You are on page 1of 11

ORI GI NAL

ARTI CLE
From ice age to modern: a record of
landscape change in an Andean cloud
forest
B. G. Valencia
1
*, D. H. Urrego
1
, M. R. Silman
2
and M. B. Bush
1
I NTRODUCTI ON
The heterogeneity of Andean cloud forest habitats contributes
signicantly to their immense biodiversity and endemism. The
combination of high diversity and habitat loss due to land
clearance for agriculture, pastoralism and forestry has resulted
in these forests being listed as conservation hotspots. Of the 25
conservation hotspots identied by Myers et al. (2000), the
tropical Andes is the one most threatened by projected climate
change (Malcolm et al., 2006). Understanding how these
systems have responded to past climate change and the scale
of pre-historical landscape transformation are important issues
in the development of appropriate conservation policy.
Seminal analyses by Thomas Van der Hammen and his
colleagues (e.g. Van der Hammen & Gonza lez, 1959; Van der
Hammen, 1974; Hooghiemstra & Van der Hammen, 2004)
showed that the forests formed in the High Plains of Bogota ,
Colombia, were replaced by grasslands during glacials. Fossil
pollen records from glaciated landscapes in Ecuador, Peru and
Bolivia (Graf, 1981; Colinvaux et al., 1997; Chepstow-Lusty
et al., 2005; Weng et al., 2006) portray modern vegetation
types assembling in the deglacial period or during the early
1
Department of Biological Sciences, Florida
Institute of Technology, Melbourne, FL, USA,
2
Department of Biology, Wake Forest
University, Winston Salem, NC, USA
*Correspondence: Bryan G. Valencia,
Department of Biological Sciences, Florida
Institute of Technology, 150 West University
Boulevard, Melbourne, FL 32901, USA.
E-mail: bguido@my.t.edu
ABSTRACT
Aim To investigate the palaeoecological changes associated with the last ice age,
subsequent deglaciation and human occupation of the central Andes.
Location Lake Pacucha, Peruvian Andes (133626 S, 731942 W; 3095 m
elevation).
Methods Vegetation assemblages were reconstructed for the last 24 cal. kyr bp
(thousand calibrated
14
C years before present), based on pollen analysis of
sediments from Lake Pacucha. An age model was established using
14
C accelerator
mass spectrometry dates on bulk sediment. Fossil pollen and sedimentological
analyses followed standard methodologies.
Results Puna brava replaced the Andean forest at the elevation of Lake Pacucha
at the Last Glacial Maximum (LGM). Deglaciation proceeded rapidly after 16 cal.
kyr bp, and near-modern vegetation was established by c. 14 cal. kyr bp. The
deglacial was marked by the range expansion of forest taxa as grassland taxa
receded in importance. The mid-Holocene was marked by a lowered lake level but
relatively unchanged vegetation. Quinoa and maize pollen were found in the
latter half of the Holocene.
Main conclusions Temperatures were about 78 C colder than present at
this site during the LGM. The pattern of vegetation change was suggestive of
microrefugial expansion rather than simple upslope migration. The mid-
Holocene droughts were interrupted by rainfall events sufciently frequent to
allow vegetation to survive largely unchanged, despite lowering of the lake level.
Human activity at the lake included a 5500-year history of quinoa cultivation and
3000 years of maize cultivation.
Keywords
Charcoal, fossil pollen, Holocene, human disturbance, Inca, Last Glacial
Maximum, maize, migration, Peruvian Andes, vegetation.
Journal of Biogeography (J. Biogeogr.) (2010) 37, 16371647
2010 Blackwell Publishing Ltd www.blackwellpublishing.com/jbi 1637
doi:10.1111/j.1365-2699.2010.02318.x
Holocene. However, none offers a complete unbroken
sequence from the Last Glacial Maximum (LGM) to the
present. Lake Titicaca (Paduano et al., 2003; Hanselman et al.,
2005; Fritz et al., 2007) provides a continuous record of puna
and super puna that spans four glacial cycles, but cloud forest
never occupied the site. At lower elevations, lakes Consuelo
(Bush et al., 2004; Urrego et al., 2005) and Chaplin (Mayle
et al., 2000) are continuous sequences that reect the down-
slope expansion of cold taxa during the full glacial, but lose the
upper cloud forest elements in the mid-deglacial period,
c. 15 cal. kyr bp (henceforth all ages are expressed as thousand
calibrated
14
C years before present; cal. kyr bp). From these
records a clear understanding has emerged that cloud forest
ecosystems have been subject to migration as glacials lowered
temperatures by c. 58 C in these settings (Van der Hammen,
1974).
A fossil pollen record from Siberia, Bolivia, suggested that
variability in precipitation was the dominant force shaping
Andean ecosystems over the last 28 cal. kyr bp (Mourguiart &
Ledru, 2003). Moreover, precipitation was inuenced by
precessional forcing that induced uctuations in lake level in
the lakes of the Altiplano (Baker et al., 2001a), central Peru
(Seltzer et al., 2002), and (with opposing signature) the lakes
of the High Plains of Bogota (Hooghiemstra et al., 1993). An
analysis of the diatom record of Lake Pacucha, Peru, demon-
strated the presence of a precessional signal that underlay early
Holocene drying c. 169 cal. kyr bp, but that a second period
of low lake levels between c. 8.2 and 6 cal. kyr bp probably had
a different explanation (Hillyer et al., 2009).
The mid-Holocene dry event caused lake levels to fall
throughout the Andes (Abbott et al., 2003). In much of the
Altiplano and the Chilean Andes this period became known as
the Zone of Archaeological Silence (Nu nez et al., 2002) as so
many sites appear to have been abandoned. However, the
complex structure of this event has only recently been resolved
as not being a continuous period of drought but one that was
exceedingly drought-prone, with intervening wet intervals
(Ekdahl et al., 2008; Hillyer et al., 2009). Ecologically, the
difference between a millennial-scale drought versus one
interrupted by intermittent moisture would have been huge
and worthy of exploration.
The greatest single ecological variable in modern Andean
ecosystems is the presence or absence of re. Cloud forests
below the immediate contact with puna seldom burn naturally
(Bush et al., 2005), and even a single re can have transfor-
mative effects (Cochrane & Laurance, 2002; Cochrane &
Barber, 2009). Fire can depress the tree line and totally
eliminate the shrubby transition between forest and grassland
(M. B. Bush, pers. obs). Where humans are actively burning
the landscape, exotic species such as Pinus and Eucalyptus are
often planted. It has been suggested that afforestation with
native species has ancient roots in the Andes and that the
Huari, and later the Inca, may have practised silviculture as
much as 1000 years ago (Chepstow-Lusty & Wineld, 2000).
Polylepis forms a native woodland that is particularly
susceptible to re (Cierjacks et al., 2008). This tree can occur
above the tree line, with some individuals growing at an
elevation of more than 5000 m (Kessler, 1995). However,
Polylepis is most abundant on scree slopes and areas where re
is rare. Whether such a constrained occurrence for this genus is
a purely recent pattern of growth is uncertain. Based on
bioclimatic tolerances and a rich endemic fauna associated
with these woodlands, it has been suggested that Polylepis
woodland may have been a dominant vegetation type across
large expanses of what are now puna grasslands (Ellenberg,
1958; Kessler, 1995; Fjeldsa & Kessler, 1996). Interest in
afforestation with native species as a winwin solution that
provides carbon sequestration and conservation habitat has led
to large-scale planting of Polylepis.
Our research hypotheses to be tested are: (1) during the
LGM plant species grew c. 10001500 m downslope of their
present range, i.e. become puna if burned or Polylepis if
unburned; (2) during deglaciation there was a steady migra-
tion of taxa upslope; (3) the mid-Holocene dry event would
produce a marked change in the ora towards drought-
tolerant taxa; and (4) substantial human impacts are evident
prior to those of the Inca. Here we test these hypotheses using
a fossil pollen and charcoal record that spans the last 24.7 cal.
kyr bp from Lake Pacucha, Peru.
THE SI TE
Lake Pacucha (Fig. 1) lies in an unglaciated section of in
the central valley of the Peruvian Andes (133626 S,
731942 W; 3095 m elevation). The origin of the lake is
not known, but we infer it to be a solution hollow. The lake is
c. 2.4 km wide and 3.3 km long with a bed that shelves gently,
reaching 30 m at its deepest point. For the bathymetry and
limnology of Lake Pacucha see Hillyer et al. (2009).
The lake is surrounded by a small marsh, and a watershed
that gradually rises from an elevation of 3095 m to a fringing
bowl of peaks that are generally at 34003500 m. The
vegetation of this area is strongly inuenced by small-scale
topography, with valleys often supporting mesic forest on the
windward slope and contrastingly xeric conditions on the lee
slope. These aspects reect exposure to direct sun and
prevailing moisture-laden winds arriving from the Amazon.
The Pacucha Basin lies in the headwater region of the
Apurimac River and is kept moist by the ow of Amazonian
air up that valley. The catchment is largely deforested, but
based on its elevation, relatively moist location and remnant
vegetation it appears probable that this basin could support
upper Andean cloud forest. The clearance is not necessarily
modern as the lake is surrounded by archaeological sites of the
Chanka (c. ad 5001400), Huari (c. ad 6001000) and Inca
(c. ad 14001533) civilizations (B. Bauer, University of Illinois
at Champagne, pers. comm.).
In general, the tree line in this section of the Andes occurs at
an elevation of between c. 3300 and 3600 m. Even at the tree
line the forests of the Peruvian Andes are still diverse, with 120
woody taxa recorded at a single site in northern Peru
(Young, 1998). Typical trees of the uppermost forest come
B. G. Valencia et al.
1638 Journal of Biogeography 37, 16371647
2010 Blackwell Publishing Ltd
from genera that produce a lot of pollen, e.g. Podocarpus,
Hedyosmum, Alnus, Weinmannia and Vallea, while even at the
family level some important components are virtually silent in
the pollen record, e.g. Lauraceae and Rubiaceae. Above the tree
line, a thin band of shrubs is often dominated by Melastom-
ataceae, Ericaceae and Asteraceae (Young, 1998), while above
3700 m puna grasses and herbs generally dominate the
landscape. Puna extends upslope and becomes puna brava at
an elevation of c. 43004500 m.
MATERI ALS AND METHODS
Between 2003 and 2005, a series of cores were raised from the
deepest portion of Lake Pacucha using a modied Colinvaux
Vohnout piston corer (Colinvaux et al., 1999) deployed from a
oating platform. Bedrock was not reached, rather we hit the
functional limit of a lightweight drilling rig with 45 m of
drillstring deployed.
Three cores (PAC-B, 11 m long; PAC-D, 11 m long; and
PAC-E, 14.5 m long) were scanned for density at 0.5 cm
(gamma attenuation) using a GEOTEK multi-sensor core
logger (GEOTEK Ltd, Daventry, Northants, UK) at the
University of Florida. Split cores were described (see Appen-
dix S1 in Supporting Information) and scanned to acquire
high-resolution digital colour images. The three cores were
then aligned and cross-correlated to form a continuous 14.3-m
long core.
The core chronology was established using
14
C accelerator
mass spectrometry dates on bulk sediment avoiding carbonate
sections (Hillyer et al., 2009; Appendix S2). Ages were cali-
brated based on Stuiver & Reimer (1993) and Stuiver et al.
(2005) for dates < 11,000 years ago, and using Fairbanks et al.
(2005) for those that were older.
Loss-on-ignition samples were placed in weighed crucibles,
and reweighed following 12 h at 105 C, 4 h at 550 C and 2 h
at 950 C (Heiri et al., 2001). Subsamples for pollen were
taken at 5-cm intervals throughout much of the core, and
at 2-cm intervals in the upper 0.5 m.
Pollen preparation followed standard protocols, including
the use of Lycopodium spores to facilitate calculations of pollen
concentration (Stockmarr, 1972; Faegri & Iversen, 1989). Three
hundred terrestrial pollen grains were counted in each sample.
A total of 158 subsamples were analysed for fossil pollen and
116 for fossil charcoal. Samples were heated in KOH before
being ltered at 180 lm (after Clark & Royall, 1996). The
residue was searched for charcoal and all pieces were docu-
mented and measured using an Olympus stereoscope at 20
equipped with a digital camera, and the images were analysed
using ImageJ software (Rasband, 2005). The resulting dia-
grams were plotted using C2 (Juggins, 1991).
Ordination of data was completed using the version of
detrended correspondence analysis (DCA) (Hill, 1979) avail-
able in pcord-4 (McCune & Grace, 2002). The ordination was
run on the 47 taxa that attained > 0.5% in any sample or that
occurred in at least 10 samples (after Birks & Gordon, 1985).
RESULTS
Stratigraphy and chronology
The chronology and stratigraphy of the 14.3-m long record
from Lake Pacucha has previously been published (Hillyer
et al., 2009). We provide a core description and the
14
C ages
and calibrations in Appendices S1 and S2. A summary diagram
of the depthage relationship, stratigraphy and loss-on-igni-
tion is presented in Fig. 2.
Figure 1 A sketch map of Peru redrawn from Hillyer et al. (2009) showing the location of Lake Pacucha relative to lakes Consuelo,
Marcacocha and Titicaca. The star represents the coring site and the depth-isoclines are expressed in metres.
Cloud forest history: from ice age to modern
Journal of Biogeography 37, 16371647 1639
2010 Blackwell Publishing Ltd
Multivariate analysis of fossil pollen data
The DCA ordination of the fossil pollen data from Lake
Pacucha polarized the Holocene from full glacial-aged samples
on axis 1 of the DCA ordination (Fig. 3). Taxa characterizing
the extremes of axis 1 were: Chenopodiaceae, Ambrosia and
Ambrosia-type A at the negative (Holocene) extreme, and
Campanulaceae (other), Lysipomia and Apiaceae at the
positive (glacial-aged) extreme (Appendix S3). Axis 2 sepa-
rated early (> 6 cal. kyr bp) from late (< 6 cal. kyr bp)
Holocene samples. Chenopodiaceae, Ambrosia, Ambrosia-type
A, Alnus and Anacardiaceae characterized the negative (late
Holocene) extreme of axis 2, while Urticaceae/Moraceae,
Alchornea and Cecropia were clustered at the positive (early
Holocene) extreme of the axis.
Local pollen zones
PAC-1 (14.3011.21 m, c. 24.716 cal. kyr BP)
Pollen concentrations in PAC-1 ranged from 2600 to
10,000 grains cm
)3
(Fig. 4). Puna brava elements were well
represented and Poaceae ranged from 40 to 60%, with
Asteraceae from 5 to 15%. Polylepis (see Appendix S4 for
taxonomic justication) was the only abundant woodland
species in PAC-1 (29.5%). With the exception of two periods
of relative scarcity at 19.2 and 17.8 cal. kyr bp, Isoetes spores
were very abundant at 90140% of the pollen sum. Charcoal
abundances remained below 4 mm
2
cm
)3
with re activity
centred around 20.5 cal. kyr bp.
Figure 2 A depthage plot of calibrated
14
C ages for the sediments of the composite core from Lake Pacucha, Peru. The trend line used for
creating the chronology is shown. Major stratigraphic patterns and loss-on-ignition data for CaCO
3
and organic matter are also shown.
Further details are provided in Appendices S1 and S2 and in Hillyer et al. (2009).
Figure 3 Results of the detrended correspondence analysis
(DCA) ordination of fossil pollen data from Lake Pacucha, Peru:
axis 1 versus axis 2. Symbols identify the age range of local pollen
zones to which samples belong. The eigenvalues for axes 1 to 3
multiplied by 100 are 17.10, 6.73 and 2.31, respectively.
B. G. Valencia et al.
1640 Journal of Biogeography 37, 16371647
2010 Blackwell Publishing Ltd
PAC-2 (11.219.93 m, c. 1613.5 cal. kyr BP)
Pollen concentration in PAC-2 ranged from 17,000 to c. 34,000
grains cm
)3
and depicted the transition from grassland to
Andean forest. Representation of Poaceae and Asteraceae
declined by 50% with respect to PAC-1, while Polylepis
and Podocarpus doubled their concentrations. Acalypha
increased to about 24%. Shrubby taxa such as Ericaceae and
Figure 4 Percentage data for selected pollen taxa (occurring at > 0.5% and in at least 10 samples) from Lake Pacucha, Peru. Also
shown are pollen inux (grains cm
)2
year
)1
), pollen concentration (grains cm
)3
) and charcoal concentration (mm
2
cm
)3
).
Cloud forest history: from ice age to modern
Journal of Biogeography 37, 16371647 1641
2010 Blackwell Publishing Ltd
Melastomataceae were poorly represented. Isoetes declined in
abundance, uctuating between c. 9% and 60%. Charcoal
frequency and area increased with respect to PAC-1; counts
reached maxima of 9 mm
2
cm
)3
.
PAC-3 (9.939.25 m, c. 13.511 cal. kyr BP)
Pollen concentrations were similar to those in PAC-2. Poaceae
abundance rose to 3155%, while representation of Polylepis,
Podocarpus and Acalypha fell. In contrast, other woodland
elements, e.g. Hedyosmum, Alchornea and Urticaceae/Mora-
ceae, increased gradually. Shrubland and herbaceous types
remained almost unaltered. Cyperaceae and Myriophyllum
became abundant at the end of PAC-3 as Isoetes declined.
Charcoal area increased in the second half of PAC-3, reaching
a maximum of 10 mm
2
cm
)3
around 11.9 cal. kyr bp.
PAC-4 (9.258.05 m, c. 116 cal. kyr BP)
Pollen concentrations varied from 5000 to 70,000 grains cm
)3
.
Some forest taxa declined with respect to PAC-3 and Polylepis
became rare. Urticaceae/Moraceae increased in importance
throughout PAC-4, with peaks of 28 and 31% at c. 10.2 and
9.7 cal. kyr bp, respectively. Cecropia showed a sharp increase
to 8%, coincident with the second peak of Urticaceae/
Moraceae (Isoe tes was virtually absent). Charcoal showed a
bimodal distribution, with peaks centred at c. 9.7 and 7.5 cal.
kyr bp with values of 15 and 14 mm
2
cm
)3
, respectively.
PAC-5 (8.050 m, c. 6 cal. kyr BP to present)
Pollen concentrations uctuated from 500 to 44,000
grains cm
)3
. Poaceae ranged from 10 to 47% and declined
towards the end of PAC-5. Ambrosia uctuated mostly
between 5 and 20% except between c. 3.06 and 0.53 cal. kyr
bp, where it remained below 5%. Ambrosia-type A ranged
between 2 and 9% from c. 4.4 to 0.5 cal. kyr bp and oscillated
under 2% in the last 500 years. Chenopodiaceae became
important between c. 5.5 and 3.2 cal. kyr bp reaching 51% at
c. 4.1. Gomphrena showed a peak of 27% at c. 2.5 cal. kyr bp,
and remained under 1% for the last 650 years. Alnus and
Anacardiaceae gained importance after c. 1.26 cal. kyr bp;
however, only Alnus had peaks at both c. 0.5 and 0.1 cal. kyr
bp. Charcoal values declined in the lower half of PAC-5 where
the highest peaks (i.e. 16 and 18 mm
2
cm
)3
) were centred at
c. 4.8 cal. kyr bp. In the upper half of this zone, all charcoal
abundance values were under 3 mm
2
cm
)3
, excepting three
peaks at 2.9, 2.4 and 1.1 cal. kyr bp.
DI SCUSSI ON
The multivariate analysis displays the substantial changes
in vegetation that took place within this system between
the full glacial and the Holocene (Fig. 3). The full glacial
vegetation is a grassland and Polylepis woodland setting,
whereas the transition into the Holocene encompasses the
expansion of Andean forest populations at the site. A marked
difference is evident in the Holocene before and after 6 cal.
kyr bp.
Conditions at the LGM
The late glacial ora of the Pacucha Basin appears to have been
a grassland vegetation close to the lake, with Polylepis growing
within the landscape but not dominating it. The data were
consistent with the formation of a fringing marsh and puna/
puna brava grassland in moist areas, with Polylepis probably
growing on drier slopes. Fire was relatively infrequent,
consistent with a relatively low biomass that did not provide
sufcient fuel for regular re. Pollen inux was so low in this
landscape, compared with later, more productive, settings, that
the local vegetation was probably the equivalent of puna brava.
In the modern system, the puna brava/puna transition is a
soft transition, generally occurring at an elevation of c. 4300
4500 m. The fossil data are consistent with the existence of
puna brava habitats around Lake Pacucha from c. 24.7 to
16 cal. kyr bp, depicted by the DCA axis 1 scores (Appen-
dix S5). Because Isoetes was abundant, and this taxon cannot
survive persistent freezing temperatures (Dejoux & Iltis, 1992),
its occurrence suggested that conditions were similar to those
close to the puna/puna brava transition rather than further
upslope. As Pacucha lay at 3095 m elevation and at the LGM
supported a puna brava-type setting, species appeared to have
moved about 1300 m downslope during the LGM. A slight
warming was evident at c. 19.5 cal. kyr bp, but the cooler
conditions returned and the main ecosystem changes associ-
ated with deglaciation were manifested at 16 cal. kyr bp. The
warming at c. 19.5 cal. kyr bp was consistent with prior reports
of deglaciation starting between 22 and 19 cal. kyr bp (Seltzer
et al., 2002), while the stronger deglacial signature at c. 16 cal.
kyr bp was consistent with the diatom record from this site
(Hillyer et al., 2009) and the fossil pollen record of Lake
Titicaca (Paduano et al., 2003).
A quantication of the cooling at the LGM can be derived by
applying the modern adiabatic lapse rate of )5.2 C per
1000 m of ascent to the 1300 m descent of vegetation, resulting
in an inferred cooling of 78 C at the LGM. Our hypothesis
of LGM cooling is supported.
Migration
An expected response of plant species to climate change is
migration. Northward (southward) migration across northern
continents and upslope (downslope) migrations on mountains
are basic predictions of warmer (colder) conditions. While this
pattern is undoubtedly true at a gross scale, the role of
microrefugia complicates it (Rull, 2009). Species can survive in
microrefugia as anomalous populations within a landscape,
perhaps because of locally favourable microclimates or the
genetics of the population, or both of these factors. Evidence
for the survival of such microrefugial populations can be found
in modern outliers, such as the stands of Thuja occidentalis L.
B. G. Valencia et al.
1642 Journal of Biogeography 37, 16371647
2010 Blackwell Publishing Ltd
well to the south of their normal range at Cedar Bog, OH, USA
(Dachnowski, 1910). That population is sheltered from hot
summer temperatures by cool spring water, creating a pocket
of boreal conditions. Similarly, microrefugia have been
suggested to be important in past migrations (McGlone &
Clark, 2005; McLachlan & Clark, 2005). The population
consequences of having embedded pockets for future dispersal
are profound, both for the genetic structure of subsequent
populations and for the potential pattern and speed of
population response to new conditions.
The long-term (centennial- to millennial-scale) migrational
patterns of species are poorly known for tropical regions.
Rather, our insights are based on the changing occurrence of
species at one or a few sites. Positive evidence, i.e. the
appearance of a species within a fossil record, allows its
abundance to be estimated. However, what does negative
evidence the disappearance of a taxon mean? Traditionally,
if a lowland species disappears from a mid-elevation site
during a time of cooling we assume that its population has
moved downslope. In this model the downslope population
would be expected to expand, but this is just an assumption.
An alternative scenario accounting for a disappearance is that
its population has dwindled below our sensitivity of detection,
but it may or may not have experienced a population
expansion elsewhere.
In the Pacucha data, taxon abundances rise and fall in this
pollen record in a largely predictable fashion. Grassland gives
way to tree-line forest/shrubland, and ultimately to an Andean
forest. If we contrast this pattern with that of Lake Consuelo
(a lake at the lower limit of modern cloud forest in Peru),
which lies 1S of and 1700 m lower than Pacucha (Bush et al.,
2004; Urrego et al., 2005), there are some internal consis-
tencies. For example, Andean forest taxa, e.g. Podocarpus,
Hedyosmum, Alnus and Weinmannia, which today occupy a
swathe of elevation c. 15002000 m broad, are present in
Consuelo in the LGM, while largely absent from Pacucha.
These ndings are entirely consistent with a migration model
of species moving downslope in response to cooling (see
Appendix S5 and also Appendix S4). Podocarpus shows a fairly
typical response of upper Andean taxa that appear in the
Consuelo record at the LGM (Fig. 5). It is noticeable, however,
that at c. 15.5 cal. kyr bp the expansion of populations upslope
must be very sudden, as it appears that for a while this mid-
Andean taxon was abundant at both high and low sites. This
observation certainly does not disprove our standard view of
migration; there are many possibilities that could account for
leads or lags in migration that would produce this pattern.
However, the data are also consistent with a microrefugial-style
expansion on the Andean ank.
A more interesting case is that of the Urticaceae/Moraceae
(Fig. 5). In the upper Andean forest, Urticaceae/Moraceae
pollen (probably primarily Urticaceae) commonly accounts for
2040%of the pollen inux to lake sediments. Although some of
this input will be long-distance transport, it appears probable
that much is locally produced. During the glacial period,
Urticaceae/Moraceae pollen was almost absent from the Pacu-
cha record, and so we would assume that the trees moved
downslope. Urticaceae/Moraceae have their highest abundance
and diversity in lowland forests (Gentry, 1988), and so a general
downslope shufing of taxa would be expected, with the higher-
elevation species replacing mid-elevation species, and so on.
At Lake Consuelo, however, while Urticaceae/Moraceae
account for 40% of the pollen in the Holocene, they account
for almost none in the Pleistocene. Ergo, the high-Andean
species in these families did not simply move downslope and
replace lowland forms.
Clearly, on the basis of just two sites, we must be very
cautious about making pronouncements about past migra-
tion, but the 12 glacial-age samples that have been published
for Amazonia and the Andes all show major reductions in
Urticaceae/Moraceae abundance. These data suggest a variant
on the microrefugial model in which some populations
collapse across wide swathes of the environmental gradient
rather than producing a see-sawing of migrating populations.
When those taxa are as abundant as the Urticaceae/Moraceae
and as important in terms of basal area, the resulting
communities will be dissimilar from those of today, i.e. no-
analogue oras.
Figure 5 A comparison of Podocarpus and Urticaceae/Moraceae
representation at lakes Consuelo (1360 m elevation) and Pacucha
(3050 m elevation).
Cloud forest history: from ice age to modern
Journal of Biogeography 37, 16371647 1643
2010 Blackwell Publishing Ltd
Possibly, a related oddity of the Pacucha pollen record was a
very high abundance of Acalypha pollen at c. 1614 cal. kyr bp.
To nd the distinctive pollen of Acalypha in the midst of the
deglacial period with percentages of c. 24% of the pollen sum
was unusual. One possible explanation was that this was a
weedy invasive species which thrived on hillsides that were
being transformed by changing climates and re regimes. The
loss of Acalypha from the record at a time when Andean
elements were expanding could be explained by competitive
exclusion. Certainly, the Acalypha-rich assemblage was so
unusual that it appears a good example of a no-analogue ora,
but in this instance based on positive as opposed to negative
evidence.
Our data do not support a simple view of downslope/
upslope migration of species. More sites would be needed
before we could reject this hypothesis, but an alternative model
of response of tropical plants to climate change is clearly worth
investigating.
Vegetation and the mid-Holocene dry event
The mid-Holocene dry event has been identied primarily
from sedimentological composition or the timing of sedi-
mentary hiatuses in Amazonian and Andean lakes (Seltzer
et al., 1995; Baker et al., 2001b; Abbott et al., 2003; Mayle &
Power, 2008). This event has usually been described as a
multi-millennial-scale drought that was time transgressive,
occurring earlier in the northern central Andes than to the
south (Abbott et al., 2003). High-resolution diatom records
from southern Peru indicate that this was a period of
generally low lake level, interrupted by relative highstands
(Ekdahl et al., 2008; Hillyer et al., 2009). The sequence of
droughts caused Lake Pacucha to exist as a mostly shallow,
somewhat saline, pool between c. 11 cal. kyr bp and 6 cal.
kyr bp. During this interval, laminated sediments were
deposited in which black gyttja alternated with carbonate-
rich muds. These ne laminations were ecologically impor-
tant as they indicated that this was not a persistent uniform
drought but a period of alternating wet and dry events. The
drought, while sufcient to lower the lake level, had a limited
impact on the vegetation, and the transition from Andean
forest relatively rich in Podocarpus and Hedyosmum to one
dominated by Urticaceae/Moraceae appeared to reect a
thermal optimum at around 11 cal. kyr bp rather than strong
drought decits. In contrast to sites located in ecotonal areas
of the Amazon lowlands (Mayle & Power, 2008), the fossil
pollen record from Pacucha did not reect either the
drought documented in the lake lowstand or the later return
of moist conditions. This result was not inconsistent with
those from other Andean lakes where the pronounced
lowstand was much more noticeable limnologically than
vegetationally (e.g. Hansen & Rodbell, 1995; Hansen et al.,
2003; Paduano et al., 2003). On this basis we reject our
hypothesis that the lake lowstands caused a substantial
vegetation change to drought-tolerant taxa in the mid-
Holocene.
Human impacts
During the droughts of the mid-Holocene there had been
relatively little vegetational change; instead the most rapid
vegetational change in the entire sequence took place at c. 6
5.5 cal. kyr bp (Urrego et al., 2009). The sudden increase in
weedy species and Chenopodiaceae, increased re frequency
and loss of arboreal species, against a backdrop of a rising lake
level, was strongly indicative of human occupation. During the
last interglacial lowstands, pollen assemblages from Lake
Titicaca were dominated by Amaranthaceae (Hanselman et al.,
2005; Gosling et al., 2008). However, at Lake Pacucha the
diatom record indicated that when Amaranthaceae became
abundant the lake was fresh not brackish. Indeed, the lack of
Amaranthaceae during the driest times strongly indicated
that these were not saltmarsh taxa. Amaranthaceae were
common ruderal weeds, but in this section of the Andes they
also included Chenopodium quinoa (quinoa). Quinoa was an
important crop and dietary staple of Andean populations prior
to the widespread adoption of maize agriculture (Cusack,
1984). Quinoa appeared to decline in importance in this
record at c. 3 cal. kyr bp, closely paralleling a decline in its
inferred abundance at Lake Marcacocha (130 km from Pacu-
cha) at c. 2.8 cal. kyr bp. At Marcacocha, Zea mays (maize)
replaced quinoa, and this pattern appeared to be repeated at
Pacucha where the rst Z. mays pollen was found at c. 3.06 cal.
kyr bp and the last record of this crop was at 0.5 cal. kyr bp.
The co-occurrence of Amaranthaceae and Ambrosia pollen at
Lake Marcacocha, Peru, was interpreted by Chepstow-Lusty
et al. (1996, 2007) to indicate quinoa cultivation and possibly
the establishment of terracing. Ambrosia was suggested to have
been used to stabilize terrace walls (Chepstow-Lusty et al.,
2002). At Pacucha, two different kinds of low-spine Asteraceae
pollen, Ambrosia and Ambrosia-type A, were recognized. Each of
these pollen types had a distinctive late Holocene distribution. A
weed of disturbance, Ambrosia rose in abundance at 6.2 cal. kyr
bp, apparently preceding the increase of Amaranthaceae at
5.5 cal. kyr bp; both increases occurred in a climatically volatile,
but overall wetter, time than the preceding millennia (Hillyer
et al., 2009). However, both these taxa declined synchronously
at c. 3 cal. kyr bp. The Ambrosia-type A, increased in abundance
at c. 4.6 cal. kyr bp just as Amaranthaceae peaked in abundance,
but continued in the record until 0.5 cal. kyr bp when it declined
abruptly as Z. mays agriculture ceased a pattern that was
strikingly similar to that of Marcacocha.
Alnus increased markedly in abundance at c. 1.2 cal. kyr bp.
Despite not having been disturbed for > 3000 years, this
landscape had not supported a substantial Alnus population
prior to 1.2 cal. kyr bp. The abrupt increase in Alnus abundance
may have indicated that these trees were being planted and
tended as a timber resource by the Huari culture. The timing of
this suggested agroforestry was matched almost exactly by
increased Alnus occurrence at Marcacocha (Chepstow-Lusty &
Wineld, 2000).
A lake at this relatively low elevation is so rare in the Andes
and so attractive to humans that it has probably had a
B. G. Valencia et al.
1644 Journal of Biogeography 37, 16371647
2010 Blackwell Publishing Ltd
continuous history of occupation. However, the decline in
crop and Alnus abundance, coinciding with resurgent Urtic-
aceae/Moraceae and reduced re frequency at c. 0.5 cal. kyr bp,
is consistent with much less intense use of the landscape
following European arrival and its associated population
collapse (Hemming, 1970).
The data from Pacucha indicate landscape transformation
in the Pacucha watershed for at least the last 4000 years,
including the cultivation of C. quinoa, Z. mays and Alnus. Our
hypothesis of substantial landscape transformation prior to the
Inca is upheld.
CONCLUSI ONS
The LGM conditions of Pacucha were wet and cold, with
puna brava conditions extending down to c. 3000 m eleva-
tion. A cooling of c. 78 C was inferred from these data. The
deglacial warming may have begun as early as 19 cal. kyr bp,
but sustained warming between 16 and 14 cal. kyr bp
produced no-analogue oras that may have resulted from
expansions of populations from microrefugia rather than
from a more traditional model of sequential upslope
migration. In the Holocene the strong drying event that
reduced the 30-m deep modern lake to a shallow saline pool
had little effect on surrounding vegetation, possibly because
the drought was frequently interrupted by wet events
allowing the persistence of existing oras. As wet conditions
resumed, human occupation was signied by inferred culti-
vation beginning at c. 5.5 cal. kyr bp. This crop was largely
replaced by maize at c. 3 cal. kyr bp. Afforestation was
suggested by a sudden increase in Alnus representation at
c. 1.2 cal. kyr bp.
ACKNOWLEDGEMENTS
We are indebted to William Gosling for camaraderie in the
eld, and subsequent discussions. Similarly, Alex Chepstow-
Lusty has been generous with his insights. This work was
funded by NSF grants 0237573 (M.B.B.) 0237684 (M.R.S.) and
by a grant from the Gordon and Betty Moore Foundation. This
is publication 03 of the Florida Institute of Technologys
Institute for Research on Global Climate Change.
REFERENCES
Abbott, M.B., Wolfe, B.B., Wolfe, A.P., Seltzer, G.O., Aravena,
R., Mark, B.G., Polissar, P.J., Rodbell, D.T., Rowe, H.D. &
Vuille, M. (2003) Holocene paleohydrology and glacial
history of the central Andes using multiproxy lake sediment
studies. Palaeogeography, Palaeoclimatology, Palaeoecology,
194, 123138.
Baker, P.A., Rigsby, C.A., Seltzer, G.O., Fritz, S.C., Lowenstein,
T.K., Bacher, N.P. & Veliz, C. (2001a) Tropical climate
changes at millennial and orbital timescales on the Bolivian
Altiplano. Nature, 409, 698701.
Baker, P.A., Seltzer, G.O., Fritz, S.C., Dunbar, R.B., Grove,
M.J., Tapia, P.M., Cross, S.L., Rowe, H.D. & Broda, J.P.
(2001b) The history of South American tropical precipita-
tion for the past 25,000 years. Science, 291, 640643.
Birks, H.J.B. & Gordon, A.D. (1985) Numerical methods in
Quaternary pollen analysis. Academic Press, London.
Bush, M.B., Silman, M.R. & Urrego, D.H. (2004) 48,000 years
of climate and forest change from a biodiversity hotspot.
Science, 303, 827829.
Bush, M.B., Hansen, B.C.S., Rodbell, D., Seltzer, G.O., Young,
K.R., Leo n, B., Silman, M.R., Abbott, M.B. & Gosling, W.D.
(2005) A 17,000 year history of Andean climatic and vege-
tation change from Laguna de Chochos, Peru. Journal of
Quaternary Science, 20, 703714.
Chepstow-Lusty, A. & Wineld, M. (2000) Inca agroforestry:
lessons from the past. Ambio, 29, 322328.
Chepstow-Lusty, A., Frogley, M.R., Bauer, B.S., Bennett, K.D.,
Bush, M.B., Chutas, T.A., Goldsworthy, S., Tupayachi
Herrera, A., Leng, M., Rousseau, D.-D., Sabbe, K., Slean, G.
& Sterken, M. (2002) A tale of two lakes: droughts, El Ninos
and major cultural change during the last 5000 years in
the Cuzco region of Peru. Environmental catastrophes and
recovery in the Holocene. Abstracts volume (ed. by S. Leroy
and I.S. Stewart), pp. 2021. Brunel University, London.
Chepstow-Lusty, A., Frogley, M.R., Bauer, B.S., Leng, M.,
Cundy, A.B., Boessenkool, K.P. & Goida, A. (2007)
Evaluating socio-economic change in the Andes using
oribatid mite abundances as indicators of domestic ani-
mal densities. Journal of Archaeological Science, 34, 1178
1186.
Chepstow-Lusty, A.J., Bennett, K.D., Switsur, V.R. & Kendall,
A. (1996) 4000 years of human impact and vegetation
change in the central Peruvian Andes with events paral-
leling the Maya record? Antiquity, 70, 824833.
Chepstow-Lusty, A.J., Bush, M.B., Frogley, M.R., Baker, P.A.,
Fritz, S.C. & Aronson, J. (2005) Vegetation and climate
change on the Bolivian Altiplano between 108,000 and
18,000 yr ago. Quaternary Research, 63, 9098.
Cierjacks, A., Salgado, S., Wesche, K. & Hensen, I. (2008)
Post-re population dynamics of two tree species in high-
altitude Polylepis forests of central Ecuador. Biotropica, 40,
176182.
Clark, J.S. & Royall, P.D. (1996) Local and regional sediment
charcoal evidence for re regimes in presettlement north-
eastern America. Journal of Ecology, 84, 365382.
Cochrane, M.A. & Barber, C.P. (2009) Climate change, human
land use and future res in the Amazon. Global Change
Biology, 15, 601612.
Cochrane, M.A. & Laurance, W.F. (2002) Fire as a large-scale
edge effect in Amazonian forests. Journal of Tropical Ecology,
18, 311325.
Colinvaux, P.A., Bush, M.B., Steinitz-Kannan, M. & Miller,
M.C. (1997) Glacial and postglacial pollen records from the
Ecuadorian Andes and Amazon. Quaternary Research, 48,
6978.
Cloud forest history: from ice age to modern
Journal of Biogeography 37, 16371647 1645
2010 Blackwell Publishing Ltd
Colinvaux, P.A., De Oliveira, P.E. & Moreno, J.E. (1999)
Amazon pollen manual and atlas. Harwood Academic Press,
New York.
Cusack, D. (1984) Quinoa: grain of the Incas. Ecologist, 14, 21
31.
Dachnowski, A. (1910) A cedar bog in central Ohio. Ohio
Journal of Science, 11, 193199.
Dejoux, C. & Iltis, A. (1992) Lake Titicaca: a synthesis of lim-
nological knowledge. Kluwer Academic Publishers, Dordr-
echt, The Netherlands.
Ekdahl, E.J., Fritz, S.C., Baker, P.A., Rigsby, C.A. & Coley, K.
(2008) Holocene multidecadal- to millennial-scale hydro-
logic variability on the South American Altiplano. The
Holocene, 18, 867876.
Ellenberg, H. (1958) Wald oder Steppe? Die naturliche Pan-
zendecke der Anden Perus. Umschau, 1958, 645681.
Faegri, K. & Iversen, J. (1989) Textbook of pollen analysis, 4th
edn. Wiley, Chichester.
Fairbanks, R.G., Mortlock, R.A., Chiu, T.-C., Cao, L., Kaplan,
A., Guilderson, T.P., Fairbanks, T.W. & Bloom, A.L. (2005)
Marine radiocarbon calibration curve spanning 0 to
50,000 years B.P. based on paired 230Th/234U/238U and
14
C dates on pristine corals. Quaternary Science Reviews, 24,
17811796.
Fjeldsa, J. & Kessler, M. (1996) Conserving the biological
diversity of Polylepis woodlands of the highlands of Peru and
Bolivia: a contribution to sustainable natural resource man-
agement in the Andes. Nordic Foundation for Development
and Ecology, Copenhagen.
Fritz, S.C., Baker, P.A., Seltzer, G.O., Ballantyne, A., Tapia,
P.M., Cheng, H. & Edwards, R.L. (2007) Quaternary glaci-
ation and hydrologic variation in the South American tro-
pics as reconstructed from the Lake Titicaca drilling project.
Quaternary Research, 68, 410420.
Gentry, A.H. (1988) Changes in plant community diversity
and oristic composition on environmental and geograph-
ical gradients. Annals of the Missouri Botanical Garden, 75,
134.
Gosling, W.D., Bush, M.B., Hanselman, J.A. & Chepstow-
Lusty, A. (2008) Glacial-interglacial changes in moisture
balance and the impact on vegetation in the southern
hemisphere tropical Andes (Bolivia/Peru). Palaeogeography,
Palaeoclimatology, Palaeoecology, 259, 3550.
Graf, K. (1981) Palynological investigations of two post-glacial
peat bogs near the boundary of Bolivia and Peru. Journal of
Biogeography, 8, 353368.
Hanselman, J.A., Gosling, W.D., Paduano, G.M. & Bush, M.B.
(2005) Contrasting pollen histories of MIS 5e and the
Holocene from Lake Titicaca (Bolivia/Peru). Journal of
Quaternary Science, 20, 663670.
Hansen, B.C.S. & Rodbell, D.T. (1995) A late-glacial/Holocene
pollen record from the eastern Andes of northern Peru.
Quaternary Research, 44, 216227.
Hansen, B.C.S., Rodbell, D.T., Seltzer, G.O., Leo n, B., Young,
K.R. & Abbott, M. (2003) Late-glacial and Holocene vege-
tational history from two sites in the western cordillera of
southwestern Ecuador. Palaeogeography, Palaeoclimatology,
Palaeoecology, 194, 79108.
Heiri, O., Lotter, A.F. & Lemcke, G. (2001) Loss on ignition as
a method for estimating organic and carbonate content in
sediments: reproducibility and comparability of results.
Journal of Paleolimnology, 25, 101110.
Hemming, J. (1970) The conquest of the Incas. Harcourt, New
York.
Hill, M.O. (1979) DECORANA a Fortran program for detr-
ended correspondence analysis and reciprocal averaging.
Ecology and systematics. Cornell University, New York.
Hillyer, R., Valencia, B.G., Bush, M.B., Silman, M.R. & Stei-
nitz-Kannan, M. (2009) A 24,700-year paleolimnological
history from the Peruvian Andes. Quaternary Research, 71,
7182.
Hooghiemstra, H. & Van der Hammen, T. (2004) Quaternary
ice-age in the Colombian Andes: developing an under-
standing of our legacy. Philosophical Transactions of the
Royal Society B: Biological Sciences, 359, 173181.
Hooghiemstra, H., Melice, J.L., Berger, A. & Shackleton, N.J.
(1993) Frequency spectra and paleoclimatic variability of the
high-resolution 301450 ka Funza I pollen record (Eastern
Cordillera, Colombia). Quaternary Science Reviews, 12, 141
156.
Juggins, S. (1991) C2 1.6. data analysis. Newcastle University,
Newcastle, UK. Available at: http://www.staff.ncl.ac.uk/staff/
stephen.juggins/.
Kessler, M. (1995) Present and potential distribution of
Polylepis (Rosaceae) forests in Bolivia. Biodiversity and con-
servation of Neotropical montane forests (ed. by S.P. Chur-
chill, H. Balslev, E. Forero and J.L. Luteyn), pp. 281294.
New York Botanical Garden, New York.
Malcolm, J.R., Liu, C., Neilson, R.P., Hansen, L. & Hannah, L.
(2006) Global warming and extinctions of endemic species
from biodiversity hotspots. Conservation Biology, 20, 538
548.
Mayle, F.E. & Power, M.J. (2008) Impact of a drier EarlyMid-
Holocene climate upon Amazonian forests. Philosophical
Transactions of the Royal Society B: Biological Sciences, 363,
18291838.
Mayle, F.E., Burbridge, R. & Killeen, T.J. (2000) Millennial-
scale dynamics of southern Amazonian rain forests. Science,
290, 22912294.
McCune, B. & Grace, J.B. (2002) Analysis of ecological
communities. MJM Software Design, Gleneden Beach,
OR.
McGlone, M.S. & Clark, J.S. (2005) Microrefugia and mac-
roecology. Climate change and biodiversity (ed. by T. Lovejoy
and L. Hannah), pp. 157159. Yale University Press, New
Haven, CT.
McLachlan, J.S. & Clark, J.S. (2005) Molecular indicators of
tree migration capacity under rapid climate change. Ecology,
86, 20882098.
Mourguiart, P. & Ledru, M.P. (2003) Last Glacial Maximum in
an Andean cloud forest environment (Eastern Cordillera,
Bolivia). Geology, 31, 195198.
B. G. Valencia et al.
1646 Journal of Biogeography 37, 16371647
2010 Blackwell Publishing Ltd
Myers, N., Mittermeier, R.A., Mittermeier, C.G., da Fonseca,
G.A.B. & Kent, J. (2000) Biodiversity hotspots for conser-
vation priorities. Nature, 403, 853858.
Nu nez, L., Grosjean, M. & Cartajena, I. (2002) Human occu-
pations and climate change in the Puna de Atacama, Chile.
Science, 298, 821824.
Paduano, G.M., Bush, M.B., Baker, P.A., Fritz, S.C. & Seltzer,
G.O. (2003) A vegetation and re history of Lake Titicaca
since the Last Glacial Maximum. Palaeogeography, Palaeo-
climatology, Palaeoecology, 194, 259279.
Rasband, W. (2005) ImageJ 1.38x. National Institutes of
Health. Available at: http://rsbweb.nih.gov/ij/.
Rull, V. (2009) Microrefugia. Journal of Biogeography, 36, 481
484.
Seltzer, G.O., Rodbell, D.T. & Abbott, M.B. (1995) Andean
glacial lakes and climate variability since the Last Glacial
Maximum. Bulletin de lInstitut Francais dE

tudes Andines,
24, 539550.
Seltzer, G.O., Rodbell, D.T., Baker, P.A., Fritz, S.C., Tapia,
P.M., Rowe, H.D. & Dunbar, R.B. (2002) Early deglaciation
in the tropical Andes. Science, 298, 16851686.
Stockmarr, J. (1972) Tablets with spores used in absolute
pollen analysis. Pollen et Spore, XIII, 615621.
Stuiver, M. & Reimer, P.J. (1993) Extended
14
C database and
revised Calib radiocarbon calibration program. Radiocarbon,
35, 215230.
Stuiver, M., Reimer, P.J. & Reimer, R.W. (2005) Calib 5.0.
Available at: http://intcal.qub.ac.uk/calib/.
Urrego, D.H., Silman, M.R. & Bush, M.B. (2005) The Last
Glacial Maximum: stability and change in an Andean cloud
forest. Journal of Quaternary Science, 20, 693701.
Urrego, D.H., Bush, M.B., Silman, M.R., Correa-Metrio, A.Y.,
Ledru, M.P., Mayle, E., Paduano, G. & Valencia, B.G. (2009)
Millennial-scale ecological changes in tropical South America
since the Last Glacial Maximum. Past climate variability from
the Last Glacial Maximum to the Holocene in South America
and surrounding regions. Past climate variability in South
America and surrounding regions (ed. by F. Vimeux, F.
Sylvestre and M. Khodri), pp. 283300. Springer, Dordrecht.
Van der Hammen, T. (1974) The Pleistocene changes of veg-
etation and climate in tropical South America. Journal of
Biogeography, 1, 326.
Van der Hammen, T. & Gonza lez, E. (1959) Historia de clima
y vegetacion del Pleistocene Superior y del Holoceno de la
Sabana de Bogota y alrededeores. Ofcial Report of the
Servicio Geologico Nacional, Department of Cundinamarca,
Bogota .
Weng, C., Bush, M.B., Curtis, J.H., Kolata, A.L., Dillehay, T.D.
& Binford, M.W. (2006) Deglaciation and Holocene climate
change in the western Peruvian Andes. Quaternary Research,
66, 8796.
Young, K.R. (1998) Composition and structure of a timberline
forest in north-central Peru. Forest biodiversity in North,
Central and South America and the Caribbean (ed. by
F. Dallmeiere and J.A. Comiskey), pp. 595613. UNESCO
and the Parthenon Publishing Group, Carnforth, UK.
SUPPORTI NG I NFORMATI ON
Additional Supporting Information may be found in the
online version of this article:
Appendix S1 Stratigraphic summary for the sediments of the
composite core PAC-B, PAC-D and PAC-E, Lake Pacucha,
Peru, from Hillyer et al. (2009).
Appendix S2 Calibrated ages derived from radiocarbon dates
for samples from Lake Pacucha, Peru, from Hillyer et al.
(2009).
Appendix S3 Biplot results of the ordination of fossil pollen
data from Lake Pacucha, Peru: DCA axis 1 versus axis 2.
Appendix S4 Ecological and taxonomic notes regarding
identication of pollen types.
Appendix S5 DCA axis 1 scores versus time results of the
ordination of fossil pollen data from Lake Pacucha, Peru.
As a service to our authors and readers, this journal provides
supporting information supplied by the authors. Such mate-
rials are peer-reviewed and may be re-organized for online
delivery, but are not copy-edited or typeset. Technical support
issues arising from supporting information (other than
missing les) should be addressed to the authors.
BI OSKETCH
Bryan G. Valencia grew up in Cusco, Peru, where he studied
ecology and botany. Currently a graduate student, he is
investigating the palaeoecology of the southern Peruvian
Andes.
Author contributions: B.V. conducted all phases of eldwork,
laboratory analyses and rst draft of the manuscript. D.H.,
M.S. and M.B. participated in the eldwork, interpretation of
data and manuscript preparation.
Editor: John Birks
Cloud forest history: from ice age to modern
Journal of Biogeography 37, 16371647 1647
2010 Blackwell Publishing Ltd

You might also like