Early (approximately 99 Mya) to Late (approximately 161 to 145 Mya) Jurassic
Pangaea, the largest landmass in the Earth's history, was nearly bisected by the equator during the Mesozoic Era. This single continent stretched latitudinally across every part of the zonal atmospheric circulation, thereby producing an extraordinary effect on global palaeoclimate (Dubiel et al., 1991; Valdes, 1993; Sellwood, Valdes & Price, 2000; Sellwood & Valdes, 2006). The supercontinent generated a mega-monsoonal atmospheric circulation in full swing during the Early and Middle Jurassic. Such atmospheric circulation led to extreme seasonality (Kutzbach & Gallimore, 1989).
In the Southern Hemisphere, latitudes north of 30°S were influenced by the ITCZ during summer, whereas southwestern South America was dominated by the movement of the Southern Panthalassea ocean semipermanent anticyclone, which shifted to the south during summer, affecting the southwest margins of Gondwana – Patagonia (Fig. 5A).
Figure 5. Latest Jurassic–earliest Cretaceous (A) Palaeolatitude and palaeogeography, and Intertropical convergence Zone (ITCZ) with Panthalassa Ocean Semipermanent Anticyclone during Southern Hemisphere summer (solid line) and winter (dashed line) (adapted from Loope, Rowe & Joeckel, 2001). B, palaeogeographical map of the southwestern Gondwana (adapted from Scherer & Goldberg, 2007). C, D, maps of Kimmeridgian/Tithonian surface wind vectors, 1120 p.p.m. CO2 simulation; (C) December, January, and February and (D) June, July, and August (sensuMoore et al., 1992a).
Download figure to PowerPoint
Moore et al. (1992a, b) used a GCM to obtain two Kimmeridgian/Tithonian (approximately 154.7 to 145.6 Mya) palaeoclimate seasonal simulations, with geologically inferred palaeotopography: one simulation used a CO2 concentration of 280 p.p.m. (pre-industrial level) and the other used 1120 p.p.m.. Increasing the CO2 four-fold warms virtually the entire planet. The greatest warming occurs over the higher latitude oceans and the least over the equatorial and subtropical regions.
During the summer, monsoon lows were centered poleward of the western Tethys Sea, near 35°S. This location is just east of the region of summer maximum temperature. Over the oceans, in middle latitudes, the Panthalassa subtropical high, which was centered near 50°S, produced surface winds that blew from the southeast over Patagonia (Fig. 5C). These winds correlated with lower precipitation than evaporation. In autumn, the Polar highs were located at 60–70°S, poleward of the Tethys Sea (Fig. 5).
The wind pattern over Patagonia during winter was opposite to that during summer (Fig. 5C, D, arrow directions within squares). During winter the Panthalassa subtropical high was centered at near 25°S. South of it, a subpolar low, centered near 60°S, dominated the south Panthalassa Ocean. These high and low pressure systems along with the continental winter monsoon high, which was located poleward of the western Tethys Sea and was centered near 40°S, were responsible for the winds from the northwest (Fig. 5D) advecting wet and warm air over Patagonia. Furthermore, in the middle troposphere the axes of the Southern Hemisphere mid-latitude storm track is located over 60°S. These factors are related to high positive net precipitation where precipitation exceeds evaporation. The model simulated precipitation rate in the region is ≥ 5 mm day−1 (Valdes, 1993) in accordance with the coal localities that are proxies for wet environments (Scotese, 2010).
Moore et al. (1992a) conclude that the Jurassic warmth evidenced by climate proxies may be explained by elevated atmospheric CO2. By contrast, Chandler et al. (1992) suggested the simulation with specified, warm SSTs was in energy balance without high atmospheric CO2, implying that a warm Jurassic climate could have been the product of enhanced poleward heat transport through the ocean. In other words, two feedback mechanisms are presumed to be primarily responsible for the warm climate over Patagonia during the Jurassic: the elimination of sea and land ice that resulted from the warm polar sea surface temperatures (SSTs) and the equatorward shift of Antarctica resulting in a decrease in surface albedo.
Cretaceous (approximately 145 to 65 Mya)
The fragmentation of Pangea began during the Jurassic. In the Late Jurassic to Early Cretaceous, the separation between Gondwana and Laurasia was well in progress, and the South Atlantic Ocean had already developed between those continents. India separated from Madagascar and raced northward on a collision course with Eurasia. North America was connected to Europe, and Australia was still joined to Antarctica. By the Late Cretaceous, the oceans had widened, and India approached the southern margin of Asia (Scotese, 2010).
The Early Cretaceous was a mild ‘Ice House’ world. There was snow and ice during the winter seasons, and cool temperate forests covered the Polar Regions. The Late Cretaceous was instead characterized by super greenhouse intervals of global warmth with ice-free continents. Globally averaged surface temperatures were 6–14 °C higher than at present (Barron, 1983) and the temperature gradient between the poles and the equator was lower than today, that is approximately 50 °C in the Northern Hemisphere and 90 °C in the Southern Hemisphere. The difference is largely a result of adiabatic cooling reflecting the elevation of Antarctica. Frakes (1999) summarized the data on estimates of Cretaceous sea surface and terrestrial temperatures.
There are four differing assumptions concerning Cretaceous temperatures and meridional gradients: (1) tropical SSTs were the same as today, although polar temperatures were warmer (5–8 °C) except when ice was present (0 to −5 °C); (2) the tropics were significantly cooler and mid-latitudes warmer than today; (3) tropical SSTs were 32–34 °C, with polar regions 10–18 °C; (4) tropical SSTs were approximately 42 °C and polar temperatures > 18 °C. Furthermore, different hypothesis can explain the drastic warming and equable high latitudes during supergreenhouse intervals of the Cretaceous and early Cenozoic. On the basis of coupled ocean–atmosphere model simulations of the middle Cretaceous, Poulsen, Gendaszek & Jacob (2003) hypothesized that the formation of an Atlantic gateway could have contributed to the Cretaceous thermal maximum. Kump & Pollard (2008) GCM simulation of Middle Cretaceous using 4 × CO2, from preindustrial atmospheric level, failed to produce the extreme high-latitude warmth implied by temperature proxy data. However, simulations with the combined increases in cloud droplet radii, which mainly affect cloud optical depth, and precipitation efficiency, resulted in a reduction in global cloud cover from 64% to 55% with optically thinner clouds that reduced planetary albedo from 0.30 to 0.24. The ensuing warming was dramatic, both in the tropics and in high latitudes, where warming was augmented by surface albedo feedback of almost vanishing snow and sea-ice cover. Otto-Bliesner, Brady & Shields (2002) altered the models by the inclusion of high-latitude forest thus changing the palaeogeography. These low-albedo forests warmed the high-latitude continents, which then transferred more heat to the high-latitude oceans, impeding sea-ice formation and warming coastal regions.
Floegel (2002) obtained scenarios for the Turonian (approximately 93.5 to 89.3 Mya) palaeogeography by GCM simulations with different orbital configurations, and 1882 p.p.m. CO2 (5 × year 2000 CO2, 7 × preindustrial CO2) concentrations. The simulated atmospheric circulation resulted in a much more complex circulation than that observed today. In this model, the tropical easterlies remain constant and strong throughout the year. However, at higher latitudes, the circulation varies with the seasons. Because of the absence of polar highs during the winter, strong westerly wind belts develop between 50°S and the high pressure zones at 30°S. Another important difference in the Froegel's scenarios lies in the strong trade winds, which developed during each hemispheric winter.
Floegel's (2002) model, in the Southern Hemisphere summer (December to February), shows the polar region under the influence of a low atmospheric pressure system and therefore the westerly winds are weak, variable, and may even reverse direction. Also, the subtropical to polar frontal systems would not exist (Fig. 6). This would result in disruption of the mid and high latitude wind systems. In the winter (June to August), the southern polar region is under the influence of a high pressure system and the westerlies are well developed (Hay, Flögel & Söding, 2005 & Hay, 2008)
Figure 6. Turonian (93.5 Mya) wind speed and pressure at sea-level for (A) December, January, February and (B) June, July, August (m/s and hPa) (© Sascha Floegel, IFM-GEOMAR, Kiel, Germany, pers. comm., sensuFloegel, 2002.).
Download figure to PowerPoint
The Maastrichtian (70.6 to 65.5 Mya) palaeo-wind scenario in Figure 7 was obtained by Bush (1997) using an atmospheric–oceanic GCM, dynamically and thermodynamically coupled, with four times the present-day value of atmospheric CO2, as indicative of Cretaceous levels. Even with these relatively new results, it is necessary to be cautious with the adopted palaeogeographical reconstruction for the Maastrichtian (Ziegler, Scotese & Barrett, 1983), used in the model, with the Drake Passage (DP) opened.
Some GCM simulations of the Cretaceous supergreenhouse also considered the DP opened as boundary conditions (Poulsen, Pollard & White, 2007; Sewall et al., 2007; Kump & Pollard, 2008; Zhou et al., 2008), whereas others considered DP closed (Haywood, Valdes & Markwick, 2004). For example, Bush & Philander (1997) had estimated a later age for the opening of DP, in the range 49 to 17 Mya; for more details and additional citations, see Cavallotto, Violante & Hernández-Molina (2011).
The earliest connection between the Pacific and Atlantic oceans at DP is controversial but important because the gateway opening probably had a profound effect on global circulation and climate (Cavallotto et al., 2011). Sijp & England (2004) studied the DP influence on climate by means of three main model simulation experiments, set up identically with the exception of bathymetry. They kept the DP closed by a land bridge between the Antarctic Peninsula and South America, DP open to a maximum depth of 690 m and DP open at its present day depth, which was modelled as an uninterrupted throughflow at depth of c. 2316 m.
The climate with DP closed is characterized by warmer Southern Hemisphere surface air temperature and little Antarctic ice. An increase in Antarctic sea ice and a cooling the Southern Hemisphere takes place on opening the DP to a shallow depth of 690 m. On fully opening the DP, the climate is mostly similar in the Southern Hemisphere to DP at 690 m.
The Bush's (1997) simulated pattern of annual mean wind stress (Fig. 7) resulted in a pattern similar to that of the present day with only some changes as a consequence of the modified continental–ocean distribution. The annual mean fields show the tropical-subtropical region circulation characterized by the ITCZ and the easterly winds occupying the latitudinal band from 30°S to 30°N. The Southern Hemisphere Westerlies occupy the latitudinal band from 40° to 70°S and the mid-latitude tropospheric jets dominate the atmospheric circulation over Patagonia (Fig. 7). The simulation shows no poleward shift of the mid-latitude westerlies in response to an ice-free planet.
In the Bush & Philander's (1997) simulation, using a coupled atmosphere–ocean GCM with quadrupled CO2, the annual mean Hadley circulation indicates a general reduction in strength of the Southern Hemisphere cells in the Cretaceous, with the south equatorial Hadley cell weakening by 20%. Subsidence in the high southern latitudes decreased dramatically in the Cretaceous in response to the elimination of the Antarctic ice sheet, which, in the present climate, induces strong subsidence poleward of 75°S. The strengths of the middle and high-latitude cells decrease as a consequence. In addition, the separation of the North and South American continents allowed stronger northeasterly trades flowing from the Tethys basin into the Pacific basin, whereas the strength of the south-easterly winds off the western coast of South America remains approximately the same.
Both Bush & Philander's (1997) and Hotinski & Toggweiler's (2003) modelling studies suggest the possibility of intensified surface circulation. The stronger upper tropospheric westerlies and stronger lower tropospheric easterlies over the tropical Pacific suggest a ‘permanent El Niño’ state, as Davies (2006) found by analyzing laminated sediment. The model results showed the global precipitation approximately 10% higher than at present. The mean annual temperatures increased and the amplitude of seasonal cycle in near-surface temperatures diminished, consequently precluding the presence of year-round snow or ice in the simulation. In high latitudes, however, there are regions that seasonally drop below freezing.
Late Palaeocene–early Eocene (approximately 57 to 53 Mya): thermal maximum
By the Late Cretaceous, the oceans had widened, and India approached the southern margin of Asia. Australia, which was attached to Antarctica, began to move rapidly northward. In the Eocene, 55 to 50 Mya, India began to collide with Asia forming the Tibetan plateau and the Himalayas (Scotese, 2010). The interval from the Late Palaeocene to the Early Eocene was one of the warmest time periods in the Earth history with small annual range. Palaeontological proxy data such as coals in the Antarctic Peninsula and Patagonia, where palm trees grew, suggest a ‘greenhouse’ world with ice free high latitudes (Zachos et al., 1993). The Late Palaeocene–Eocene Thermal Maximum (LPTM) was a dramatic, short-term (approximately 170 kya) global warming event that occurred at approximately 55 Mya (Röhl et al., 2007). The LPTM proxy data predict warm (6–12 °C) deep ocean and polar temperatures, near-modern (23–27 °C) tropical SSTs (Crowley & Zachos, 2000), and higher-than-modern greenhouse gas levels (pCO2 400–4000 p.p.m.) (Pearson & Palmer, 2000). No significant ice accumulation existed at high latitudes in either hemisphere. The ice-free state during LPTM was characterized by a much higher annual-mean temperature and a greatly reduced seasonal cycle along with a lower equator-to-pole surface temperature gradient than is known for any other period in the Cenozoic for which data exist (Abbot & Tziperman, 2008). There is substantial evidence that a weakening of the gradient occurred rapidly and synchronously with a perturbation to the global carbon cycle (Huber & Sloan, 1999). Zeebe, Zachos & Dickens (2009) also suggest that, in addition to the direct CO2 forcing, other hitherto unknown processes and/or feedbacks that must have caused a substantial portion of the warming during the Palaeocene–Eocene Thermal Maximum. The LPTM rapid and short lived change in the pole to equator temperature gradient probably involves a rapid atmospheric and ocean circulation reorganization.
In the Huber & Sloan (1999) simulation of the LPTM climate conditions, the surface winds for January (Fig. 8A) presented mid-high latitude easterlies centered on 50° to 60°S and a weak westerly wind belt around 30°S. The presence of mid-high latitude easterlies is not the result of an expansion of the tropical easterlies so much as an expansion of the polar easterlies. This wind scenario, characterized by relatively strong easterlies over Patagonia and a weakened wind system over the Pacific Ocean likely generated air mass advection from the Atlantic inducing convection and precipitation over land. Such conditions are consistent with the suggested rich subtropical vegetation that existed over a large portion of Eocene Patagonia with the presence of megathermal families such as palms, other taxa with broader climatic requirements such as conifers, cycads, and Ginkgoales (Wilf et al., 2005; Iglesias, Artabe & Morel, 2011; Quattrocchio et al., 2011), as well as the coals in the east Patagonia (Scotese, 2010).
Figure 8. The seasonal maps of the surface wind vectors for the Late Palaeocene–Eocene Thermal Maximum (LPTM) case for (A) January and (B) July. Vector length indicates the magnitude of wind speed in proportion to the sample vector (sensuHuber & Sloan, 1999).
Download figure to PowerPoint
The strong high latitude precipitation in the model scenarios (not shown) was a reflection of the warm high latitude SSTs, as a result of the fact that the removed seasonal sea-ice from the polar seas enhanced local evaporation and precipitation. The reversal of the ‘polar cell’ caused by the deep high latitude convection is a fundamental disruption of the normal atmospheric general circulation. During winter time (Fig. 8B), the winds display the same easterlies pattern over Patagonia but weaker than in summer.
Tortonian (approximately 11 to 7 Mya) after the Eocene/Oligocene cooling
The Late Miocene represents a transitional phase between the Eocene greenhouse climate and the Quaternary icehouse situation. Ice had begun accumulating at the South Pole during the Late Eocene (approximately 37.2 to 33.8 Mya) culminating in the Pleistocene ice age. The widespread glaciation of Antarctica and the associated shift toward colder temperatures near the Eocene/Oligocene boundary (approximately 34 Mya) represents one of the most fundamental reorganizations of the global climate system recognized in the geological record (DeConto & Pollard, 2003). Concordantly, the atmospheric circulation pattern also changed (Broccoli & Manabe, 1992) and the climatic conditions in the region of Patagonia were strongly affected.
Several mechanisms were connected with Oligocene cooling, including changes in continental distribution (Barron, 1985), plateau uplift (Kutzbach et al., 1989; Hay et al., 2002), oceanic gateways (Sijp, England & Toggweiler, 2009), and the decrease of atmospheric CO2 (DeConto & Pollard, 2003). Scher & Martin (2006) suggest that the Drake Passage opened before the Tasmanian Gateway, implying the late Eocene establishment of a complete circum–Antarctic pathway. However, note that Cavallotto et al. (2011) considered that the Tasman path opened before. The Antarctic Circumpolar Current thermally isolated Antarctica, leading to the growth of the Antarctic Ice Sheet. Toggweiler & Bjornsson (2000) suggest that this opening reduced the poleward heat transport in the high-latitude Southern Hemisphere cooling high-latitude surface temperatures by several degrees (in the range 0.8–4 °C).
The Late Miocene Tortonian (approximately 11 to 7 Mya) period, was characterized by intensive Antarctic glaciations and the beginning of glaciations in the North Atlantic region. Steppuhn et al. (2006) run an atmospheric general circulation model coupled to a simple mixed-layer ocean model representing only the uppermost layer of the ocean to calculate the palaeo-SST independently. The boundary conditions and the altitudes of the mountain chains had to be adapted to the Tortonian time interval. The height of the Tibetan Plateau was assumed to have reached approximately 50% of its present height. The Control run represents the present-day situation. Model simulations indicate that differences between the Tortonian and Control runs are less significant in the Southern than in the Northern Hemisphere. Cooling occurs over Antarctica appearing more intense during summer than during winter. Thus, the Tortonian seasonality is reduced at high latitudes. Well developed semi-permanent anticyclones occurred over oceans at subtropical latitudes affecting the equatorial boundary of Patagonia.
In the south Pacific anticyclone, the Tortonian run (Fig. 9A) is stronger than the Control run, as can be seen in the differences between Tortonian and Control run (Fig. 9B) and, together with the atmospheric circulation over the Pacific Ocean, they suggest a permanent El-Niño state. The latitudes of the Patagonian region are dominated by the westerlies (Fig. 9A). The easterly arrows over the high-midlatitudes, in the difference between the Tortonian and the present Control run (Fig. 9B), means less westerlies stress than at present.
Figure 9. The mean annual wind stress field corresponding to the (A) Tortonian run and (B) the difference between the Tortonian and the present Control run. The magnitude of the reference arrow is 1 m s−1 for (A) and 0.25 m s−1 for (B) (sensuSteppuhn et al., 2006).
Download figure to PowerPoint
The uplift of mountain ranges and plateaus during the Neogene and the Cenozoic has been considered by some geologists to be either the direct or the indirect cause of the ‘climatic deterioration’ leading to the late Neogene glaciations (Ruddiman et al., 1997). The South American Andes, formed mainly through crustal thickening associated with Cenozoic subduction and convergence between the Nazca Plate and the South American Plate (Folguera et al., 2011).
The significant surface uplift of the northern Andean plateau, on the order of 2.5 ± 1 km, occurred between approximately 10.3 and 6.4 Mya. Bershaw et al. (2010) suggested that, by approximately 8 Mya in the northern Altiplano and 3.6 Mya in the southern Altiplano, both regions had reached high elevation and established a latitudinal rainfall gradient similar to modern conditions. Ramos & Ghiglione (2008) broadly described the Patagonian Andes uplift, which occurred during the Neogene.
The large influence of the Andes on regional climate has led to the speculation that surface uplift caused past climate changes. The uplift of the southern Patagonian Andes forms a pronounced topographic barrier to atmospheric circulation in the Southern Hemisphere Westerlies, and caused one of the most drastic orographic rain shadows on Earth. Carbon isotope data from the pedogenic carbonate samples demonstrate that this led to strong desertification in the eastern foreland and, presumably, strongly increased precipitation rates on the windward western side of the mountains (Blisniuk et al., 2006; Aragón et al., 2011). Poulsen, Ehlers & Insel (2010) studied the influence of the Southern Andes uplift on the South American climate, precipitation and atmospheric circulation, using an isotope-tracking atmospheric general circulation model. The resulting low atmospheric perturbation from three GCM experiments with prescriptions of Andean elevations corresponds to Andean elevations of 250 m (LOW) (Fig. 10A), intermediate height of 50% of contemporary altitude (Fig. 10B), and the modern elevations (Fig. 10C). Surface uplift modifies the low-level circulation. The Andes block the zonal atmospheric flow and intensify the low-level flow and vapor transport via the South American low-level jet which intensifies. Early uplift of the Andes also leads to southeastward displacement of the subtropical high, away from the growing plateau, and the mid-latitude westerlies migrate southward and become stronger.
Figure 10. Simulated summer low-level (800-mbar) circulation for the three uplift scenarios: (A) low elevation (LOW), (B) intermediate elevation (INT) and (C) modern elevation (MOD). Vectors are missing in (C) where the flow intersects high topography. The difference in 800-mb wind speed (m s−1) from the control (MOD) case (LOW–MOD; INT–MOD) with positive (grey shaded) values represent faster flow and negative (dotted) values represent slower flow. The circle indicates the center of the subtropical high; arrows indicate low-level circulation and American low-level jet (SALLJ) (sensuPoulsen et al., 2010).
Download figure to PowerPoint
Pliocene (approximately 5.3 to 2.5 Mya)
The Pliocene spans a time frame from approximately 5.3 to 2.5 Mya. The epoch incorporates the time interval in which the Earth experienced a transition from relatively warm climates to the prevailing cooler climates of the Pleistocene (e.g. Lisiecki & Raymo, 2007).
The Middle Pliocene is the most recent interval of geological time during which global climate was significantly warmer than at present took place (Crowley, 1996). By contrast to other Pre-Quaternary intervals, the boundary conditions for the Pliocene are fairly well-known. The Mid-Pliocene is recent enough that the continents and ocean basins had almost reached their present geographical configuration. The average of the warmest times during the Mid-Pliocene suggests a globally warmer world, in which atmospheric CO2 concentrations were likely higher than the pre-industrial value (Jansen et al., 2007).
The Early to Middle Pliocene temperature structure of the equatorial Pacific Ocean would have been more zonally uniform than it normally is today and more like the modern Pacific Ocean in an El Niño state (Molnar & Cane, 2002). According to Philander & Fedorov (2003), approximately 3 Mya, the thermocline was sufficiently shallow for the winds to bring cold water from below the thermocline to the surface in certain upwelling regions. Feedback mechanisms could account for the amplification of the Earth's response to periodic variations in obliquity of the Earth's axis (at a period of 41 kyr) and could involve ocean–atmosphere interactions of the type associated with El Niño and also mechanisms by which high-latitude surface conditions can influence the depth of the tropical thermocline. Therefore, the Pliocene paradox is the persistence of El Niño conditions, which was the major factor in the warmth of the Early Pliocene. Fedorov et al. (2006) found that the El Niño-like permanent conditions contributed to global warming, first by causing the disappearance of stratus clouds from the eastern equatorial Pacific Ocean, thus lowering the planetary albedo, and second by increasing the atmospheric concentration of water vapor, a powerful greenhouse gas. Brierley et al. (2009) reconstructed the latitudinal distribution of SST approximately 4 Mya, during the early Pliocene. The reconstruction shows that the meridional temperature gradient between the equator and subtropics was greatly reduced, implying a vast poleward expansion of the oceanic tropical warm pool with enormous impacts on the Pliocene climate, including a slowdown of the atmospheric Hadley circulation.
Simulated 850 hPa wind fields (Jiang et al. (2005) show weakening Middle Pliocene Westerlies during summer, whereas, over the Atlantic and Pacific oceans, semi-permanent anticyclones appears to have been strengthened and poleward shifted. During winter in the south Pacific Ocean, the westerlies are strengthened and only slightly reinforced over Patagonia, compared with today.
The Quaternary since 2.5 Mya and the Last Glacial Maximum (LGM; approximately 20 kyr BP)
The warm equable climates were, after the Pliocene, replaced by recurring ice ages. Raymo et al. (1996) attempted to explain this replacement appealing to changes in CO2 in the atmosphere. Estimates of partial pressures of CO2 in the atmosphere of the past, however, show only modest changes during the past 50 Mya (Pearson & Palmer, 2000). The additional cause, suggested by Sloan, Crowley & Pollard (1996), is the reduced heat transport by the Atlantic Ocean. Other forcing mechanisms included the fact that the obliquity cycle became more prominent approximately 3 Mya. Haug & Tiedemann (1998) propose that the progressive increase in obliquity amplitudes was the final trigger for amplification and continuation of the long-term expansion of the Northern Hemisphere ice sheets after the necessary preconditions were met 4.6 to 3.6 Mya by formation of the Isthmus of Panama. Cane & Molnar (2001) instead, proposed that the closing of the Indonesian Seaway, with New Guinea approaching the equator and with much of Halmahera emerging in the last 5 Myr, had more effect on the climates of the Pacific and Indian Oceans than did the closing of the Panamanian isthmus or changes in thermohaline circulation.
As a consequence of these forcing effects and the cyclic variations in eccentricity, axial tilt, and precession of the Earth's orbit, mathematically theorized by Milankovitch (Crowley & North, 1999), an alternating sequence of glacial and interglacial conditions in approximately 100 kyr cycles characterized the second half of the Quaternary. The strongly affected atmospheric circulation displayed patterns that differed significantly between glacial and interglacial periods. These differences may have been similar to those that exist between the LGM at approximately 20 kyr BP and interglacial warm periods represented by the extremely Holocene Optimum.
The LGM is conventionally defined from sea-level records as the most recent interval in Earth history when global ice sheets reached their maximum integrated volume (Mix, Bard & Schneider, 2001; Clark et al., 2009). Information on precipitation and surface temperature distributions over South America during the LGM, as reconstructed from pollen, lake levels, ocean cores, speleothems, and plant microfossils, indicates levels of aridity that may be interpreted as plant available moisture, precipitation minus evaporation, or some other related variable (Cusminsky et al., 2011). Changes that occurred through the LGM and the Holocene over South America are widely described in Vimeux, Sylvestre & Khodri (2009).
Patagonia was evidently affected by changes in the westerlies and in the position of the southern border of the semi permanent Atlantic and Pacific anticyclones. These changes, which took place during the LGM and the glacial-interglacial transition, are however still poorly understood. The hypothesis that the westerlies shifted north during the LGM has been inferred from terrestrial palaeoecological and marine records off Chile (Heusser, 1989; Stuut & Lamy, 2004). In response to Heusser (1989); Markgraf (1989) proposed that the westerlies shifted to high latitude, giving rise to a long-standing controversy regarding the interpretation of pollen records in Patagonia and the westerlies behavior (Markgraf et al., 1992). However, most of the analyses by proxy data suggest an equatorward shift and intensification of this atmospheric circulation system during the LGM (Lamy et al., 1998, 1999, 2001; Moreno & León, 2003; Valero-Garcés et al., 2005).
The climate modelling results are ambiguous and somewhat contradictory. Although simulations by Kitoh, Murakami & Kiode (2001) and Shin et al. (2003) indicate a poleward shift in the surface westerlies, those by Kim, Flato & Boer (2003) indicate a shift towards the equator, and experiments by Otto-Bliesner et al. (2006) show no shift in the position of the maximum westerlies. Recently, Rojas et al. (2009) analyzed four coupled ocean-atmosphere simulations carried out by the Palaeoclimate Modelling Intercomparison Project Phase 2. They did not find a definite ‘shift’ in the westerly circulation, but suggest a general decrease in surface windspeeds in the Southern Ocean and sub-Antarctic sectors. This decline, in practice, could induce a similar effect as the hypothesized equatorward shift of the southern margin of the southern westerlies.
The Holocene: the Optimum approximately 9000 to 6000 years BP and the last millennium
The Holocene climate probably resulted from predictable changes in the Earth's orbit and from those which caused the end of the LGM. These orbital changes would have had a maximum heating effect in the Northern Hemisphere climate at 9000 years BP when axial tilt was 24° and the nearest approach to the Sun took place during the boreal summer. The changes in the climatic forcings affected the global atmospheric circulation (Crowley & North, 1999; Issar, 2003). In Patagonia, the westerlies shifted poleward toward their modern position as the climate warmed following the LGM (Toggweiler, Russell & Carson, 2006). This atmospheric change produced dried conditions (i.e. decreased precipitations and lake levels) in northern Patagonia and central Chile (Galloway, Markgraf & Bradbury, 1988; Heusser, 1990; Villagrán & Varela, 1990; Villa-Martínez & Villagrán, 1997; Lamy et al., 2001; Jenny, Wilhelm & Valero-Garcés, 2003).
Wagner et al. (2007) model simulation for the Middle-Holocene period between 7.0 and 4.5 kyr BP suggest negative pressure anomalies increased with latitude during summer leading to stronger westerlies over Patagonia than the pre-industrial climate (Fig. 11, left). Winter conditions were very different from those in the summer during this period (Fig. 11, right). These were characterized by a weakening of the zonal winds and reduced gradients in the polar frontal jet as indicated by the positive pressure anomalies located at 60°S over the southwestern part of the South American southern tip.
Figure 11. The sea level pressure (SLP, in hPa) difference between the mid-Holocene and the pre-industrial climate for December, January, February (DJF, left) and June, July, August (JJA, right) (sensuWagner et al., 2007).
Download figure to PowerPoint
Renssen et al. (2005) examined potential climate changes through the Holocene using a coupled atmosphere, sea ice, ocean, and vegetation model to simulate atmospheric circulation variables for Southern Hemisphere mid-high latitudes from 9000 years BP to the present conditions.
Their results indicate that, over the Southern Ocean, the Early and Middle-Holocene temperatures were higher than at present in all seasons (Renssen et al., 2005). The wind speed average of over 45–60°S and the meridional temperature gradient between 23–30°S and 50–60°S (Fig. 12) were closely linked, showing stronger westerlies with an increased meridional temperature gradient. The westerlies weakened in winter from 9 to 5 kyr BP, and then remained nearly constant up to the last 2 kyr and increased weakly in the last millennium. In the other seasons, and especially for summer, the wind had opposite behavior.
Figure 12. Simulated anomalies time series, through the last 9000 years, as deviations from the preindustrial mean (1000/250 years BP) for (A) wind speed (m s−1) over 45–60°S and (B) meridional gradient (°C), 20–30°S minus 50–60°S (sensuRenssen et al., 2005).
Download figure to PowerPoint
The last millenium
Two opposing climate anomalies took place during the last millenium, the so-called Medieval Warm Period (MWP, approximately 9th to 13th Centuries AD, Lamb, 1982; Jansen et al., 2007) and the Little Ice Age (LIA) from approximately 1450 AD to 1890 AD. The LIA comprised three main cold stages, the first beginning at approximately 1650 AD, the second at 1770 AD, and the third at 1850 AD. They were separated by intervals of slight warming. The cooler temperatures were caused by a combination of large volcanic eruptions, less solar activity during the Spörer (1460–1550), Maunder (1645–1715), and Dalton (1790–1830) minima and the inherent climate variability (Crowley & North, 1999; Jansen et al., 2007).
Evidence of MWP and LIA effects over Patagonia has been recovered from different localities, including Laguna Aculeo in Central Chile (Jenny et al., 2002; von Gunten et al., 2009); the Jacaf Channel of Chilean Northern Patagonia (Rebolledo et al., 2008; Sepulveda et al., 2009) Laguna Potrok Aike (Zolitschka et al., 2004; Haberzettl et al., 2005); and Valle de Andorra, Ushuaia (Mauquoy et al., 2004). Glacier retractions following the LIA have been documented for the Laguna San Rafael and the glacier of the same name (Araneda et al., 2007), for the North Patagonian Icefields (Hielo Patagonico Norte; Harrison, Winchester & Glasser, 2007), for the Proglacial Lago Frías at Monte Tronador in northern Patagonia (Rabassa, Rubulis & Suárez, 1979; Villalba et al., 1990), and for the Glaciar Río Manso, also at Monte Tronador (Masiokas et al., 2010), among others.
The atmospheric circulation at this time appears to have been characterized by an increase of the meridional component at high latitudes of the Southern Hemisphere (Kreutz et al., 1997) and by a cooling at the South Pole, resulting in stronger zonal westerlies (Mosley-Thompson et al., 1990). Palaeoecological reconstruction and model simulation for LIA in southern Patagonia show that the precipitation fluctuation can be attributed to latitudinal northward shift of the westerlies (Mayr et al., 2007; Meyer & Wagner, 2008). Recently, Maenza & Compagnucci (2010) obtained scenarios of the anomalies between the Maunder Minimum and the last two decades of the 20th Century. The results show the zonal wind component increased over Central and Northern Patagonia and decreased in the southern tip during the warm semester, whereas, during the colder semester, the wind increased all over Patagonia, reaching a maximum over the southern tip. Varma et al. (2010) used a global climate model and proxies reflecting changes in the position of the westerlies, to infer that, centennial-scale periods of lower solar activity caused equatorward shifts of the westerlies during the past 3000 years.
At present time, the climatic anomalies in northern Patagonia could be related to El Niño (Garreaud et al., 2009) and the whole region appears to be affected by the Southern Annular Mode. This mode is the most important pattern of the atmospheric circulation variability for the middle and high latitudes in the Southern Hemisphere.
The atmospheric circulation at surface over the southern cone of South America corresponds to the behavior synthetically described above (Fig. 3). Among the daily atmospheric circulation types obtained by Compagnucci & Salles (1997) the mean field or basic flow, with the westerlies prevailing over Patagonia, explains more than 50% of the total variance. The remaining variance is explained by different synoptic features representing an increased meridional component (e.g. passage of low and high pressure systems) relative to the basic flow. The variance explained by the basic flow during winters of El Niño years is lower than during winters of La Niña years (Compagnucci & Vargas, 1998).
The westerly component of the low-level flow ahead of the cold front is largely blocked by the Andes, producing a northerly jet along the coast of south-central Chile. This results in a drastic precipitation change between the sides of the Andean Cordillera, affecting temperature as well (Garreaud, 2009; Garreaud et al., 2009). On the basis of instrumental observations along with downscale large-scale signals and up-scale local environmental changes, Garreaud, Lopez & Rojas (2010) described how the lower level (850 hPa) zonal flow (U) strongly modulates precipitation (pp) and surface air temperature (T) (Fig. 13). The mean annual value of zonal wind at 850 hPa correlation coefficients with precipitation (Fig. 13A) and temperature (Fig. 13B) show that the year-to-year changes in the zonal flow are strongly and positively (negatively) correlated with annual mean precipitation changes to the west (east) to the Andes. The seasonal variation of the precipitation dependence to the zonal wind appears to be weak. Furthermore, the surface air temperature relationship exhibits seasonal and geographical dependence. Stronger than normal westerlies flow are associated with milder than normal winters in western Patagonia but colder than normal summers in eastern Patagonia.
Figure 13. Mean annual values of the correlation coefficients between the zonal wind component at 850 hPa and (A) the total precipitation and (B) the surface air temperature for the second half of the 20th Century [obtained from R. D. Garreaud, pers. comm.; sensuGarreaud et al. (2010)].
Download figure to PowerPoint