Introduction
The Southern Hemisphere Westerly Winds (SWW) constitute an
important zonal circulation system that dominates the dynamics of Southern
Hemisphere mid-latitude climate. Furthermore, they influence the global ocean
circulation through wind-driven upwelling of deep water in the Southern Ocean
and may play a significant role in the global climate system through the
control of the CO2 budget in the Southern Ocean (Anderson et al.,
2009; Toggweiler et al., 2006; Varma et al., 2011). The understanding of the
variability and the impact of various forcings on the SWW has been discussed
by the study of different proxy and modelling approaches especially at
millennial time scales during the Holocene (e.g. Fletcher and Moreno, 2011;
Kilian and Lamy, 2012; Lamy et al., 2001, 2010; Varma et al., 2012; Whitlock
et al., 2007). Little, however, is known about climatic changes in the
Southern Hemisphere in comparison to the Northern Hemisphere due to the low
density of proxy records, and adequate chronology and sampling resolution to
address environmental changes of the last 2000 years (Moy et al., 2009;
Villalba et al., 2009). Nevertheless, the few available records point towards
significant fluctuations in both temperature and precipitation occurring
during this period (Jones and Mann, 2004; Masiokas et al., 2009; Tonello
et al., 2009). On this time scale orbital boundary conditions only changed
slightly and thus internal variability, solar and volcanic forcing played
a dominant role before the humans became noticeable (Jones and Mann, 2004;
Wilmes et al., 2012).
The “Little Ice Age” (LIA) usually refers to climatic anomalies over the
Northern Hemisphere between the 13th and mid-19th century
(750–150 calyrsBP). The LIA is well documented in northern
Europe and North America, where a huge variety of chronicles, historical
documents, proxy-based reconstructions and also temperature measurements
indicate cooler and wetter conditions (Meyer and Wagner, 2008). Within LIA,
a period with even lower temperatures was the Maunder Minimum (MM;
AD 1645–1715/305–235 calyrsBP). Proxy and modelling studies
point to a prominent influence of solar forcing causing the MM (Eddy, 1976;
Zorita et al., 2004). At the beginning of the last millennium, a period of
warmer conditions, especially over Europe, has been documented: the so-called
Medieval Warm Period (MWP; ca 9th–13th
centuries/1150–750 calyrsBP; Jones et al., 2001; Osborn and
Briffa, 2006). Recently, Neukom et al. (2010, 2011, 2014) points to a number
of climatic variations occurring during the last millennium in Southern South
America. The authors showed that the Southern Hemisphere response to external
forcing may be delayed in approximately two centuries respect to Northern
Hemisphere medieval times with high temperatures and coherent extreme cool
conditions in both hemispheres around AD 1600 (350 calyrsBP).
Pollen records derived from lakes and bogs represent one of the most abundant
paleoclimate archives in South America. Since the pioneering work by Auer
(1933, 1958), many studies have reconstructed the ecological and climatic
history over the Pleistocene and Holocene periods at millennial time scale
(e.g. Heusser and Heusser, 2006; Mancini et al., 2008; Markgraf et al., 2003;
Moreno et al., 2009). There are few pollen based paleoenvironmental
reconstruction with highly-precise chronology in Patagonia for the last
millennia (Fletcher and Moreno, 2012b; Huber and Markgraf, 2003a; Moreno
et al., 2014; Whitlock et al., 2006; Wille et al., 2007). These authors
presented different Patagonia climatic variability scenarios for the last
2000 years. Moy et al. (2009) and Kilian and Lamy (2012) suggest that the
different signal shown in these data set could be attributed to the location
of the records in different ecological environments; the depositional
environment, and local differences in the sensitivity of eastern Andean
vegetation ecotones to changes in precipitation.
Since 2009, new pollen and charcoal records from bog and lakes in northern
and southern Patagonia at the east side of the Andes have been published with
an adequate calibration of pollen assemblages related to modern vegetation
and ecological behaviour (Bamonte and Mancini, 2011; Bamonte et al., 2014;
Echeverria et al., 2014; Iglesias, 2013; Iglesias and Whitlock, 2014;
Iglesias et al., 2012, 2014; Mancini, 2009; Marcos et al., 2012a, b; Sottile
et al., 2012; Sottile, 2014). In this work we improve the chronological
control of some eastern Andean previously published sequences and integrate
pollen and charcoal dataset available east of the Andes to interpret possible
environmental and SWW variability at centennial time scales. Through the
analysis of modern and past hydric balance dynamics we compare these
scenarios with other western Andean SWW sensitive proxy records for the last
2000 years.
Modern eastern Andean Patagonia environmental setting
Climate
Most of Patagonia is dominated by air masses coming from the Pacific Ocean.
The Patagonian region is located between the semipermanent anticyclones of
the Pacific and the Atlantic oceans at approximately 30∘ S and the
subpolar low pressure belt at approximately 60∘ S (Prohaska, 1976).
The strong, constant west winds (westerlies) are dominant across the region.
The seasonal movement of the low and high pressure systems and the
equatorward ocean currents determine the precipitation pattern. During
winter, the subpolar low is more intense. This situation, combined with the
equatorial displacement of the Pacific High Pressure System and with ocean
temperatures that are higher than the continental temperatures, leads to an
increase in precipitation during this season. The northeastern and the
southeastern parts of the region are additionally affected by air masses
coming from the Atlantic Ocean. This Atlantic influence results in a more
even seasonal distribution of precipitation in this part of Patagonia
(Paruelo et al., 1998).
The Andes play a crucial role in determining the climate of Patagonia. The
north–south distribution of the mountains imposes an important barrier for
humid air masses coming from the Pacific Ocean. Most of the water in these
maritime air masses is dropped on the Chilean side, and air becomes hotter
and drier through adiabatic warming as it descends on the Argentine side of
the Andes (Fig. 1a). The westerlies are strongest during austral summer,
peaking between 45 and 55∘ S. During austral winter, the jet
stream moves into subtropical latitudes (its axis is about 30∘ S)
and the low-level westerlies expand equatorward but weaken, particularly at
∼ 50∘ S (Garreaud et al., 2009) (Fig. 1b).
Over Patagonia, the inter-annual correlation between precipitation and zonal
wind at 850 hPa (U850) using annual means exhibits positive values
increasing from Pacific to a maximum along the Chilean coast and the western
slope of the Andes (r(P,U850)∼0.8), a sharp transition just to
the east of the mountain ridge and negative values over the Argentinean
Patagonia. During years with stronger than average westerly flow features
increased precipitation to the west of the Andes and decreased precipitation
over the lowlands to the east. The marked west-east precipitation gradient
over Patagonia is always present but it is slightly less in those years with
weaker than average westerly flow aloft (Garreaud et al., 2013).
When averaged over the year, an ENSO warm event (positive multivariate ENSO
index values) is associated with an overall decrease in the strength of the
wind field and a slight reduction in precipitation in western Patagonia (Moy
et al., 2009). Northern Patagonia exhibits an overall reduction in summer
precipitation and warmer surface air temperature. Of particular relevance is
the frequent occurrence of long-lived, tropospheric deep anticyclonic
anomalies west of the southern tip of South America (below 40∘ S and
centered at 50∘ S, 100∘ W) during El Niño years
(Rutllant and Fuenzalida, 1991). These phenomena favour a northward
displacement of the storm tracks between 33 and 39∘ S (Garreaud and
Aceituno, 2007; Garreaud et al., 2009; Montecinos et al., 2000; Moreno
et al., 2010). During negative multivariate ENSO index values (La
Niña like), the South Pacific anticyclone strengthen and moves southward
(Aceituno, 1988). Similarly, during the positive phase of Southern Annular
Mode (SAM), the SWW intensifies and moves southward, decreasing precipitation
above 48∘ S, favouring the occurrence of forest fires between 39 and
48∘ S (Holz and Veblen, 2011; Mundo et al., 2013; Veblen
et al., 1999; Villalba et al., 2012).
Northern Patagonia vegetation
Eastern Andean communities in Northern Patagonia between 40 and
44∘ S present four major transitions. The first (ca 72∘ W)
from tree/epiphyte species rich Valdivian rainforest to structurally more
simple poor species Nothofagus-dominated forests. This transition
zone coincides approximately with eastern areas of low Andean longitudinal
valleys and where precipitation drops below ca
3000–2500 mmyr-1. A second sharp transition occurs further east
(ca 71.6∘ W) where the continuous Nothofagus forest cover
breaks up giving rise to first patchy but further east more extensive species
rich shrublands composed of heliophyllous species (Iglesias et al., 2014).
This transition occurs where annual precipitation drops below ca
1800 mmyr-1. Finally, a third transition takes place at ca
71–71.2∘ W where easternmost small outpost trees population
(Nothofagus pumilio and Austrocedrus chilensis) intermingle
within the Patagonian steppe matrix. This transition coincides with rainfall
areas below ca 600–800 mmyr-1 (Iglesias et al., 2014). South of
44∘ S, Austrocedrus chilensis disappear and only
Nothofagus tree patches intermingle between steppe patches (Veblen
et al., 1997). Patagonian grass and shrub steppes cover plains and plateaus
eastward ∼70∘ W between 600 and 300 mmyr-1, with
a significant decrease on above-ground vegetation cover following
precipitation gradient (León et al., 1998). Below
300 mmyr-1, Patagonian steppe is replaced by “Monte” shrubland
vegetation (Fig. 1a). Monte shrub communities are arranged as two-phase
mosaic composed by a phase of perennial grasses and shrub-dominated patches
alternating with sparse cover (Bisigato et al., 2009).
Southern Patagonia vegetation
South of 47∘ S, the forest communities impoverished due to the low
temperatures of the growing season. Mixed evergreen-deciduous forest of
Nothofagus betuloides and N. pumilio develop on eastern
Andean lowland areas with annual precipitation above 800 mmyr-1
(Mancini et al., 2008). Between 1000 and 600 mm of annual
precipitation closed deciduous forest of N. pumilio develop from the
tree line to lowlands. These closed forest communities become progressively
open with tree patches of N. pumilio and N. antarctica with
high cover of tall xerophytic shrubs and grass species between 600 and
400 mmyr-1. Eastward between 400 and 200 mmyr-1
a grass steppe covers a narrow and discontinuous strip along the extra-Andean
and the Patagonian plateau and the southeastern tip of the continent
dominated by Festuca spp., cushions plants and isolated shrub
patches (Boelcke et al., 1985; Mancini et al., 2012). At the Patagonian
plateau, the shrub steppe distribution is primarily related to the
availability of water which is actually controlled by unpredictable
precipitation inputs, runoff redistribution and edaphic diversity and is
clearly reflected by the vegetation differences between the plateaus and
valley and ravines (“cañadones”) (Mancini et al., 2012).
Fire regime
The occurrence of wildfires is largely controlled by climatic variability
through its action of modifying fine fuel build up rates and fuel
desiccation. On the easternmost Patagonian communities where steppe
bunchgrasses dominate, fires are limited by fuel amounts and continuity
(Kitzberger, 2012; Sottile et al., 2012). Because fine fuels (grasses) are
highly responsive to precipitation pulses, during rainy growing seasons,
systems that normally do not spread efficiently due to lack of fuel loads
suddenly become more prone for developing large fires (Morgan et al., 2003).
Years with high net primary productivity and rainy springs/summers have also
been highlighted as factors favouring fire occurrence in Monte shrubland
communities (Hardtke, 2014).
Further west in the transition or higher in altitude, in the realm of the
tall Nothofagus forests fine fuels are less important and coarse
fuels that require long drying periods dominate. Here fires are exclusively
associated to strong droughts lasting several months, beginning during the
winter, the time when soils are replenished with water (Kitzberger, 2012).
Whenever dry winter-springs associate with warm summers, wet forests ignite
and spread fire without significant natural fire breaks (Mermoz et al.,
2005). These strong drought events not only produce larger fires but also
more severe events that create conditions that provide less regeneration
opportunities to obligate seed dispersed species (such as N. dombeyi
or N. pumilio; Kitzberger et al., 2005) and more opportunities for
the rapid expansion of resprouting shrubland species. Markgraf and Anderson
(1994) postulated that even though lightning are scarce in southern
Patagonia, they might have been more frequent in the past under different
climatic conditions as fire ignition sources.
Material and methods
In order to reconstruct the past 2000 years of environmental variability on
different landscapes of eastern Andean Patagonia, we selected continuous
pollen and charcoal records from lakes and peat-bogs (Table 1) where data
sets fulfil some qualitative criteria explained as follows:
Dataset availability: pollen records previously published and available at
Neotoma Paleoecology Database (http://www.neotomadb.org) and
pollen/charcoal records from Paleoecology and Palynology Lab database
(UNMdP-IIMyC, CONICET).
Chronology and temporal resolution: proxy data series must have a chronology
based on more than 2 dating for the last 2300 yrs BP. The time
series should at least have a mean sampling resolution of one sample
200 yrs-1.
Also, the sites selected for this work, fulfil more than four criteria of
2 K proxy records for paleoclimate reconstructions according to the
PAGES-2 K criteria (see Supplement, Sect. 3, for details).
We constructed past pollen-based paleohydric balance indices. Main pollen
taxa were considered suggesting above/below hydric availability at every
site, following paleoecological and modern pollen-vegetation calibrations
highlighted on previous published works (Bamonte and Mancini, 2011; Bamonte
et al., 2014; Bianchi and Ariztegui, 2012; Echeverria et al., 2014; Iglesias,
2013; Iglesias and Whitlock, 2014; Iglesias et al., 2012, 2014; Mancini,
2007, 2009; Mancini et al., 2012; Marcos and Mancini, 2012; Marcos et al.,
2012a; Paez et al., 2001; Sottile et al., 2012; Sottile, 2014). Each
paleohydric balance was calculated as the standardized ratio between the sum
(in percentages) of positive hydric availability taxa and the sum of negative
hydric availability taxa (see Supplement, Sect. 4 for details).
Standarization of every ratio was calculated by subtracting the mean and
dividing by the standard deviation. In order to highlight the general trend
of every site index, we apply a locally weighted scatterplot 0.2 smoothing
spline (Cleveland, 1979, 1981) and plotted the 95 % confidence band based
on a 999 bootstrap replicate technique. Modern hydric balance of every site
(Table 1) was compared to paleohydric values in reference to the pollen
samples with an age of ca AD 1900 of every record (preventing possible
changes on pollen spectra related to European settlement).
Also, composite pollen-based indices for Northern and Southern Patagonia were
performed using all dataset available for each region. In order to highlight
the general positive/negative trends of every region index, we applied
a locally weighted 0.2 smoothing spline and plotted the 95 % confidence
band based on a 999 bootstrap replicate technique.
Results
Northern Patagonia
Northern Patagonian pollen based paleohydric balance allow us to reconstruct
past variability especially in terms of seasonality. Assuming that Northern
Patagonian forest and Monte shrubland development, are favoured by
spring-summer rain, positive (negative) values suggest above (below) average
spring-summer precipitation. Nothofagus-Austrocedrus forest
and Nothofagus forest/steppe transitions records present mainly
negative values between 1600 and 750 calyrsBP (Fig. 2b–d). Since
750 calyrsBP, there is a raising trend to positive paleohydric
values peaking ca 250–300 calyrsBP (Fig. 2b–d). On the
contrary, Lake Trébol present the opposite trend during the last
2000 yrs. The comparison of past paleohydric balance to modern hydric
balance suggest >493.7 mmyr-1 in Lake Trébol; <8.60 mmyr-1; <143.2 mmyr-1 in Lake Mosquito and
<268.4 mmyr-1 in Mallín Pollux between 1600 and
750 calyrBP.
Even though Bajo de la Quinta shows mainly negative values, its general
paleohydric trend follows general forest and forest-steppe transition records
behaviour, showing the major paleohydric values after 750 yrs BP
(Fig. 2e). A comparison with modern hydric balance values, suggests Bajo de
la Quinta registered paleohydric values <-516.3 mmyr-1
between 1600 and 750 calyrsBP.
Fire activity presents an opposite behaviour between Andean communities and
Monte shrubland. The highest Charcoal accumulation rates (CHAR) are
registered between 2000 and 750 calyrsBP in Andean sites while
the highest CHAR values in Bajo de la Quinta occur after
500 calyrsBP. Mallín Pollux and Bajo de la Quinta also
register high CHAR values for the last 100 years, which might be related to
European settlements.
Southern Patagonia
Southern Patagonian pollen dataset were classified into two categories (local
and regional, sensu Jacobson and Bradshaw, 1981) in response to the pollen
source area and the variables selected to calculate past paleohydric balance
index. Local dataset category involves pollen records that register past
local vegetation variations. These records present a high relationship
between surrounding deposition site vegetation pollen indicators and modern
pollen samples assemblages (PAA, PAB, MPD, LT, CV). Thus interpretation of
the paleohydric balance index from this sites may be related to changes on
local conditions. Regional category includes records that on recent pollen
samples present higher amounts of pollen types reaching from longer distances
(>3 km southwestward) than pollen from surrounding areas of the
deposition site (CF and RR). Thus, we interpret regional paleohydric balance
indices not as changes on hydric balance in a single site but throughout the
forest-steppe ecotone region.
Southern Patagonian Forest and Forest-steppe ecotone paleohydric indices
present positive values between 2000 and 750 calyrsBP, suggesting
above average water availability on Andean communities (Fig. 3). On the
contrary steppe records present mainly negative values suggesting dry
conditions on extra-andean areas (Fig. 3).
Comparison with modern hydric balance values for pollen records registering
local environmental variability, suggest higher than modern hydric balance
values for PAA, PAB (>104.5 and 67.2 mmyr-1, respectively)
previous to 750 calyrBP. Steppe sites suggest values similar to
modern ones in MPD (∼-163.2 mmyr-1), higher than modern
values in LT (>-146.6 mmyr-1) and lower than modern values in
CV (<-303.7 mmyr-1).
After 750 calyrsBP, Forest and forest-steppe sites exhibit
a decreasing trend in paleohydric balance indices (Fig. 3). PAA and PAB
indices suggest paleohydric values <104.5 and <67.2 mmyr-1, respectively. Steppe sites exhibit the opposite
paleohydric trend toward positive values. The three steppe sites suggest
significant higher than present values of hydric balance (MPD >163.2 mmyr-1; LT >146.6 mmyr-1; CV >-303.3 mmyr-1). Fire activity exhibit synchronous CHAR patterns
especially between 2000–1700 and 750–250 calyrsBP in southern
Patagonian charcoal records.
Discussion
Controls over hydric balance in
Northern and Southern Patagonia
The late Holocene changes in paleohydric balance reconstructed from Northern
Patagonian records could be interpreted in terms of latitudinal variation of
the SWW belt, using the modern latitudinal distribution of precipitation
seasonality over Patagonia (Fig. 1b) as analogous. Thus, when modelling all
Northern Patagonian dataset we perform a composite pollen-based Northern
Patagonia SWW belt latitudinal variation index between 40 and
45∘ S (Fig. 4c). This pollen-based index displays high
precipitation seasonality before 750 calyrsBP. Such a high
seasonality likely suggests a more poleward position of the SWW belt,
reflecting similar to present day precipitation seasonality (Fig. 4).
Nevertheless, Lake Trébol shows high values of paleohydric balance index
before 750 calyrsBP and lower values since 750, around present
day hydric values. These pattern joint to the general trend of most northern
Patagonian paleohydric balance indices, may reflect intense SWW during winter
favouring higher precipitation amounts over areas close to the Andean divide
linked to a steeper west-to east precipitation gradient that soften up to
present condition since 750 calyrsBP.
Since 750 calyrsBP the Northern Patagonia pollen based index
shows a remarkable decrease in precipitation seasonality peaking between 400
and 200 calyrsBP (Fig. 4c). This low seasonality period likely
reflects a northward expansion of the SWW favouring increased spring-summer
precipitation near the Andes. The similar paleohydric balance of Bajo de la
Quinta (Fig. 2e) at the Atlantic coast to those of forest environments
suggest that between 400 and 200 cal yrs BP, Atlantic humid air
masses reached the continent probably under weak SWW (Marcos et al., 2012a,
2014). Therefore we can interpret dominant summer-like conditions in terms of
hydric balance in northern Patagonia between 1600 and
750 calyrsBP and winter-like conditions between 750 and
200 calyrsBP.
During the last 200 calyrsBP, there is a remarkable decrease in
the Northern Patagonia pollen based index suggesting higher than before
precipitation seasonality toward present day conditions between 40 and
45∘ S. Even though pollen spectra might be biased for the last
100 years, the decreasing trend in Northern Patagonia pollen based index,
precedes European arrival (Fig. 4c).
Precipitation seasonality inferences coincide with centennial fire activity
in northern Patagonia. We found an antiphase behaviour of fire occurrence
between western and eastward environments. During southward displacement of
SWW, fire activity increases on forest communities likely related to coarse
fuel desiccation and low biomass availability on eastern Monte shrublands
avoiding fire propagation (Fig. 2e). On the contrary, during periods of
winter-like conditions, fire activity increases on Monte shrublands, likely
related to an increase in biomass favoured by Atlantic Humid air flow masses.
Iglesias and Whitlock (2014) presented northern Patagonia biomass burning
general trends since the last 18 000 calyrsBP and compared them
to environmental and archaeological information. They interpret that
variations in indigenous population densities were not associated with
fluctuations in regional or watershed-scale fire occurrence, suggesting that
climate–vegetation–fire linkages in northern Patagonia evolved with minimal
or very localized human influences before European Settlement (Iglesias and
Whitlock, 2014). On the Atlantic coast, archaeological records suggest high
anthropogenic activity ca 1000 calyrsBP with a decreasing trend
up to present day (Marcos and Ortega, 2014). Thus, patterns of fire activity
increase since ca 500 calyrsBP in Bajo de la Quinta are likely
related to climate variability and lightning sources.
The late Holocene changes in paleohydric balance reconstructed from Southern
Patagonian records could be interpreted in terms of intensity variation of
the SWW belt, using the modern latitudinal distribution of precipitation
seasonality over Patagonia (Fig. 1b) as analogous. These sequences are not
significantly affected by seasonal variability but mainly affected by changes
on SWW intensity (Garreaud et al., 2013). During years with stronger than
average SWW precipitation increased to the west of the Andes and decreased
over the lowlands to the east (Garreaud et al., 2013). Therefore we expect
that hydric balance increases in forest areas (especially those with present
day positive hydric balance values) and decreases in grass steppe
extra-Andean environments. Conversely, the marked west–east precipitation
gradient is slightly less in those years with weaker than average westerly
flow, thus we expect lower than average hydric balance values on forest areas
and higher than average hydric balance values in grass steppe extra-Andean
environments. Atlantic humid air masses probably increase hydric balance
values on steppe records next to the Atlantic coast during periods of weaker
westerlies (Agosta et al., 2015).
Thus, when modelling all Southern Patagonian datasets we perform a composite
pollen-based Southern Patagonia SWW intensity variation index between 48 and
52∘ S (Fig. 4) by considering forest and forest-steppe
ecotone index values and inverse steppe index values. Figure 4 shows the
scatterplot dataset and smothering spline of Local and Regional records from
southern Patagonia. This pollen-based index displays intense SWW before
750 calyrsBP and weaker SWW since 750 calyrsBP,
peaking ca 500–600 calyrsBP (Fig. 4). The Southern Patagonia
index increases slightly toward ca 250–300 calyrsBP suggesting
an intensification pulse of the SWW. Since then, the Southern Patagonia index
values decreases to modern ones, thus we interpret a slight weakening of the
SWW up to modern conditions.
In contrast to Northern Patagonia regional fire behaviour, Southern Patagonia
fire activity trends on forest and steppe communities are synchronous. The
maximum fire activity in southern Patagonia occurs during weaker westerlies
(on steppe environments especially previous to 1600 cal yrs BP,
Fig. 2). Therefore we interpret an antiphase behaviour between northern and
southern forest communities and an inphase behaviour of fire occurrence in
extra-andean steppe and Monte shrublands.
Anthropogenic fires may represent an extra driving factor favouring fire
activity between 1000 and 2000 calyrsBP in southwestern Patagonia
due to the more intense and extensive archaeological signal registered for
this area (Franco et al., 2004). However, fire activity registered between
250 and 750 cal years BP is probably related to natural lightning
sources since archaeological signal decreases during the last
1000 calyrsBP in southwestern Patagonia related to an eastward
population migration (Franco et al., 2004).
Comparison with western Andean precipitation and SWW belt records
The timing of major SWW changes in latitudinal shift and intensity recorded
by the pollen-based Eastern Andean Northern and Southern Patagonia indices
performed in this work at 750 calyrsBP (1200 AD) roughly
corresponds to a major reorganization of the climate system throughout the
world, which is frequently associated to the Little Ice Age originally
described in the Northern Hemisphere. Here, we compare our inferred-SWW
variation during the last 2000 years to western Andean regional precipitation
and SWW reconstructions.
Fletcher and Moreno (2012) studied a pollen and charcoal record from Laguna
San Pedro (38∘ S, Fig. 1) located on the western side of the Andes
and performed a Nothofagus vs. Poaceae (N/P) index to infer
changes in humidity during the last 1500 years. The N/P index shows similar
behaviour to our pollen-based Northern Patagonia SWW index (Fig. 4a). Indeed,
a brief peak on both indices is registered ca 1100/1400 calyrsBP,
suggesting a short period of lower precipitation seasonality under a long
term trend of higher precipitation seasonality at both sides of the Andes
range. The charcoal record from Laguna San Pedro coincides with eastern
Andean Northern Patagonia fire activity during the last 2000 years.
Bertrand et al. (2014) performed a precipitation seasonality index by
analysing past two millennia sedimentation changes at Quitralco fjord
(46∘ S, Fig. 1). The authors suggests a poleward-shifted SWW belt
between 1350 and 750 cal yrs BP, followed by a gradual shift towards
the equator between 750 and 450 calyrsBP, and stabilization in
a sustained northward position between 450 and 0 calyrsBP
(Fig. 4b). The most recent return to a slightly poleward shifted SWW recorded
at Quitralco fjord, is in agreement with recent trends observed in
climatological data (Bertrand et al., 2014). The coincidence between
Bertrand's sedimentation based seasonality index and our composite
pollen-based Northern Patagonia SWW belt index supports the reliability of
our Northern Patagonian proposed past environmental and climate variability
scenarios. Similarly, other marine records shows increases on precipitation
of SWW origin between 750 and 200 calyrBP at 41∘ S (Lamy
et al., 2001) and 44∘ S (Sepulveda et al., 2009).
In Southern Patagonia, Lake Potrok Aike (52∘ S) is located where
precipitation is negatively correlated with westerly wind strength
(Habertzettl et al., 2005). These authors inferred increased lake levels
associated to easterly humid flows during weaker westerlies between 490 and
0 calyrsBP. Further south, the MA1 stalagmite record
(53∘ S, Fig. 1) also provides evidence for a decrease in annual
precipitation, and therefore a weakening of the westerlies, since
1000 calyrsBP (Schimpf et al., 2011, Fig. 4d) synchronously with
our composite pollen-based Southern Patagonia SWW belt index. Similarly, the
sediment record from Lago Fagnano (Waldmann et al., 2010; Fig. 1) suggests
a decrease in precipitation of westerly origin, represented by a decrease in
iron supply between 750 and 100 calyrsBP (Fig. 4e). These
independent records and Koffman et al. (2013) interpretations of westerlies
strength throughout changes in the grain-size of dust particles in the WAIS
Divide ice core at Antarctic Peninsula, supports the sensitivity of our
Southern Patagonia SWW belt composite pollen-based index to environmental
variability.
The slight intensification of the SWW belt ca 300 calyrsBP,
coincide with major glaciers advances in southern Patagonia (Aniya, 2013;
Masiokas et al., 2009; Mercer et al., 1982; Strelin et al., 2008; Wenzens,
1999) and a Southern Hemisphere extreme cold period inferred by Neukom
et al. (2014). Therefore the synergic direction of low temperatures and an
increase in hydric balance may have favoured Maunder Minimum glacier advances
in Southern Patagonia.
Changes in SWW belt and possible forcing mechanisms
Our SWW belt reconstruction suggest southward intensified westerlies since ca
1600 calyrsBP including the MCA (1150–750 calyrsBP)
and northward weaker westerlies during LIA (750–150 calyrsBP,
Fig. 4c). During LIA, atmospheric cooling in the Southern Hemisphere would
have caused a northward shift of the SWW and contraction of the Southern
Hemisphere Hadley Cell (Koffman et al., 2013). General circulation model
(GCM) experiments have shown that the latitudinal extent of the Hadley cell
circulation is sensitive to changes in global surface temperatures, with
warmer temperatures causing an expansion of the Hadley cell (Frierson et al.,
2007). These changes in the Hadley cell width are likely driven by shifts in
the latitude where baroclinic eddies begin to occur; as surface temperatures
warm, the transition from baroclinic stability to instability shifts
poleward, driving the eddy-driven Southern Hemisphere storm track southward
(Frierson et al., 2007; Lu et al., 2010). This proposed mechanism implies
that the SWW respond to surface temperature changes on decadal to centennial
timescales (Koffman et al., 2013). The mechanism proposed above differs from
the seesaw-type redistribution of heat between the hemispheres that was
invoked to explain the migration of the SWWB during the last deglaciation
(Anderson et al., 2009; Toggweiler, 2009). This suggests that the SWWB may
respond to different forcing mechanisms at different timescales (Bertrand
et al., 2014). Varma et al. (2011) presented proxy and model evidence that
centennial-scale variability in the position of the SWW is significantly
influence by fluctuations in solar activity during the past 3000 years. They
argued that periods of lower solar activity were associated with annual-mean
northward shifts of the SWW, whereas periods of higher solar activity were
linked to annual-mean poleward displacements of the SWW.
Finally, our results coincide with other inferences predominantly from
sea-surface temperature and modelling data about ENSO activity over the last
1500 years, where during the MCA, La Niña like or weak El Niño
conditions and probably positive SAM dominates in Southern South America
(Graham et al., 2010; Mann et al., 2009; Rein et al., 2004; Seager et al.,
2007). On the contrary, during LIA dominated more intense El Niño like
conditions and negative SAM values (Mann et al., 2009; Rein et al., 2004;
Villalba et al., 2012). The marked decreased in our Northern Patagonia pollen
based index suggesting a southward shift of the SWW belt storm track during
the last decades coincides with modern climate data measurements (Archer and
Caldeira, 2008; Hu and Fu, 2007) linked to the poleward migration of the
descending branch of the Hadley cell (Villalba et al., 2012).