 269, 1541–1549 (1995). | Article | PubMed | ISI | ChemPort |
  3. Zhou, M. & Heusser, C. J. Late-glacial palynology of the Myrtaceae of Chile. Rev. Palaeobot. Palynol. 91, 283–315 (1996). | Article | ISI |
  4. Bennett, K. D., Haberle, S. G. & Lumley, S. H. The last glacial-Holocene transition in Southern Chile. Science 290, 325–328 (2000). | Article | PubMed | ISI | ChemPort |
  5. Denton, G. H. & Hendy, C. H. The age of the Waiho Loop glacial event. Science 271, 668–670 (1995).
  6. Denton, G. H. & Hendy, C. H. Younger Dryas age advance of Franz Josef Glacier in the Southern Alps of New Zealand. Science 264, 1434–1437 (1994). | Article | PubMed | ISI |
  7. Singer, C., Shulmeister, J. & McLea, B. Evidence against a significant Younger Dryas cooling event in New Zealand. Science 281, 812–814 (1998). | Article | PubMed | ISI | ChemPort |
  8. Newnham, R. M. & Lowe, D. J. Fine-resolution pollen record of late-glacial climate reversal from New Zealand. Geology 28, 759–762 (2000). | Article | ISI |
  9. Björk, S. et al. An event stratigraphy for the Last Termination in the North Atlantic region based on the Greenland ice-core record: a proposal by the INTIMATE group. J. Quat. Sci. 13, 281–292 (1998). | ISI |
  10. Villagrán, C. Vegetatiosgeschichtliche und pflanzensoziologische Untersuchungen im Vicente Perez Rosales National park (Chile). Diss. Bot. 54, 1–165 (1980). | ISI |
  11. Páez, M. M., Villagrán, C. & Carrillo, R. Modelo de la dispersión polínica actual en la región templada chileno-argentina de Sudamérica y su relación con el clima y la vegetación. Rev. Chil. Hist. Nat. 67, 417–434 (1994). | ISI |
  12. Mercer, J. H. Glacial history of southernmost South America. Quat. Res. 6, 125–166 (1976). | Article | ISI |
  13. Mercer, J. H. Chilean glacial chronology 20,000 to 11,000 carbon-14 years ago: some global comparisons. Science 172, 1118–1120 (1972).
  14. Porter, S. C. Pleistocene glaciation in the southern lake district of Chile. Quat. Res. 16, 263–292 (1981). | Article | ISI | ChemPort |
  15. Lusk, C. H. Gradient analysis and disturbance history of temperate forests of the coast range summit plateau, Valdivia, Chile. Rev. Chil. Hist. Nat. 69, 401–411 (1996). | ISI |
  16. Lusk, C. H. Long-lived light-demanding emergents in southern temperate forests: the case of Weinmannia trichosperma (Cunoniaceae) in southern Chile. Plant Ecol. 140, 111–115 (1999). | Article | ISI |
  17. Ramírez, C. in Monte Verde, a Late Pleistocene Settlement in Chile Vol. I, Paleoenvironment and Site Context 53–87 (Smithsonian Institution Press, Washington, 1989).
  18. Dillehay, T. Monte Verde, a Late Pleistocene Settlement in Chile. Vol. II, The Archaeological Context and Interpretation (ed. Gould, R. A.) (Smithsonian Institution Press, Washington, 1997).
  19. Ariztegui, D., Bianchi, M. M., Masaferro, J., Lafargue, E. & Niessen, F. Interhemispheric synchrony of late-glacial climatic instability as recorded in proglacial Lake Mascardi, Argentina. J. Quat. Sci. 12, 333–338 (1997). | Article | ISI |
  20. Charles, C. D., Lynch-Stieglitz, J., Ninnemann, U. S. & Fairbanks, R. G. Climate connections between the hemispheres revealed by deep sea sediment core/ice core correlations. Earth Planet. Sci. Lett. 142, 19–27 (1996). | Article | ISI | ChemPort |
  21. Sowers, T. & Bender, M. Climate records covering the last deglaciation. Science 269, 210–214 (1995). | Article | PubMed | ISI | ChemPort |
  22. Broecker, W. S. & Denton, G. H. The role of ocean-atmosphere reorganizations in glacial cycles. Quat. Sci. Rev. 9, 305–341 (1989). | ISI |
  23. Broecker, W. S. Massive iceberg discharges as triggers for global climate change. Nature 972, 421–424 (1994). | Article |
  24. Broecker, W. S. Paleocean circulation during the last deglaciation: A bipolar seesaw? Paleoceanography 13, 119–121 (1998). | Article | ISI |


We thank I. Hajdas, W. Beck and T. Jull for their contribution to the development of the radiocarbon chronology of Canal de la Puntilla and Huelmo sites. This work was supported by the Office of Climate Dynamics of NSF, NOAA, the National Geographic Society, the Geological Society of America, the EPsCOR program of NSF, and a Fondecyt grant.