El Niño-Southern Oscillation signal in the world`s highest

Palaeogeography, Palaeoclimatology, Palaeoecology 281 (2009) 309–319
Contents lists available at ScienceDirect
Palaeogeography, Palaeoclimatology, Palaeoecology
j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / p a l a e o
El Niño-Southern Oscillation signal in the world's highest-elevation tree-ring
chronologies from the Altiplano, Central Andes
Duncan A. Christie a,b,⁎, Antonio Lara a,b, Jonathan Barichivich a, Ricardo Villalba c,
Mariano S. Morales c, Emilio Cuq a
a
b
c
Laboratorio de Dendrocronología, Facultad de Ciencias Forestales, Universidad Austral de Chile, Casilla 567, Valdivia, Chile
Forest Ecosystem Services under Climatic Fluctuations (Forecos), Universidad Austral de Chile, Valdivia, Chile
Departamento de Dendrocronología e Historia Ambiental, Instituto Argentino de Nivología, Glaciología y Ciencias Ambientales, IANIGLA, C.C. 330, 5500 Mendoza, Argentina
a r t i c l e
i n f o
Article history:
Received 23 June 2007
Accepted 19 November 2007
Available online 9 September 2008
Keywords:
ENSO variability
Dendroclimatology
Tropical Andes
Highest elevation treeline
Altiplano
Polylepis tarapacana
a b s t r a c t
El Niño-Southern Oscillation (ENSO) is the largest source of inter-annual variability operating in the earth's
climate system with enormous ecological, social, and economic impacts. As the instrumental record of ENSO
variability is limited to the past 150 yr, the search for proxies sensitive to ENSO has become a priority task in
recent years. It is vital to expand the sparse network of annually resolved and exactly dated ENSO-sensitive
proxies to entirely capture the spatial variability of the ENSO coupled ocean-atmospheric phenomenon.
Several tree-ring records have been used to reconstruct past ENSO variability, however, none of them are
from tropical South America. The Polylepis tarapacana woodlands form the world's highest elevation treeline.
Trees grow on the Altiplano plateau between 4000–5200 m elevation (16°–23° S). The climate variability in
this climatically transitional tropical–subtropical region is strongly modulated by ENSO. Two chronologies of
P. tarapacana along the Western Cordillera in the tropical central Andes were analyzed to determine the
strength of the ENSO signal present in these tree-ring records.
The growth of P. tarapacana has a strong common signal amongst trees in a single site and between sites
across the Altiplano. Ring formation is induced by climatic conditions during the previous and current
growing season (December–February), at the time of the strongest ENSO influences on the Peruvian Coast.
Tree growth variations show opposite relationships with climate between consecutive growing seasons.
During the current growing season, ring-width variations show positive and negative relationships with
temperature and precipitation, respectively. Both tree-ring chronologies are significantly correlated with
austral spring–summer (August–February) SST in the Niño3.4 region and show oscillatory modes within the
classical ENSO bandwidth. Our results support the idea that P. tarapacana chronologies from the southcentral tropical Andes provide high-resolution records extremely sensitive to ENSO in the tropical Pacific, and
represent an important component to be considered in future multiproxy ENSO reconstructions.
© 2008 Elsevier B.V. All rights reserved.
1. Introduction
El Niño-Southern Oscillation (ENSO) is one of the most prominent
sources of interannual variability operating in the earth's climate
system (Trenberth and Caron, 2000). ENSO is associated worldwide
with extreme weather conditions such as heavy snowstorms, floods,
droughts and cyclone activities, having large ecological (Holmgren
et al., 2001), social (Bouma et al., 1997) and economic impacts (Chen
et al., 2001). As the instrumental record of ENSO variability is limited
to the past 100–150 yr, the search for proxies sensitive to ENSO has
⁎ Corresponding author. Laboratorio de Dendrocronología, Facultad de Ciencias
Forestales, Universidad Austral de Chile, Casilla 567, Valdivia, Chile. Tel.: +56 63 293791
221566; fax: +56 63 221230.
E-mail address: [email protected] (D.A. Christie).
URL: http://www.dendrocronologia.cl.
0031-0182/$ – see front matter © 2008 Elsevier B.V. All rights reserved.
doi:10.1016/j.palaeo.2007.11.013
become a priority task in the scientific community in recent decades.
Long records of ENSO variability are needed to characterize the nonstationary behavior of ENSO prior to the instrumental period, to
compare previous with recent-century ENSO variability, and to
develop predictive models of ENSO in the future.
A number of proxy records have been used to reconstruct ENSO
variability including documentary archives (Quinn and Neal, 1983;
Ortlieb, 2000), corals (Evans et al., 2002), ice cores (Thompson et al.,
1984), and tree rings (Stahle et al., 1998; D'Arrigo et al., 2005). A few
multiproxy-based reconstructions have also been conducted (Mann et al.,
2000; Gergis and Fowler, 2006). However, as Stahle et al. (1998) and
Gergis et al. (2006) indicate, the network of ENSO-sensitive proxies is still
in its infancy with regards to its temporal length, exactly annual datingresolution, and spatial distribution, particularly within South America.
In order to successfully achieve high-quality ENSO reconstructions
in the future, it is a priority to increase the geographical distribution of
proxy records to include new areas around the tropical Pacific. Tree-
310
D.A. Christie et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 281 (2009) 309–319
rings are one of the best high-resolution proxy climate indicators and
exhibit some of the strongest statistical relationships with instrumental climate records (Jones et al., 1998). However, the search for
ENSO-sensitive tree species and sites is still ongoing. At present,
highly sensitive ENSO tree-ring chronologies have been developed for
Indonesia (Berlage, 1931; D'Arrigo et al., 1994), Thailand (Buckley et
al., 1995), New Zealand (Fowler et al., 2000), and North America
(Stahle et al., 1998; Cleaveland et al., 2003). Although the potential to
infer past ENSO behavior has been explored in tree-ring records from
extratropical South America (Lough and Fritts, 1985; Villalba, 1994),
no ENSO-sensitive tree-ring chronologies have been produced for the
eastern tropical Pacific, where ENSO variability has a direct impact on
the local climate (Garreaud et al., 2009-this issue).
During the past few decades, the study of ENSO influences on the
climate of the tropical Central Andes in South America has become a
major issue for the local and international scientific community due to
the negative effects of El Niño events on water resources in the region
(Aceituno, 1988; Vuille, 1999; Vuille et al., 2000; Garreaud and
Aceituno, 2001; Francou et al., 2003). Glacier mass balance, rain and
snowfall variability in the Central Andes are strongly modulated by
ENSO (Vuille et al., 2000; Francou et al., 2004). Most water for human
consumption, agriculture, and mining come from the Andes, conferring on these mountains the role of regional “Water Towers”
(Viviroli et al., 2003; Messerli et al., 2004). Under the future scenarios
of increasing temperatures (Bradley et al., 2006) and the lack of
knowledge about interactions between global warming and ENSO,
additional information on ENSO behavior under different air temperature regimes is urgently required (Trenberth and Hoar, 1997).
In the tropical central Andes, and particularly in the Western
Cordillera in Bolivia and Chile, are located the world's highest
elevation woodlands formed by Polylepis tarapacana trees (Braun,
1997). This unique species grows up to 5200 m, reaches an age of more
than 600 yr, and dead wood remains on the ground for centuries due
to the arid dominant climate (Argollo et al., 2004) (Fig. 1). These
dendrochronological characteristics, in combination with its particular location in a directly ENSO-influenced tropical region, make P.
tarapacana particularly suitable for reconstructing past ENSO variability. The main goals of this study are 1) to develop two tree-ring
chronologies of P. tarapacana at high elevations along the Western
Andean Cordillera on the Altiplano, 2) to investigate the relationship
between ENSO and the Altiplano local climate, 3) to determine the
strength of ENSO signal on the tree-ring chronologies, and 4) to
evaluate the potential of P. tarapacana proxy records in estimating
past changes in the coupled ocean–atmospheric phenomenon of the
tropical Pacific.
2. Setting and climate of the Altiplano
The tropical central Andes act as a formidable obstacle for the
tropospheric circulation which generates two contrasting regions, the
tropical lowlands to the east and the Pacific coastal deserts to the west
(Garreaud et al., 2003). The Altiplano is a high-altitude arid plateau at
above ~4000 m, on average 300 km wide and located at the southern
sector of the tropical central Andes between the Eastern and Western
Andean Cordillera (15° S–23° S). This high-altitude region separates
the dry Pacific from the wet Atlantic influences (Fig. 2). The Western
Andean Cordillera is characterized by the presence of high mountains
and volcanoes that reach more than 6500 m above sea level (ASL).
The Altiplano is located within the “Arid Diagonal”, a narrow band
of scarce precipitation spanning South America from the east-coast of
Patagonia up to northern Peru (Abraham et al., 2000). The climate of
the Altiplano is characterized by a reduced seasonality in temperature
but a marked seasonality in precipitation with cool–dry winters and
cool–wet summers (Garreaud et al., 2003; Vuille et al., 2000).
Precipitation over the Altiplano is highly related to the upper-air
circulation with an easterly (westerly) zonal flow favoring wet (dry)
Fig. 1. Polylepis tarapacana woodlands at 4750 m on the slope of the Guallatiri volcano
(6071 m) in the Altiplano.
conditions (Garreaud et al., 2003). Summer (DJF) rainfall is of a
convective nature and represents more than 75% of the total annual
precipitation (Aceituno, 1988). Seasonal climate variability depends
on the intensity and location of the Bolivian High, an upper-air high
pressure system which develops in the upper troposphere south and
southeast of the Altiplano (Lenters and Cook, 1997; Garreaud et al.,
2009-this issue). DJF precipitation is driven by the southern migration
of the maximum convection zone across the Amazon Basin and along
the Andes in response to the South American monsoon circulation
(Vera et al., 2006). Precipitation events are associated with localconvective storms resulting from moisture advection from the
continental Amazon Basin into the Altiplano (Aceituno, 1993;
Garreaud, 2000).
Interannual climate variability is mainly related to changes in the
mean zonal wind over the Altiplano largely modulated by sea-surface
temperature (SST) changes in the tropical Pacific (Vuille et al., 2000;
Garreaud and Aceituno, 2001; Bradley et al., 2003). Marked climatic
anomalies occur during the ENSO events (Vuille, 1999; Vuille et al.,
2000; Garreaud and Aceituno, 2001; Bradley et al., 2003). During the
warm phase of ENSO (El Niño), the Bolivian High is displaced
northward leading to reduced precipitation and increased temperatures. Conversely, during the cold ENSO (La Niña) events, the easterly
flow is enhanced resulting in increased cloud-cover and convective
precipitation, especially during afternoons and evenings (Aceituno,
1993; Garreaud and Wallace, 1997).
With regard to current Altiplano temperature trends, instrumentalbased measurements indicate that temperatures have been increasing
at a rate of 0.10 °C–0.11 °C per decade since 1939. This rate of warming
has more than tripled over the last 25 yr (0.32°–0.34 °C per decade),
and temperatures during the 1997–98 El Niño event were the warmest
during the last six decades (Vuille and Bradley, 2000). Predictive
atmospheric circulation models suggest that maximum temperatures
D.A. Christie et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 281 (2009) 309–319
311
Fig. 2. Location of the northern (Volcán Guallatiri 18°28′ S/69°04′ W) and southern (Cerro Granada 22°32`S / 66°35`W) study sites in the Western Andean Cordillera in the Altiplano.
in the lower troposphere in this region will substantially increase
during the next 100 yr (Bradley et al., 2006).
3. Methods
3.1. ENSO and local climate
To asses the relationships between ENSO and climate variability in
our study sites (Fig. 2), we utilized a combination of meteorological
station records across the Altiplano and sea surface temperatures
(SSTs) from the Niño-3.4 region (5° N/5° S, 170°/120° W) (Table 1).
Long, continuous and homogeneous meteorological records in the
central Andes are sparse, especially for temperature. For the northern
sector of our study area, precipitation variations at Charaña, Chucuyo,
Chungará, Cotakotani, Guallatiri, Isla Blanca and Parinacota (Fig. 2; see
also Solíz et al., 2009-this issue) were combined to create a regional
rainfall record. Monthly standard deviations were determined for each
station and then averaged among stations to produce a regional
record. Most temperature records close to the northern tree-ring
chronology were short and discontinuous. However, as the spatial
variability of air temperature anomalies over the Altiplano is not very
large (Vuille et al., 2000), we used the Oruro record for comparison
with SSTs. Short and fragmentary temperature records from Pacollo,
Sajama, Charaña, and Chungará, (Fig. 2) which are closer to the
northern tree-ring site, were pooled together to characterize the
monthly temperature variations shown in Fig. 6. For the southern
chronology, we utilized the long, continuous, precipitation and
temperature records from La Quiaca.
Relationships between Niño-3.4 SST with temperature and
precipitation records were evaluated on a monthly basis using
correlation analysis. The effect of ENSO events on local temperature
and precipitation variations was assessed using Superposed Epoch
analysis – SEA – (Mooney and Duval, 1993). SEA compares precipitation and temperature time series with a list of ENSO events (El Niño
and La Niña years). For each event, a five-year window was considered
which includes 2 yr leading up to and 2 yr following the event. The
five-year windows for all the events were superimposed and averaged
to obtain the mean pattern of temperature and precipitation related to
ENSO events. Confidence limits were calculated using 1000 Monte
Carlo simulations.
We objectively defined ENSO-year events following the criteria
outlined by Trenberth (1997). Consequently, El Niño (La Niña) events
were defined as those showing, from October to March, 5-month
running means of the SST anomalies in the Niño-3.4 region exceeding
±0.4 °C. We choose the October to March season to include the time
of the year which encompasses the peak of ENSO, the season with the
largest influences of tropical Pacific SST on the Altiplano climate and
the time concurrent with P. tarapacana growing season (December to
March). The Southern Hemisphere tree-ring dating convention
(Schulman, 1956) assigns to annual rings the date of the year in
which radial growth begins. Thus, we assigned to ENSO events the
dates of the year in which the corresponding growing season began.
For example, the 5-month running means exceeding + 0.4 °C were
registered from October 1997 to March 1998 and assigned to the 1997
ENSO event. Table 2 shows the list of ENSO events used in this study.
3.2. Tree-ring chronology development
The genus Polylepis comprises 26 species distributed along the
Andes Cordillera from Venezuela to northern Argentina (8° N–32° S).
P. tarapacana grows in the Altiplano between 16° and 23°S (Kessler
and Schmidt-Lebuhn, 2006). We sampled two P. tarapacana populations separated latitudinally by about 500 km, located on the west and
east slopes of the Western Andean Cordillera. The Guallatiri volcano
woodland is in Chile at 4600–4700 m on the western slope of the
Cordillera (18°45′ S 69°10′ W), whereas the Cerro Granadas population is located in Argentina at 4500–4750 m on the opposite slope
Table 1
Meteorological stations in the Altiplano used for comparision with Sea Surface
Temperature (SST) in the tropical Pacific and Polylepis tarapacana tree growth.
Climatic variable
Location
Elevation (m)
Country
Period
SST
Temperature
Oruro
Charaña
Sajama
Pacollo
Chungará
La Quiaca
Precipitation
Charaña
Isla Blanca
Oruro
Cotakotani
Chucuyo
Parinacota
Chungará
Guallatiri
La Quiaca
Niño 3.4 region
–
–
1870–2001a
17°57′ S/67°08′ W
17°35′ S/69°26′ W
18°06′ S/68°53′ W
18°10′S/69°29′ W
18°17′ S/69°07′ W
22°07′ S/65°36′ W
3706
4059
4220
4050
4500
3442
Bolivia
Bolivia
Bolivia
Chile
Chile
Argentina
1953–1998b
1951–1998b
1975–1981b
1978–1993b
1986–1993b
1911–1990c
17°35′ S/69°26′ W
17°36′ S/69°36′ W
17°57′ S/67°08′ W
18°11′ S/69°14′ W
18°13′ S/69°20′ W
18°14′ S/69°12′ W
18°17′ S/69°07′ W
18°30′ S/69°10′ W
22°07′ S/65°36′ W
4059
4500
3706
4450
4200
4390
4500
4280
3442
Bolivia
Chile
Bolivia
Chile
Chile
Chile
Chile
Chile
Argentina
1948–1998b
1970–1989b
1948–1998b
1962–1992b
1960–1994b
1952–1994b
1962–1994b
1969–1994b
1903–1993c
Source: a Hurrell et al. (2008); b Vuille et al. (2000); c Servicio Meteorológico Nacional,
Argentina.
312
D.A. Christie et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 281 (2009) 309–319
(22°32′ S 66°35′ W). In this paper, we will refer to the Guallatiri and
Granada records as the north and south chronologies, respectively
(Fig. 2).
Due to the eccentric patterns of radial growth and the presence of
twisted trunks, we collected wedges and cross-sections from branches
of living individuals and dead wood, respectively. Samples were
prepared following standard dendrochronological techniques as outlined in Stokes and Smiley (1968). Growth rings were visually crossdated to the year of ring formation (Fritts, 1976). As mentioned above,
we followed the Schulman (1956) convention for the Southern
Hemisphere, which assigns dates of annual rings to the year in which
radial growth started. Annual rings were measured under a binocular
stereoscope with a Bannister-type measuring machine (0.001 mm
accuracy) connected to a computer (Robinson and Evans, 1980). The
cross-dating quality was checked with the computer program
COFECHA (Holmes, 1983), which calculates correlation coefficients
between individual tree-ring series as a tool for identification of absent
or false rings.
Ring-width measurements were standardized to remove variability in the time series not related to climate such as tree ageing or
forest disturbances (Cook et al., 1990). Prior to standardization, the
variance of each series was stabilized using a data adaptive power
transformation based on the local mean and standard deviation of
each tree-ring series (Cook and Peters, 1981; Cook and Peters, 1997).
Standardization was accomplished by fitting cubic smoothing splines
with a 50% frequency-response cutoff width equal to 67% of the series
length for each individual tree-ring series. During the standardization
process indices were obtained by differences between power
transformed ring width measurements and fitted splines (Cook and
Peters, 1997).
Tree-ring chronologies were produced with the ARSTAN40c
program (Cook and Krusic, 2006). Residual instead of standard
chronologies were used in our analysis. By using autoregressive
modeling to remove the serial autocorrelation in the original tree-ring
series (Cook et al., 1990), the residual chronologies satisfy the
statistical assumptions required for establishing the statistical
significance of the climate–growth relationships. Residual chronologies also emphasize the high-frequency domain contained in the time
series, which could be related to ENSO (Gilman et al., 1963).
Finally, the quality of the tree-ring chronologies was assessed by
using the statistic Expressed Population Signal (EPS). To calculate EPS,
we use a statistically demanding 10-year window with an overlap of
5 yr between adjacent windows. The EPS measures the strength of the
common signal in a chronology over time and quantifies the degree to
which a particular chronology portrays the hypothetically perfect
chronology (Cook et al., 1990). While there is no level of significance
for EPS per se, values above 0.85 are generally accepted as showing a
Table 2
El Niño and La Niña events for the period 1870–1999 following Trenberth (1997).
El Niño
1876
1877
1880
1884
1885
1887
1888
1895
1896
1899
1900
1902
1904
1905
1911
1913
1914
La Niña
1918
1919
1923
1925
1929
1930
1935
1939
1940
1941
1947
1951
1952
1953
1957
1958
1963
1965
1968
1969
1972
1976
1977
1979
1982
1986
1987
1990
1991
1992
1993
1994
1997
1870
1871
1872
1873
1874
1875
1878
1879
1882
1886
1889
1890
1892
1893
1894
1897
1898
1903
1906
1908
1909
1910
1916
1917
1920
1922
1924
1933
1938
1942
1945
1949
1950
1954
1955
1956
1964
1967
1970
1971
1973
1974
1975
1983
1984
1985
1988
1995
1998
1999
good level of common signal fidelity between trees (Wigley et al.,
1984). If a given period in the tree-ring chronology drops below this
threshold, it should be viewed more critically when trying to make
inferences about climate.
3.3. Tree-rings and climate
To identify the relationships between climatic factors and tree
growth we computed correlation functions between tree-ring width
indices and monthly mean temperature, monthly mean Niño-3.4 SST
and total monthly precipitation for each study site (Blasing et al.,
1984). Tree radial growth is influenced by climatic conditions several
months prior to ring formation, therefore, we examined the climate
growth relationships over the period starting with June from the prior
growing season to May in the current year of ring formation (Fritts,
1976). The effects of warm (El Niño) and cold (La Niña) ENSO events
on tree growth were evaluated using the SEA.
For the common period between the tree-ring chronologies, their
frequency domains were determined using Maximum Entropy
Spectral Analysis – MESA – (Ghil et al., 2002). In order to expand
the tree-ring chronologies into time-frequency space and localize
intermittent periodicities in the records, we performed Continuous
Wavelet Transform analysis – WT – (Torrence and Compo, 1998).
Additionally, causal relationships in the time-frequency space
between ENSO variability and P. tarapacana growth were examined
employing Cross Wavelet Transform analysis – XWT – (Grinsted et al.,
2004). XWT shows the common power and relative phase between
time-series in the time-frequency space, to determine whether they
present a consistent phase relationship, and consequently whether
they are physically related (Jevrejeva et al., 2003). Monte Carlo
methods were used to assess the statistical significance against red
noise backgrounds (Grinsted et al., 2004).
Finally, to determine the atmospheric and oceanic features more
closely related to P. tarapacana growth we calculated correlation maps
between the tree-ring chronologies and monthly averaged 2.5° × 2.5°
gridded SST, Sea Level Pressure (SLP), and 250 hPa zonal wind (U250)
time series derived from the NCEP reanalysis global database (Kistler
et al., 2001). In order to determine the strongest temporal relationships between P. tarapacana growth and the variables of interest, we
compared several seasonal combinations of SST, SLP, and UA using the
map correlation routines available at the Climate Diagnostics Center,
National Oceanic and Atmospheric Administration (CDC, NOAA).
4. Results and discussion
4.1. ENSO influences on local climate
At our study sites, temperature (T°) and precipitation (Pp) co-vary
in anti-phase on seasonal and annual time-scales. Correlation between
monthly mean temperature and monthly total precipitation is strongly
negative especially during the summer season (DJF), with values up to
−0.71 and − 0.61 during January at the north and south study site,
respectively (P b 0.001; n = 46 and 80, respectively). At annual timescales, the relation between mean T° and total Pp is − 0.65 and −0.6 at
the north and south study site, respectively (P b 0.001; n = 46 and 80,
respectively). These results are consistent with previous studies on
large-scale spatial patterns in the Altiplano using these climatic
parameters (Vuille et al., 2000; Vuille and Keimig, 2004).
At monthly time-scales, relationships between Niño-3.4 SST,
temperature and precipitation are variable. During the austral
summer period Niño-3.4 SST is positively correlated with temperature
and negatively with precipitation (Fig. 3). It is in this season that about
85% of the total annual precipitation occurs at our study sites. The
regional atmospheric circulation in summer favors the advection of
water vapor from the Amazon basin at times when the annual rings of
P. tarapacana formed (Garreaud, 2000).
D.A. Christie et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 281 (2009) 309–319
313
across the Altiplano (Fig. 5; Table 3). The EPS values remain above the
threshold of 0.85, even when a statistically-demanding 10-year window
comparison is used to evaluate the common signal between trees in a
single site. The two P. tarapacana chronologies, at 500 km apart on
opposite sides of the Western Andean Cordillera, are remarkably similar
in their oscillations (r = 0.6; p b 0.001, period 1830–1999). The similarity
in growth patterns between the two P. tarapacana chronologies suggest
that the records share large-scale environmental patterns, as it is shown
in a companion study based on a network of 14 chronologies of P.
tarapacana across the Altiplano (Solíz et al., 2009-this issue).
The north and south chronologies show contrasting responses to
temperature in the previous (negative) and the current (positive)
growing seasons (Fig. 6). This opposite relationships between tree
growth and climate for consecutive growing seasons is difficult to
explain on a biological basis. Argollo et al. (2004) and Morales et al.
(2004) suggested that P. tarapacana growth is largely influenced by the
water balance during the previous growing season. They also indicated
that the reversal in the response of growth to temperature between two
consecutive years is due, to some extent, to the properties of the
temperature variations in the Altiplano, characterized by a persistent
oscillation centered at 2.3 yr. In consequence, if cool (warm) summers
are consistently followed in most cases by warm (cool) summers,
opposite relationships between radial growth of P. tarapacana and
temperature are expected for consecutive growing seasons. The
counter–intuitive relationship between tree growth and climate for
consecutive growing seasons may be also partly due to the semiarid
Fig. 3. Sliding monthly correlations between the Niño-3.4 SST and lag1 month
precipitation and mean temperature from the north and south study sites in the Altiplano
(n = number of correlated pairs). Dashed horizontal lines indicate statistical significance at
the 95% confidence level. The 12 month year was set according to the P. tarapacana growing
year (June–May). The grey vertical bar indicates the tree growing season.
Significant correlations between tropical Pacific SST and summer
temperature on both sides of the Western Cordillera reflect the spatial
uniformity in the temperature variability across the Altiplano (Vuille
et al., 2000). Conversely, the relationships between Niño-3.4 SST and
precipitation are weaker (Fig. 3) and somewhat different between sites.
However, Vuille et al. (2000) who identified the main spatiotemporal
modes of interannual summer precipitation from the Altiplano and
tropical Pacific based on principal components analysis concluded that
precipitation variability in this region is related primarily to ENSO.
At the interannual time-scale, there is a clear modulation of ENSO
events on summer climatic conditions in the Altiplano (Fig. 4).
Temperature anomalies are closely related to ENSO events at both
study sites, whereas precipitation has a more modest response. In
general, El Niño events tend to cause warm and dry summers, whereas
La Niña events are related to cold and rainy summers (Aceituno, 1988;
Vuille et al., 2000; Garreaud and Aceituno, 2001). During El Niño
events (warm-ENSO phase), the westerly-dry flow on the Altiplano
increases in intensity forcing the Bolivian High to northern latitudes.
In consequence, the eastern-wet influences are reduced over the
entire region leading to generally warm-dry summers (Vuille et al.,
1998; Garreaud et al., 2003). Conversely, when there are negative
anomalies in the tropical Pacific SSTs (cold-ENSO phase), the easterly
flow increases the moisture transport from continent producing cold–
cloudy–wet summers dominated by afternoon convective storms
(Aceituno, 1993).
4.2. Tree-growth response to local climate
The tree-ring width variations of P. tarapacana show a strong
common signal between individuals at a single site and between sites
Fig. 4. Superposed epoch analysis (SEA) comparing summer season (DJFM) temperature
(a), and precipitation (b) at the north and south study sites during El Niño and La Niña years.
The X-axis represents a set of 5 yr, from 2 yr before the ENSO event to 2 yr following it. The
dotted horizontal lines represent the 95% confidence levels after 1000 Monte Carlo iterations.
The period from 1953–97 was used for the north, and in the south 1991–89 and 1911–97 were
used for temperature and precipitation, respectively. The ENSO events utilized are listed in
Table 2. About 90% of annual rainfall occurs over this region during the DJFM season.
314
D.A. Christie et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 281 (2009) 309–319
Fig. 5. The prewhitened Polylepis tarapacana ring-width chronologies from Volcán Guallatiri (north) and Cerro Granada (south). The Expressed Population Signal (EPS) statistic
calculated for 10-year windows and sample depth trough time are shown for each chronology. The r value indicates the correlation coefficient between chronologies.
climate and the extreme high elevation of the P. tarapacana woodlands.
However, additional studies are needed to gain insight on this particular
response of P. tarapacana radial growth to temperature in the Altiplano.
In this study, we related P. tarapacana radial growth during the
current growing season with SST in the tropical Pacific. In the
companion study from Solíz et al. (2009-this issue), the potential of
P. tarapacana chronologies to reconstruct spatial precipitation patterns
across the Altiplano is assessed based on the positive relationship
between precipitation and radial growth during the previous growing
season. As summer variations in precipitation and temperature in the
Altiplano are strongly anti-correlated and vary in anti-phase, the radial
growth of P. tarapacana is simultaneously recording seasonal changes
in both climatic parameters. In consequence, tree growth reflects
changes in both temperature and precipitation during summer.
4.3. The ENSO signal on P. tarapacana growth
Monthly tropical Pacific SSTs from the Niño-3.4 region and
P. tarapacana growth are significantly correlated from July previous
to the growing season to March during the current year of ring
formation. This period encompasses the peak ENSO season (November–
February), when warm and cold events typically mature in the
equatorial Pacific (Fig. 6). The strong relationship between
P. tarapacana chronologies and ENSO is consistent with the documented
association between variations in the tropical Pacific SSTs and those of
summer temperature and precipitation across the Altiplano. Correlation
coefficients reach higher significance when SST variability in the
equatorial Pacific precedes temperature and precipitation variations in
the Altiplano by 1–2 months (Fig. 3; Vuille et al., 2000).
El Niño and La Niña events are well recorded in our P. tarapacana
chronologies. Summer warm and cold conditions during El Niño and
La Niña events are related to above- and below-average ring-width
indexes during the current growing season (Fig. 7). However, in the
instrumental records there are some warm-dry (cold-wet) summers
in the Altiplano not directly related to El Niño (La Niña) events
(Garreaud et al., 2003). In consequence, not all the wide (narrow)
rings represented in our chronologies should be inferred as past El
Niño (La Niña) events. During the common period between the
Altiplano tree-rings and Niño-3.4 SST time series (130 yr; Fig. 8), 87%
of the 15 highest tree-ring indices of each chronology correspond to El
Niño events. In the case of the 15 lowest tree-rings indices, 80% and
87% correspond to La Niña events at the north and south study sites,
respectively.
The P. tarapacana chronologies capture most of the high-frequency
variability present in the Niño-3.4 SSTs (Fig. 8). Strong El Niño events
such as 1877, 1888, 1896, 1918, 1925, 1965, 1972, 1987 and 1997 are
coincident in both chronologies with tree-ring indices above the
mean. Strong and sustained El Niño conditions as observed during
1904–05 and 1940–41, which were associated with sustained
droughts in the Altiplano, negatively affected growth of P. tarapacana
resulting in ring widths below the long-term mean. Differences in
chronology responses to El Niño during the 1900, 1912, 1982 events,
registered as higher indices at the north chronology, show the major
sensitivity of the west slope climate of the Western Cordillera to El
Niño induced climate anomalies, a feature previously described by
Garreaud et al. (2003). Strong La Niña events such as those of 1874,
1886, 1892, 1910, 1916, 1942, 1971, 1998 and 1999 are clearly identified
in both chronologies by indices well below the mean. Nevertheless,
this relationship was not observed in all cases such as the 1950 La Niña
event.
Correlation coefficients between the P. tarapacana chronologies
and Niño-3.4 SST are similar, in some cases higher, than those reported
between the Texas–Mexico (Tex–Mex) chronologies and the winter
SOI (Stahle et al., 1998). The Tex–Mex chronologies register one of the
strongest ENSO signal detected in tree-ring data worldwide. Correlations with SOI are significantly higher during January and February,
Table 3
Descriptive statistics for the tree-ring chronologies from Polylepis tarapacana used in
this study.
Site
Correlation within series
North 0.67
South 0.57
Mean sensitivity
Variance in first eigenvector (%)
0.277
0.307
52.2
38.1
D.A. Christie et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 281 (2009) 309–319
315
Fig. 6. Mean monthly temperature (°C) and precipitation (mm) ± 2 standard error and correlations functions based on prewhitened chronologies showing the relationships between
P. tarapacana radial growth and monthly precipitation, temperature, and Niño-3.4 SST from the prior to the current year of ring formation (n and n′ = length in years of the
temperature and precipitation records, respectively. N = number of years used in the correlation analyses). Dashed horizontal lines indicate statistical significance at the 95%
confidence level. Grey vertical bars indicate the previous and current P. tarapacana growing season, respectively.
which temporally coincided with the season of higher correlations
between P. tarapacana chronologies and Niño-3.4 SST (August to
February, Fig. 8). This feature suggest that spatial patterns of past
ENSO variability and teleconnections can be explored using tree-ring
chronologies from North and South America showing similar
seasonal-response to climate anomalies from ENSO.
The significant oscillatory modes in both P. tarapacana chronologies occur within the high frequency domain. MESA indicates that a
large percentage of the variance is centered within the “classical”
ENSO bandwidth with peaks that exceed the 95% confidence limit at
2.1–3.1, 3.6–4.1, and 6.1–6.3 yr. A multi-decadal peak at 36 yr is
registered in the north chronology. The WT, which allows the
expansion of the tree-ring chronologies in the time-frequency space
and therefore can reveal intermittent periodicities, shows similar
spectral peaks to those identified by MESA (Fig. 9). The WT also
revealed strong temporal similarities within the ENSO band between
the chronologies, and a multi-decadal peak of 36 yr in the interval
1890–1930 for the north chronology. In the WT of the chronologies,
the oscillatory periods capturing the largest variance are in agreement
with the significant common spectral power between the chronologies and the instrumental Niño-3.4 SST revealed by the XWT (Fig. 9).
The XWT indicates large covariance between the P. tarapacana
chronologies and the Niño-3.4 SST within the classical ENSO
bandwidth. The north and south XWT are remarkably similar,
demonstrating again the strong regional signature of ENSO on the P.
tarapacana chronologies. Significant intervals within the north and
south XTW show mostly consistent in-phase behavior between the
Niño-3.4 SST variability and P. tarapacana growth, as indicated by the
vector directions (Grinsted et al., 2004). These results are consistent
with the previous SEA and correlations analysis between the
chronologies and Niño-3.4 SST. In general, the results demonstrate
that the tree-ring chronologies contain tropical Pacific forced compo-
nents represented by signals within the ENSO band, and that P.
tarapacana growth is physically linked to tropical Pacific SST variability.
The spatial correlation patterns between the P. tarapacana chronologies and the instrumental gridded SST, U250 and SLP provide insights
on how ENSO-related large-scale climatic variability affects P. tarapacana
Fig. 7. A superposed epoch analysis (SEA) is illustrating comparing tree-ring index
departures for northern and southern P. tarapacana residual (prewhitened) chronologies during El Niño and La Niña years for the period 1870–1999. The ENSO events
utilized are listed in Table 2. The X-axis represents a set of 5 yr from 2 yr before to 2 yr
following the ENSO event. We use as ENSO event years 50 El Niño and 50 La Niña
corresponding to the 1870–1999 period according to Trenberth (1997). The dotted
horizontal lines represent the 95% confidence levels resulting from 1000 Monte Carlo
iterations. The large (small) tree-ring indices during El Niño (La Niña) years are
significant different at the 95% confidence level for both chronologies.
316
D.A. Christie et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 281 (2009) 309–319
Fig. 8. Comparison between the northern and southern prewhitened P. tarapacana chronologies from the Altiplano and the instrumental 5-month running mean of monthly SST
anomalies for the Niño-3.4 region from 1870 to 1999. Black (grey) shading shows tree-ring indices above (below) the long-term mean, and for the Niño-3.4 SST record the + (−)
0.4 °C threshold indicating El Niño (La Niña) events according to Trenberth (1997). Correlation coefficients between the tree-ring chronologies and the Niño-3.4 SST anomalies
(August to February) are indicated.
growth in both study sites (Fig. 10). These spatial patterns mostly
resemble El Niño-like conditions, supporting the relationships between
ENSO conditions and P. tarapacana growth. The spatial correlation fields
for August–February SST show the classical ENSO pattern with warmer
anomalies in the central and eastern equatorial Pacific related to P.
tarapacana above-average growth. With regard to U250, the patterns are
characterized by a reduction of the westerly flow at 250 hPa in a narrow
equatorial band over the central Pacific and an enhancement between
10°–25° across the Pacific domain, which in turn reduces the weteasterly influences over the Altiplano during El Niño events (Vuille et al.,
2000). Finally, the spatial structure of the December–February SLP is
characterized by the well-known dipole over the eastern equatorial
Fig. 9. The wavelet power spectrum (Morlet) of the northern and southern tree-ring chronologies (WT), and the cross-wavelet transform (XWT) between each tree-ring chronology
and Niño-3.4 SST (August–February period). Black thick contours indicate the 95% significance level using the red noise model, and the cone of influence is shown as a lighter shade.
Vectors indicate the relative phase relationship between the Niño-3.4 SST and the tree-ring chronologies (a horizontal arrow pointing right and left implies in-phase and anti-phase
relationships respectively, and pointing up means the second series lag the first by 90°. All significant sections in the XTW show in-phase relationships between the Niño-3.4 SST and
P. tarapacana chronologies.
D.A. Christie et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 281 (2009) 309–319
Pacific and the western Pacific-eastern Indian Ocean (the tropical Warm
Pool). The spatial correlation patterns between SST, U250 and SLP and
the tree-ring chronologies are consistent with previous SEA and
correlation analyses between the tree-ring records and Niño-3.4 SSTs
(Figs. 7 and 8), which indicate that the P. tarapacana growth is
influenced by atmospheric circulation modulated by the tropical Pacific.
5. Conclusions and prospects
We have identified the ENSO signal in the world's highest elevation
P. tarapacana tree-ring chronologies. The chronologies have been
developed in the tropical south-central Andes, a region where the
interannual climatic variability is strongly modulated by ENSO. The
growth of P. tarapacana is influenced by the December–February
anomalies in temperature and precipitation, which in turn are
strongly connected with the ENSO cycles. The P. tarapacana tree-ring
records show a large common signal within individuals in a single site,
and between sites across the Altiplano. These records are significantly
correlated with instrumental ENSO-related records such as the Niño-
317
3.4 SST, and they capture the dominant spectral oscillations within the
classical ENSO bandwidth.
P. tarapacana is a long-lived tree species whose individuals reach
more than 600 yr of age. Abundant death and sub-fossil material are
presently being used to extend the chronologies for the past 1000 yr.
As P. tarapacana grows 800 km along the Andes from 16°S to 23°S, it
offers the unique opportunity to develop a large multi-century
network of tree-ring chronologies all above 4000 m in the tropics
(Argollo et al., 2004; Solíz et al., 2009-this issue).
Our results indicate that the P. tarapacana chronologies from the
Altiplano represent high-resolution proxies extremely sensitive to
ENSO in the Pacific basin. These new ENSO-sensitive tree-ring records
from the tropical Andes offer a unique opportunity to geographically
expand the sparse network of annually resolved and exactly dated
ENSO proxies. Our records should be considered in the development
of new multi-proxy ENSO reconstructions in order to adequately
capture the past spatial variability of this coupled ocean–atmospheric
phenomenon. More research is needed to gain insights into the
complexity of the climate–P. tarapacana growth relationships, and to
Fig. 10. Spatial correlation fields between the P. tarapacana prewhitened chronologies and 2.5° × 2.5° gridded monthly averaged December–February 250 hPa zonal wind (U250),
December–February Sea Level Pressure (SLP), and August–February Sea Surface Temperature (SST). U250, SLP and SST data for the interval 1949–1999 are from NCEP–NCAR
reanalysis. Study sites are indicated by red dots.
318
D.A. Christie et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 281 (2009) 309–319
increase the spatial distribution and temporal extent of the tree-ring
records which will provide consistent patterns of climate variability in
the Altiplano and their teleconnections worldwide.
Acknowledgements
This work was carried out with the aid of a grant from the InterAmerican Institute for Global Change Research (IAI) CRN I # 03 and CRN
II # 2047 which is supported by the US National Science Foundation
(Grant GEO-0452325), the ICM project P04-065-F Chilean Ministry of
Planning, the Argentinean Agency for Promotion of Science (PICTR 123)
and CONICET. The first author was also supported by a doctoral
scholarship from the Chilean Science Research Council (CONICYT). We
are grateful to M. Vuille from the U. of Massachusetts for providing
meteorological data records, D. Shea from the NCAR Climate and Global
Dynamics Division for the Niño-3.4 region SST record extracted from
Hurrell et al. (2008), and P. Štěpánek and M. Masiokas for AnClim
program and RespoSum_Avg time series analysis routine, respectively. J.
P. Francois, G. Lara and J. Moya are thanked for “high altitude” field
assistance. We acknowledge the comments from D. Stahle, R. Garreaud
and one anonymous reviewer which significantly improved this manuscript, and CONAF (Chilean Forest Service) for permission to collect P.
tarapacana samples in Lauca National Park-UNESCO Biosphere Reserve.
References
Abraham, E.M., Garleff, K., Liebricht, H., Regairaz, A.C., Schaebitz, F., Squeo, F.A., Stingl, H.,
Villagrán, C., 2000. Geomorphology and paleoecology of the arid diagonal in
southern South America. Z. Angew. Geol. SH1, 55–61.
Aceituno, P., 1988. On the functioning of the Southern Oscillation in the South American
sector. Part I: Surface climate. Mon. Weather Rev. 116, 505–524.
Aceituno, P., 1993. Elementos del clima en el Altiplano Sudamericano. Rev. Geofís. 44,
37–55.
Argollo, M., Solíz, C., Villalba, R., 2004. Potencialidad dendrocronológica de Polylepis
tarapacana en los Andes centrales de Bolivia. Ecol. Boliv. 39, 5–24.
Berlage, H., 1931. On the relationship between thickness of tree rings of Djati and rainfall
on Java. Tectona 24, 939–953.
Blasing, T.J., Solomon, A.M., Duvick, D.N., 1984. Response functions revisited. Tree-Ring
Bull. 44, 1–15.
Bouma, M.J., Kovats, R.S., Goubert, S.A., Cox, J.S.H., Haines, A., 1997. Global assessment of
El Niño's disaster burden. Lancet 350, 1435–1438.
Bradley, R.S., Vuille, M., Hardy, D., Thompson, L.G., 2003. Low latitude ice cores record Pacific
sea surface temperatures. Geophys. Res. Lett. 30, 1174. doi:10.1029/2002GL016546.
Bradley, R.S., Vuille, M., Diaz, H.F., Vergara, W., 2006. Threats to water supplies in the
tropical Andes. Science 312, 1755–1756.
Braun, G., 1997. The use of digital methods in assessing forest patterns in an Andean
environment: the Polylepis example. Mt. Res. Dev. 17, 253–262.
Buckley, B.M., Barbetti, M., Watanasak, M., D'Arrigo, R., Boonchirdchoo, S., Sarutanon, S.,
1995. Dendrochronological investigations in Thailand. IAWA J. 16, 393–409.
Chen, C., McCarl, B., Adams, R., 2001. Economic implications of potential ENSO
frequency and strength shifts. Clim. Change 49, 147–159.
Cleaveland, M., Stahle, D., Therrell, M., Villanueva-Diaz, J., Burns, B., 2003. Tree-ring
reconstructed winter precipitation and tropical teleconnections in Durango, Mexico.
Clim. Change 59, 369–388.
Cook, E.R., Peters, K., 1981. The smoothing spline: a new approach standardizing forest
interior tree-ring series for dendroclimatical studies. Tree-Ring Bull. 41, 45–53.
Cook, E.R., Peters, K., 1997. Calculating unbiased tree-ring indices for the study of climatic
and environmental change. Holocene 7, 361–370.
Cook, E.R., Krusic, P.J., 2006. Program ARSTAN: a tree-ring standardization program
based on detrending and autoregressive time series modeling, with interactive
graphics. Tree-Ring Laboratory, Lamont Doherty Earth Observatory, Columbia Univ.,
Palisades, N.Y.
Cook, E.R., Briffa, K., Shiyatov, S., Mazepa, V., Jones, P.D., 1990. Data analysis. In: Cook, E.R.,
Kairiukstis, L.A. (Eds.), Methods of dendrochronology: applications in the environmental sciences. Kluwer, Dordrecht, pp. 97–162.
D'Arrigo, R.D., Jacoby, G.C., Krusic, P.J., 1994. Progress in dendroclimatic studies in
Indonesia. Terrestr., Atm. Oceanogr. Sci. 5, 349–363.
D'Arrigo, R.D., Cook, E.R., Wilson, R.J., Allan, R., Mann, M.E., 2005. On the variability of
ENSO over the past six centuries. Geophys. Res. Lett. 32. doi:10.1029/2004GL022055.
Evans, M., Kaplan, A., Cane, M., 2002. Pacific sea surface temperature field reconstruction from coral δ18O data using reduced space objective analysis. Paleoceanography
17, 7/1-7/13.
Fowler, M., Palmer, J., Salinger, J., Ogden, J., 2000. Dendroclimatic interpretation of treerings in Agathis australis (Kauri): 2. Evidence of a significant relationship with
ENSO. J. Roy. Soc. N. Z. 30, 277–292.
Francou, B., Vuille, M., Wagnon, P., Mendoza, J., Sicart, J.E., 2003. Tropical climate change
recorded by glacier in the central Andes during the last decades of the twentieth century:
Chacaltaya, Bolivia, 16°S. Geophys. Res. Lett. 108, 4154. doi:10.1029/2002JD002959.
Francou, B., Vuille, M., Favier, V., Cáceres, B., 2004. New evidence for an ENSO impact on
low latitude glaciers: Antizana 15, Andes of Ecuador, 0°28′S. Geophys. Res. Lett. 109.
doi:10.1029/2003JD004484.
Fritts, H.C., 1976. Tree Rings and Climate. Academic Press, London.
Garreaud, R., 2000. Intraseasonal variability of moisture and rainfall over the South
American Altiplano. Mon. Weather Rev. 128, 3346–3379.
Garreaud, R., Wallace, J.M., 1997. The diurnal march of convective cloudiness over the
Americas. Mon. Weather Rev. 125, 3157–3171.
Garreaud, R., Aceituno, P., 2001. Interannual rainfall variability over the South American
Altiplano. J. Climate 14, 2779–2789.
Garreaud, R., Vuille, M., Clement, C.A., 2003. The climate of the Altiplano: observed
current conditions and mechanism of past changes. Palaeogeogr. Palaeoclimatol.
Palaeoecol. 194, 5–22.
Garreaud, R., Vuille, M., Compagnucci, R., Marengo, J. 2009. Present-day South American
climate. Palaeogeogr. Palaeoclimatol. Palaeoecol. 281, 180–195 (this issue).
Gergis, J., Fowler, A., 2006. How unusual was late twentieth century El Niño-Southern
Oscillation (ENSO)? Assessing evidence from tree-ring, coral, ice-core and
documentary archives, A.D. 1525–2002. Adv. Geosci. 6, 173–179.
Gergis, J., Braganza, K., Fowler, A., Mooney, S., Risbey, J., 2006. Reconstructing El NiñoSouthern Oscillation (ENSO) from high-resolution palaeoarchives. J. Quat. Sci. 21,
707–722.
Ghil, M., Allen, M.R., Dettinger, M.D., Ide, K., Kondrashov, D., Mann, M.E., Robertson, A.W.,
Saunders, A., Tian, Y., Varadi, F., Yiou, P., 2002. Advanced spectral methods for
climatic time series. Rev. Geophys. 40, 1–41.
Gilman, D.L., Fuglister, F.J., Mitchell Jr., J.M., 1963. On the power spectrum of “red noise”.
J. Atm. Sci. 20, 182–184.
Grinsted, A., Moore, J.C., Jevrejeva, S., 2004. Application of the cross wavelet transform
and wavelet coherence to geophysical time series. Nonlinear Proc. Geophys. 11,
561–566.
Holmes, R.L., 1983. Computer-assisted quality control in tree-ring dating and measurements. Tree-Ring Bull. 43, 69–75.
Holmgren, M., Scheffer, M., Ezcurra, E., Gutiérrez, J.R., Mohren, G.M.J., 2001. El Niño
effects on the dynamics of terrestrial ecosystems. Trends Ecol. Evol. 16, 89–94.
Hurrell, J.W., Hack, J.J., Shea, D., Caron, J.M., Rosinski, J., 2008. A new sea surface
temperature and sea ice boundary data set for the Community Atmosphere Model.
J. Climate. 21, 5145–5153.
Jevrejeva, S., Moore, J.C., Grinsted, A., 2003. Influence of the Arctic Oscillation and El
Niño-Southern Oscillation (ENSO) on ice conditions in the Baltic Sea: the wavelet
approach. J. Geophys. Res. 108, 4677. doi:10.1029/2003JD003417.
Jones, P.D., Briffa, K.R., Barnett, T.P., Tett, S.F.B., 1998. High-resolution palaeoclimatic
records for the last millennium: Integration, interpretation and comparison with
general circulation model control run temperatures. The Holocene 8, 455–471.
Kessler, M., Schmidt-Lebuhn, A.N., 2006. Taxonomical and distributional notes on Polylepis
(Rosaceae). Org. Divers. Evol. 1, 1–10.
Kistler, R., Kalnay, E., Collins, W., Saha, S., White, G., Woollen, J., Chelliah, M., Ebisuzaki,
W., Kanamitsu, M., Kousky, V., Van den Dool, H., Jenne, R., Fiorino, M., 2001. The
NCEP–NCAR 50-year reanalysis: monthly means CD-ROM and documentation. Bull.
Am. Meteorol. Soc. 82, 247–267.
Lenters, J.D., Cook, K.H., 1997. On the origin of the Bolivian High and related circulation
features of the South American climate. J. Atm. Sci. 54, 656–677.
Lough, J.M., Fritts, H.C., 1985. The Southern Oscillation and tree rings: 1600–1961. J. Clim.
Appl. Meteorol. 24, 952–966.
Mann, M.E., Bradley, R.S., Hughes, M.K., 2000. Long-term variability in the ENSO and
associated teleconnections. In: Diaz, H.F., Markgraf, V. (Eds.), El Niño and the Southern
Oscillation: Multiscale Variability and Global and Regional Impacts. Cambridge Univ.
Press, New York, pp. 357–412.
Messerli, B., Viviroli, D., Weingartner, R., 2004. Mountains of the world: vulnerable
water towers for the 21st century. Ambio 13, 29–34.
Mooney, C.Z., Duval, R.D., 1993. Bootstrapping: a nonparametric approach to statistical
inference. Sage University Paper series on quantitative applications in the social
sciences. Sage Publications, Newbury Park, CA.
Morales, M., Villalba, R., Grau, R., Paolini, L., 2004. Rainfall-controlled tree growth in
high elevation subtropical treelines. Ecology 85, 3080–3089.
Ortlieb, L., 2000. The documentary historical record of El Niño events in Peru: An update
of the Quinn record (sixteenth through nineteenth centuries). In: Diaz, H.,
Markgraf, V. (Eds.), El Niño and the Southern Oscillation: multiscale variability
and global and regional impacts. Cambridge Univ. Press, NY, pp. 207–295.
Quinn, W.H., Neal, V.T., 1983. Long-term variations in the Southern Oscillation, El Niño,
and the Chilean subtropical rainfall. Fish. Bull. 81, 363–374.
Robinson, E.J., Evans, R., 1980. A microcomputer-based tree ring measuring system.
Tree-Ring Bull. 40, 59–64.
Schulman, E., 1956. Dendroclimatic changes in semiarid America. Univ. of Arizona Press,
Tucson.
Solíz, C., Villalba, R., Argollo, J., Morales, M.S., Christie, D.A., Moya, J., Pacajes, J. 2009.
Spatio-temporal variations in Polylepis tarapacana growth across the Bolivian
Altiplano during the 20th century. Palaeogeogr. Palaeoclimatol. Palaeoecol. 281,
296–308 (this issue).
Stahle, D.W., D'Arrigo, R.D., Krusic, P.J., Cleveland, M.K., Cook, E.R., Allan, R.J., Cole, J.E.,
Dunbar, R.B., Therrel, M.D., Gay, D.A., Moore, M.D., Stokes, M.A., Burns, B.T.,
Villanueva-Diaz, J., Thompson, L.G., 1998. Experimental dendroclimatic reconstruction of the Southern Oscillation. Bull. Am. Meteorol. Soc. 79, 2137–2152.
Stokes, M.A., Smiley, T.L., 1968. An introduction to tree-ring dating. Univ. Chicago Press,
Chicago.
Thompson, L.G., Mosley-Thompson, E., Morales-Arnao, B., 1984. El Niño-Southern
Oscillation events recorded in the stratigraphy of the tropical Quelccaya ice cap,
Peru. Science 226, 50–53.
D.A. Christie et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 281 (2009) 309–319
Torrence, C., Compo, G.P., 1998. A practical guide to wavelet analysis. Bull. Am. Meteorol.
Soc. 78, 61–79.
Trenberth, K.E., 1997. The definition of El Niño. Bull. Am. Meteorol. Soc. 78, 2771–2777.
Trenberth, K.E., Hoar, T.J.,1997. El Niño and climate change. Geophys. Res. Lett. 24, 3057–3060.
Trenberth, K.E., Caron, J.M., 2000. The Southern Oscillation revisited: Sea level
pressures, surface temperatures and precipitation. J. Climate 13, 4358–4365.
Vera, C., Higgins, W., Amador, J., Ambrizzi, T., Garreaud, R., Gochis, D., Gutzler, D.,
Lettenmaier, D., Marengo, J., Mechoso, C.R., Nogues-Paegle, J., Silva Dias, P.L., Zhang,
C., 2006. Towards a unified view of the American monsoon systems. J. Climate 19,
4977–5000.
Villalba, R., 1994. Tree-ring and glacial evidence for the Medieval Warm Epoch and the
Little Ice Age in southern South America. Clim. Change 26, 183–197.
Viviroli, D., Weingartner, R., Messerli, B., 2003. Assessing the hydrological significance of
the world's mountains. Mt. Res. Dev. 23, 32–40.
Vuille, M., 1999. Atmospheric circulation over the Bolivian Altiplano during dry and wet
periods and extreme phases of the Southern Oscillation. Int. J. Climatol. 19,1579–1600.
319
Vuille, M., Bradley, R.S., 2000. Mean annual temperature trends and their vertical
structure in the tropical Andes. Geophys. Res. Lett. 27, 3885–3888.
Vuille, M., Keimig, F., 2004. Interannual variability of summertime convective cloudiness
and precipitation in the central Andes derived from ISCCP-B3 data. J. Climate 17,
3334–3348.
Vuille, M., Hardy, D.R., Braun, C., Keimig, F., Bradley, R., 1998. Atmospheric precipitation
events on Sajama Ice Cap, Bolivia. Geophys. Res. Lett. 103, 11,191–11,204.
Vuille, M., Bradley, R.S., Keimig, F., 2000. Interannual climate variability in the Central
Andes and its relation to tropical Pacific and Atlantic forcing. Geophys. Res. Lett. 105,
12447–12460.
Wigley, T.M.L., Briffa, K.R., Jones, P.D., 1984. On the average value of correlated time
series, with applications in dendroclimatology and hydrometeorology. J. Clim. Appl.
Meteorol. 23, 201–213.