plants

Article
The Evolution of Agrarian Landscapes in the Tropical Andes
Courtney R. Shadik 1 , Mark B. Bush 1, *, Bryan G. Valencia 2 , Angela Rozas-Davila 1 , Daniel Plekhov 3 ,
Robert D. Breininger 1 , Claire Davin 4 , Lindsay Benko 1 , Larry C. Peterson 5 and Parker VanValkenburgh 6

1 Institute of Global Ecology, Florida Institute of Technology, Melbourne, FL 32901, USA;

[email protected] (A.R.-D.); [email protected] (L.B.)
2 Facultad de Ciencias de La Tierra y Agua, Universidad Regional Amazónica Ikiam, Tena 150150, Ecuador;
[email protected]
3 Department of Anthropology, Portland State University, Portland, OR 97201, USA; [email protected]
4 Department of Mathematics, University of Richmond, Richmond, VA 23173, USA;
[email protected]
5 Rosenstiel School of Marine, Atmospheric, and Earth Science, University of Miami, Miami, FL 33149, USA;
[email protected]
6 Department of Anthropology, Brown University, Providence, RI 02915, USA;
[email protected]
* Correspondence: [email protected]

Abstract: Changes in land-use practices have been a central element of human adaptation to Holocene
climate change. Many practices that result in the short-term stabilization of socio-natural systems,
however, have longer-term, unanticipated consequences that present cascading challenges for human
subsistence strategies and opportunities for subsequent adaptations. Investigating complex sequences
of interaction between climate change and human land-use in the past—rather than short-term causes
and effects—is therefore essential for understanding processes of adaptation and change, but this
approach has been stymied by a lack of suitably-scaled paleoecological data. Through a high-
resolution paleoecological analysis, we provide a 7000-year history of changing climate and land
management around Lake Acopia in the Andes of southern Peru. We identify evidence of the onset
of pastoralism, maize cultivation, and possibly cultivation of quinoa and potatoes to form a complex
agrarian landscape by c. 4300 years ago. Cumulative interactive climate-cultivation effects resulting in
Citation: Shadik, C.R.; Bush, M.B.;
erosion ended abruptly c. 2300 years ago. After this time, reduced sedimentation rates are attributed
Valencia, B.G.; Rozas-Davila, A.;
to the construction and use of agricultural terraces within the catchment of the lake. These results
Plekhov, D.; Breininger, R.D.; Davin,
provide new insights into the role of humans in the manufacture of Andean landscapes and the
C.; Benko, L.; Peterson, L.C.;
VanValkenburgh, P. The Evolution of
incremental, adaptive processes through which land-use practices take shape.
Agrarian Landscapes in the Tropical
Andes. Plants 2024, 13, 1019. https:// Keywords: climate change; fossil pollen; lake sediment; terracing; crops; pastoralism; Sporormiella;
doi.org/10.3390/plants13071019 charcoal; XRF

Academic Editor: Dénes Lóczy

Received: 15 February 2024

Revised: 4 March 2024 1. Introduction
Accepted: 12 March 2024 The tropical Andes are both a biodiversity and a conservation hotspot, as defined
Published: 3 April 2024 on the basis of high plant endemicity and substantial habitat loss [1,2]. Tropical Andean
landscapes have a deep history of human modification as revealed through studies of fossil
pollen and charcoal [3,4], spores [5,6], and phytoliths [7–9], coupled with archaeology [10–12].
Such studies have provided evidence of early human impacts such as the extirpation of
Copyright: © 2024 by the authors.
megafauna and the introduction of frequent fires [13]. Consequently, the modern landscape,
Licensee MDPI, Basel, Switzerland.
especially above tree-line, is widely regarded as a landscape manufactured by past human
This article is an open access article
distributed under the terms and
actions [14–16].
conditions of the Creative Commons
1.1. An Historical Overview of Early Settlement and Cultivation
Attribution (CC BY) license (https://
creativecommons.org/licenses/by/ The earliest archaeological evidence of people occupying the high Andes of Peru
4.0/). comes from open-air workshops where stone tools were fashioned [17]. These sites are

Plants 2024, 13, 1019. https://doi.org/10.3390/plants13071019 https://www.mdpi.com/journal/plants

Plants 2024, 13, 1019 2 of 19

dated to about 12,500–11,500 calibrated radiocarbon years before present (hereafter cal BP).
The initial settlers of the Andes were hunter-gatherers, but archaeological data point to
a transition toward more sedentary lifestyles about 8000 cal BP [18]. Coastal indigenous
societies had cultivated domesticated squash and beans by 10,000 cal BP [19], but Andean
cultivation of crops, such as maize, potato and quinoa, appears to have expanded about
c. 7000 cal BP [20,21]. Domestication of livestock may have gone hand-in-hand with crop
cultivation. In southern Peru, llama and alpaca domestication may have begun as early
as 7000 cal BP [22], possibly triggering a synergy between the corralling of camelids and
the cultivation of weedy amaranths (pseudo-cereals such as quinoa). These plants may
have been favored by disturbed ground and the high nutrient availability in soils enhanced
with dung in and around corrals [23]. By about 4000 cal BP, reliance on domesticated
camelids and crops such as Zea mays (maize) and Chenopodium quinoa (quinoa) had become
widespread [21,24]. While we have a basic timeline of the introduction of these practices,
there is no single record that shows their development and the ecological consequences of
that land use.

1.2. Human Responses to Holocene Climate Change and the Temporal Scale of Study
Within this timeframe of plant and animal domestication, frequent and intense
droughts associated with the Mid-Holocene Dry Event (MHDE) of c. 9000–4000 cal BP
induced the lowest lake levels of the last 100,000 years [25] and probably impacted many
aspects of human life [26,27]. Between c. 4000 years ago and the European invasion of the
1500s, many societies made investments in extensive irrigation and drainage canals, raised
fields and terraces to regulate water supplies and safeguard soils [28–30]. The respective
roles of climate change and social factors in driving humans to invest in agriculture and
modify landscapes has been actively debated [31–33]. Much of this debate has emphasized
short-term interactions between humans and their environments and provided mostly
monocausal explanations for their resulting changes. Missing from such explanations has
been consideration of the influence of antecedent land-use practices and how their legacies
create opportunities and obstacles for later land use [34]. To build robust models of these
processes, we face the considerable challenge of capturing interactions between social
and environmental factors over long timespans (millennia), while also attending to the
processes through which human agents make decisions. Creating such models has been
complicated for both paleoecologists and archaeologists, as paleoecological studies are
often too temporally coarse to capture short-term decision-making, while archaeological
research tends to prioritize the study of single periods, missing longer-term trajectories. In
addition, land-use practices often fail to leave behind materials that can be easily dated.
Where remains (such as terrace walls, corrals, and canal courses) are visible, archaeologists
may be able to infer that certain land-use practices occurred but then have trouble dating
them due to repeated practices of maintenance and modification, including the re-use of
soils and building materials. Thus, in moving from short-term monocausal explanations to
more complex models of long-term change, untangling specific relationships of cause and
effect becomes more difficult and resulting models risk losing explanatory power [35,36].
Here, we present a 7000-year history that reveals the timing and trajectory of climate
and land management around Lake Acopia in the Andes of southern Peru. Specifically, we
ask: How did climate change influence land use around Lake Acopia? And can we detect
adaptations to environmental change in the way the land was used?

2. Site Description, Materials and Methods

Lake Acopia (Figure 1) lies at 3715 m above sea level (hereafter masl), approximately
70 km southeast of Cusco, Peru. This lake lies within the region of the Wari (c. AD
600–1000; 1350–950 cal BP) and Inka (c. AD 1400–1533; 550–417 cal BP) empires and
is likely to have been occupied for a much longer period. The pre-Wari history of this
region is poorly known, and yet it is considered to lie within the cradle of plant and
animal domestication [21]. The climate of this part of the Andes is strongly seasonal with
 2024). The basin is bordered by the town of Acopia to the north, potato fields and pastures
to the east and south, and a steeper slope to the west that supports Eucalyptus—a
nonnative tree that has only been commonly cultivated in this landscape during the past
60 years [37]. Although Poaceae are the most abundant cover, and Eucalyptus is the com-
Plants 2024, 13, 1019 monest large woody species in the basin today, taxa such as Vallea, Schinus, and a 3wide of 19
diversity of Asteraceae and Fabaceae would have been important components of the nat-
ural vegetation. In 2008, we mapped the bathymetry of Lake Acopia, collected using a
hand-held
a strong dry sonar.
seasonThelasting
lake has
froma closed basin with a The
May–September. mostly
meanflatannual
bottom,precipitation
averaging aroundin the
20
Cusco area is 1278 mm, with a mean annual temperature of approximatelythat
m deep except for a depression in the west-central portion of the basin 7.8 ◦reaches
C (https: a
98 m depth (Figure 2). As the deepest point in Lake Acopia was
//en.climate-data.org/south-america/peru/cusco/cusco-1016/#climate-graph; accessed beyond the limits of our
coring rig, and
26 February suchThe
2024). a sheer
basindepression
is borderedwouldby theinvite
town slumping,
of Acopia to wetheselected
north, apotato
flat shelf at
fields
aand
depth of 19.5 m for coring (Figure 2a).
pastures to the east and south, and a steeper slope to the west that supports Eucalyptus—
A sediment
a nonnative core has
tree that wasonly
raisedbeenfrom the coring
commonly point (Figure
cultivated in this2a) using a Colinvaux-
landscape during the
Vohnout piston corer from a raft of rubber boats [38]. Eight drives,
past 60 years [37]. Although Poaceae are the most abundant cover, and Eucalyptus totaling 7.74 m isof the
re-
covered sediment, were acquired before hitting a methane pocket in
commonest large woody species in the basin today, taxa such as Vallea, Schinus, and a the sediment, halting
the coring
wide process.
diversity All cores and
of Asteraceae wereFabaceae
sealed inwould
the field
haveand sent
been back to Florida
important Institute
components of theof
Technology where they were stored at 4 °C. The split cores were described
natural vegetation. In 2008, we mapped the bathymetry of Lake Acopia, collected using a and one of the
two halvessonar.
hand-held subsampled.
The lake Eight bulk sediment
has a closed basin withsamples
a mostlywere shippedaveraging
flat bottom, to Woods Hole
around
Oceanographic
20 m deep except Institute NOSAMS (National
for a depression Ocean Sciences
in the west-central portion Accelerator
of the basin Mass
thatSpectrom-
reaches a
etry)
98 m for AMS
depth (accelerator
(Figure 2). As mass spectrometry)
the deepest point indating and onewas
Lake Acopia to Direct
beyond AMS.
the The
limitspackage
of our
rbacon
coring rig, and such a sheer depression would invite slumping, we selected a flat shelfcali-
version 2.2 [39] was used to create the Acopia age model using the IntCal20 at a
bration
depth ofcurve
19.5 m[40,41].
for coring (Figure 2a).

−5-

− 10 -
Latitude
O

− 15 -
Elevation
6500 m

0
− 75 − 70
O
Longitude

Figure 1. Relative location of Lake Acopia (red star) in relation to the Andes and other sites mentioned
Figure 1. Relative location of Lake Acopia (red star) in relation to the Andes and other sites men-
in text.
tioned in text.
A sediment core was raised from the coring point (Figure 2a) using a Colinvaux-
Vohnout piston corer from a raft of rubber boats [38]. Eight drives, totaling 7.74 m of
recovered sediment, were acquired before hitting a methane pocket in the sediment, halting
the coring process. All cores were sealed in the field and sent back to Florida Institute of
Technology where they were stored at 4 ◦ C. The split cores were described and one of the
two halves subsampled. Eight bulk sediment samples were shipped to Woods Hole Oceano-
graphic Institute NOSAMS (National Ocean Sciences Accelerator Mass Spectrometry) for
AMS (accelerator mass spectrometry) dating and one to Direct AMS. The package rbacon
version 2.2 [39] was used to create the Acopia age model using the IntCal20 calibration
curve [40,41].
Loss-on-ignition at 108 subsampled locations along the core followed the protocol of
Heiri et al. [42]. Subsamples were heated in a muffle furnace to 550 ◦ C for 4 h to determine
organic content, and at 950 ◦ C for 2 h to quantify carbonate content.
Plants
Plants 2024,
2024, 13,
13, x1019
FOR PEER REVIEW 4 of 21
4 of 19

Figure 2. (a) The bathymetry of Lake Acopia relative to modern agricultural settings and terracing.
Figure 2. (a) The bathymetry of Lake Acopia relative to modern agricultural settings and terracing.
The star marks the coring location. (b) Google Earth image of Lake Acopia. Trees shown to the left
The star marks the coring location. (b) Google Earth image of Lake Acopia. Trees shown to the left of
of the lake are the exotic Eucalyptus.
the lake are the exotic Eucalyptus.
Loss-on-ignition at 1083 subsampled locations along the core followed the protocol of
A total of 108, 0.5-cm sediment subsamples were analyzed for pollen, Sporormiella,
Heiri et al. [42]. Subsamples were heated in a muffle furnace to 550 °C for 4 h to determine
and maize presence at the same locations as LOI. Pollen samples were prepared using
organic content, and at 950 °C for 2 h to quantify carbonate content.
standard palynological procedures [43]. Pollen grains and spores were identified with a
A total of 108, 0.5-cm3 sediment subsamples were analyzed for pollen, Sporormiella,
Zeiss Axioskop photomicroscope at ×630 magnification using pollen atlases [44,45], the
and maize
modern presence
pollen at the
collection at same
Florida locations
Tech andasthe LOI. Pollen samples
Neotropical pollen were
key andprepared
database using
[46].
standard palynological procedures [43]. Pollen grains and spores
After initial counts were made, samples were filtered through a 55-µm mesh to facilitate were identified with a
Zeiss Axioskop
extended countingphotomicroscope
for large grains at [47]
×630such
magnification
as Zea maysusing pollen
(maize). Zeaatlases [44,45], the
was identified as
modern
Poaceae grains >80 µm. Sporormiella spores were quantified as % of the pollen database
pollen collection at Florida Tech and the Neotropical pollen key and sum and
[46]. After initial counts
as concentrations (sporeswere permade,
cm3 ).samples
CONISS were filtered athrough
software, a 55-µm
multivariate mesh totool
analysis facil-
to
itate extended counting for large grains [47] such as Zea mays (maize).
determine significant changes in a time-series, was run on all pollen data [48] to define the Zea was identified
as Poaceae
local pollengrains
zones.>80 µm. Sporormiella spores were quantified as % of the pollen sum and
as concentrations (spores per
Sediment subsamples cm3cm
of 0.5 ). CONISS software,
3 for charcoal a multivariate
analysis were takenanalysis tool to
continuously de-
every
termine
centimetersignificant
along the changes
core. Thein a time-series,
subsamples was wererun on allwith
filtered pollen datathrough
water [48] to define
a 180-µmthe
local pollen zones.
mesh to separate macrocharcoal from other material within the sediment. The particles on
Sediment
the mesh were subsamples
transferred to of a0.5 cm3dish
petri for charcoal analysis and
for identification werequantification.
taken continuously every
All charcoal
centimeter along the core. The subsamples were filtered with
was identified at ×20 magnification on an Olympus stereoscope, and surface area was water through a 180-µm
mesh to separate
quantified using macrocharcoal
ImageJ version1.54a from other material
software [49]. within the sediment. The particles on
the mesh were transferred
The archive to a petri
half of the core dish forwith
was scanned identification
an Avaatech and quantification.
(Dodewaard, The All charcoal
Netherlands)
was
XRF identified
core scanner at [50]
×20 atmagnification
the Universityon ofan Olympus
Miami, stereoscope,
Rosenstiel School ofand surface
Marine, area was
Atmospheric,
quantified using ImageJ
and Earth Science. version1.54acore
Each refrigerated software [49]. to room temperature before analysis
was brought
The archive
to reduce half of the
condensation on thecoresediment
was scanned with
surface. Theancore
Avaatech
surface(Dodewaard, The Nether-
was then gently scraped
lands)
with aXRF glasscore scanner
slide, [50] atathe
to present freshUniversity
surface for of Miami,
scanning. Rosenstiel School
Each core wasofcovered
Marine,withAt-
mospheric,
0.4-µm-thick and Ultralene ® to prevent
Earth Science. Each contact
refrigerated
between core thewasXRF
brought to room
detector temperature
and the sediment
surface.
before The XRF
analysis detector
to reduce read elemental
condensation composition
on the at 10 kVThe
sediment surface. andcore
30 kV at a current
surface was then of
1000 µA
gently for 20 swith
scraped every 2 mmslide,
a glass alongtothe core. a fresh surface for scanning. Each core was
present
coveredMajor
withchanges
0.4-µm-thickin mean erosion,
Ultralene ® to using
prevent the Ti record,
contact betweenwerethedetermined
XRF detector using the
and the
“changepoint” package in R version 2.2.2 with the BinSeg
sediment surface. The XRF detector read elemental composition at 10 kV and 30 kV at a method [51]. The number
of meaningful
current of 1000 µA change
for 20points
s every was2 mmdetermined
along thefrom core. the diagnostic plot created in the
“changepoint”
Major changes package whereerosion,
in mean the addition
using ofthea Titrue change
record, point
were improves using
determined modelthe fit.
Changes in erosion
“changepoint” package patterns, as indicated
in R version by the
2.2.2 with thetrue changepoints,
BinSeg method [51]. wereTheanalyzed
number of in
meaningful change points was determined from the diagnostic plot created in the
 “changepoint” package where the addition of a true change point improves mod
Changes in erosion patterns, as indicated by the true changepoints, were analyzed in
bination with maize presence, charcoal, and Sporormiella influx as proxies of human
Plants 2024, 13, 1019 scape disturbance. 5 of 19
For assessing the density and composition of terraces within the Acopia basin
employed a combination of extensive field survey and satellite imagery analysis.
surveys
combination with took
maize place during
presence, the summer
charcoal, of 2022 and
and Sporormiella involved
influx targeted
as proxies visits to several
of human
of the Acopia
landscape disturbance. basin to characterize and document terrace morphologies and ongoing
These visits were paired with intensive digitization of all visible
For assessing the density and composition of terraces within the Acopia basin, weterraces in publicly
able satellite imagery.
employed a combination of extensive field survey and satellite imagery analysis. Field
surveys took place during the summer of 2022 and involved targeted visits to several
3. Results
parts of the Acopia basin to characterize and document terrace morphologies and ongoing
The modern
uses. These visits were landscape
paired with has
intensive at least 233
digitization of km of extant
all visible terracing
terraces within its app
in publicly
mately
available satellite 33 km2 catchment (Figure 3), the vast majority of which appears to be anci
imagery.
based on the construction styles used and the frequency of artifacts found on the su
3. Results around the terraces.
The modern landscape hasthe
Four out of at nine
least bulk
233 km of extant
sediment 14Cterracing
dates werewithin its approximately
accepted and used to create th
33 km2 catchment
model (Table 1; Figure 4). The five dates that were rejected as outliers cameon
(Figure 3), the vast majority of which appears to be ancient—based from the u
the construction
120styles
cm ofused and the
sediment. Thefrequency
resultingof artifacts
age model found on the
provides surface
a basal age around
of c. 7140 (calib
the terraces. radiocarbon years before present; hereafter cal BP).

a) b)

Figure 3.
Figure 3. (a) Terracing (a) Terracing
mapped withinmapped withinof
the watershed the watershed
Lake Acopia of Lake Acopia
(derived (derived
from Google andfrom
BingGoogle and
imagery), black outline shows approximate location of photograph in Panel (b). (b) modern terracing modern t
imagery), black outline shows approximate location of photograph in Panel (b). (b)
ing beside Lake Acopia, Peru.
beside Lake Acopia, Peru.

Four out Table

of the1.nine
Results
bulkofsediment
AMS dating
14 Cofdates
eightwere
bulk sediment
acceptedsamples
and usedfrom Lake Acopia,
to create the agePeru. Ages
calibrated using rbacon [52] and the IntCal20 calibration curve [40].
model (Table 1; Figure 4). The five dates that were rejected as outliers came from the upper
Lab I 120 cm of sediment.
Depth (cm) The 14 resulting
C (yr BP)age
± 1model
σ provides
cal BP 2σa Range
basal age ofMedian
c. 7140 (calibrated
Age Cal BP 13

radiocarbon
OS-135406 * years
7 before present;
5070 hereafter
± 40 cal BP).
5660 – 5897 5817 −27
OS-125490 * 71 4500 ± 20 5159 – 5287 5167 −2
Table 1. Results of AMS dating of eight bulk sediment samples from Lake Acopia, Peru. Ages were
DAMS-052062 * 79 3814 ± 34 4089 – 4396 4203
calibrated using rbacon [52] and the IntCal20 calibration curve [40].
OS-126479 * 81 2610 ± 20 2698 – 2757 2749 −24
Lab I OS-135405
Depth (cm)* 120
14 C (yr BP) ± 1 σ 3140 ±cal BP
252σ Range
3222 – 3383 3368 13 C
Median Age Cal BP −27
OS-135406 *
OS-125489
7 5070
156
± 40
2480 5660± 20 – 2633 5897– 2701 5817 2589−27.51 −2
OS-125490 * OS-125503
71 4500 365.5
± 20 3070 5159 ± 20 – 3283 –
5287 3345 5167 3291 −25.5 −25
DAMS-052062 * OS-125504
79 3814 578
± 34 4180 4089± 20 – 4570 4396– 4823 4203 4728 −25
OS-126479 * 81
OS-125505 2610 ±
698 20 5610 2698± 25 – 6298 2757– 6407 2749 6372−24.35 −26
OS-135405 * 120 3140 ± 25 3222 – 3383 3368 −27.23
* Indicates rejected radiocarbon date.
OS-125489 156 2480 ± 20 2633 – 2701 2589 −25.4
OS-125503 365.5 3070 ± 20 3283 – 3345 3291 −25.84
OS-125504 578 4180 ± 20 4570 – 4823 4728 −25.91
OS-125505 698 5610 ± 25 6298 – 6407 6372 −26.69
* Indicates rejected radiocarbon date.
Plants 2024, 13, x FOR PEER REVIEW 6 of 21
Plants 2024, 13, 1019 6 of 19

Figure
Figure 4. Bayesianchronology
4. Bayesian chronologyofof thethe sediment
sediment core
core raised
raised fromfrom
LakeLake Acopia.
Acopia. Themodel
The age age model
is de-
is developed
veloped using using
rbaconrbacon
[52] and[52]
the and the IntCal20
IntCal20 calibration calibration
curve [40].curve [40].
Panel (a) showsPanel
the (a) shows
Markov the
Chain
Markov Chainiterations
Monte Carlo Monte Carlo
withiterations with little
little structure amongstructure amongiterations
neighboring neighboring iterations(b)
(desirable); (desirable);
provides
(b)
the provides
distributionthe of
distribution of accumulation
accumulation rates within the rates within
core and the corememory
(c) the and (c) strength,
the memory whichstrength,
deter-
mines determines
which how much how influence
muchpreceding
influence samples
preceding have on the
samples interpolated
have curve. In bcurve.
on the interpolated and c,Inthe green
b and c,
curves
the greenarecurves
the priors, and
are the greyand
priors, histograms are the posterior
grey histograms distributions.
are the posterior (d) The(d)
distributions. age-depth model
The age-depth
for Lake
model forAcopia showsshows
Lake Acopia calibrated 14C dates as transparent blue shapes whose heights display the 1
calibrated 14 C dates as transparent blue shapes whose heights display
σ range of the calibrated ages, and widths display the most probable range of calibrated ages used
the 1 σ range of the calibrated ages, and widths display the most probable range of calibrated ages
in the age model iterations. Darker greys indicate more likely calendar ages. Grey dotted lines depict
used in the age model iterations. Darker greys indicate more likely calendar ages. Grey dotted lines
95% confidence intervals. The red dotted line depicts the best model based on the weighted mean
depict
age for95%
eachconfidence
depth. Theintervals. The red
purple circle dotted
depicts an line
age depicts
based on theanbest model
exotic basedXon
species. the weighted
s mark the cali-
mean age for each depth. The purple
brated ages of dates rejected as outliers. circle depicts an age based on an exotic species. ‘X’s mark the
calibrated ages of dates rejected as outliers.
The last reliable 14C age is at c. 2590 cal BP, and ages in the upper 120 cm are derived
The last reliable 14 C age is at c. 2590 cal BP, and ages in the upper 120 cm are derived
by a simple interpolation between that point and the surface, taken to be 2008 when the
by a simple interpolation between that point and the surface, taken to be 2008 when the
lake was cored. The identification of nonnative Eucalyptus pollen in the topmost sample
lake was cored. The identification of nonnative Eucalyptus pollen in the topmost sample
(1 cm depth), which only became widespread in the Andes after c. 1960 [37], confirms that
(1 cm depth), which only became widespread in the Andes after c. 1960 [37], confirms that
the top of the core is modern.
the top of the core is modern.
Pollen and
Pollen and spores
spores were
were well-preserved
well-preserved and and 136
136 pollen
pollen types
types were
were differentiated.
differentiated. The
The
CONISS analysis produced three statistically significant local pollen
CONISS analysis produced three statistically significant local pollen zones (ACP–1–3). zones (ACP–1–3).
• ACP–1 (768–543 cm; c. 7100–4450 cal BP)
• ACP–1
Sediments(768–543
at thecm; c. 7100–4450
base of the zone calwere
BP) composed of clays with sand rich in car-
Sediments
bonates and Caat(Figures
the base5 of the6).
and zone were
These composed
sediments of clays
gave way with sand
to dark rich inorganic-rich
brown, carbonates
and
(~30%Cacarbon)
(Figureslaminated
5 and 6). These
gyttjasediments
at c. 6380gave way
cal BP to dark
(Figure 5).brown,
The XRForganic-rich
data for S(~30% car-
approxi-
bon) laminated gyttja at c. 6380 cal BP (Figure 5). The XRF data for S approximated
mated that of C derived from the LOI analysis, peaking at c. 5580 cal BP before gradually that of
C derived from the LOI analysis, peaking at c. 5580 cal BP before gradually
decreasing. The elemental analysis (Figure 6) revealed very similar profiles for the terri- decreasing. The
elemental analysisAL,
genous elements: (Figure
Si, K,6)Ti,revealed veryand
Fe, Co, Cu, similar profiles for
Zn (hereafter the terrigenous
terrigenous elements:
elements). These
AL, Si, K,all
elements Ti,increased
Fe, Co, Cu, and Zn (hereafter
in concentration abruptlyterrigenous
c. 5350 calelements).
BP. A further These elements
transition all
to less-
 Plants 2024, 13, x FOR PEER REVIEW 7 of 21
Plants 2024, 13, 1019 7 of 19

carbon rich
increased gyttjas intercalated
in concentration abruptlywith tan clays
c. 5350 occurred
cal BP. A furthernear the top to
transition of less-carbon
the zone asrich
sedi-
mentation rates increased c. 4700 cal BP (Figure 5). The pollen concentrations
gyttjas intercalated with tan clays occurred near the top of the zone as sedimentation rates increased
throughout
increased this cal
c. 4700 zone,
BP reaching a peak
(Figure 5). of c. 300,000
The pollen grains per
concentrations cm3 c. 4500
increased cal BP (Figure
throughout this
7). Poaceae (c. 70%) dominated the pollen spectra with
3 Plantago, Amaranthaceae,
zone, reaching a peak of c. 300,000 grains per cm c. 4500 cal BP (Figure 7). Poaceae Ambrosia,
(c.
and dominated
70%) other Asteraceae as thespectra
the pollen next most
withabundant types.
Plantago, Zea was not found,
Amaranthaceae, and Sporormiella
Ambrosia, and other
spores were
Asteraceae rare.
as the Charcoal
next was abundant
most abundant types. throughout
Zea was not much
found,ofandthisSporormiella
zone. The spores
samples
from ACP–1 had predominantly negative values on both Axis 1 and 2 of
were rare. Charcoal was abundant throughout much of this zone. The samples from ACP–1 the DCA (Figure
8).
had predominantly negative values on both Axis 1 and 2 of the DCA (Figure 8).

Plants 2024, 13, x FOR PEER REVIEW 8 of 21

Figure 5. 5.Stratigraphy
Figure Stratigraphyofofthe Lake
the Acopia
Lake Acopiasediment
sedimentscaled
scaledbyby
age.
age.The
Theright
rightthree panels
three portray
panels portray
sedimentation rates and loss-on-ignition estimates of % organic and % carbonate content.
sedimentation rates and loss-on-ignition estimates of % organic and % carbonate content.

Figure6.6.Results
Figure Resultsof
ofthe
theX-ray
X-rayfluorescence
fluorescenceanalysis
analysisof
ofselected
selectedelements
elementsin
inthe
thesediments
sedimentsofofLake
Lake
Acopia,
Acopia,Peru.
Peru.
Plants 2024, 13, x FOR PEER REVIEW 9 of 21

Plants 2024, 13, 1019 8 of 19

Figure 7. The common fossil pollen types of Lake Acopia. Pollen assemblages through time showing
Figure 7. The common fossil pollen types of Lake Acopia. Pollen assemblages through time showing
taxa
taxawith
with>2%>2%abundance
abundance inin
five or more
five samples.
or more Empty
samples. lineslines
Empty on plots ×5 exaggeration.
showshow
on plots Puna
×5 exaggeration.
vegetation shown in black, forest (non-local) vegetation shown in green, aquatics shown in blue,
Puna vegetation shown in black, forest (non-local) vegetation shown in green, aquatics shown in
and spores shown in brown. The dashed red line on the Sporormiella plot marks 2% abundance. Tick
blue, and spores shown in brown. The dashed red line on the Sporormiella plot marks 2% abundance.
marks on maize plot show sampling resolution for extended maize; + indicates maize presence (red
+Tick marksmaize
indicates on maize plotinshow
found initialsampling
300 grainresolution for extended maize; + indicates maize presence
pollen count).
(red + indicates maize found in initial 300 grain pollen count).

• ACP–2 (542–139 cm; c. 4450–2090 cal BP)

Sediments in ACP–2 were all finely laminated clays and gyttjas with low (~10%)
carbon content (Figure 5). The rate of sedimentation increased markedly c. 3200 cal BP, but
then declined to near its lowest level of the record at c. 2300 cal BP. The concentrations of
terrigenous elements were volatile but showed a marked temporary decline at c. 2300 cal
BP (Figure 6). The pollen spectra were broadly similar to those of the preceding zone except
for higher abundances of Amaranthaceae, Asteraceae, and Alnus, with peaks of c. 41, 22%,
and 10%, respectively (Figure 7). Samples from ACP–2 were largely segregated from all
other samples by having negative values on DCA Axis 1 and positive values on Axis 2
(Figure 8) due to Zea being first recorded at c. 4400 cal BP, with no further record until c.
3700 cal BP, whereafter it was found regularly (Figure 7). Sporormiella was also consistently
present at above 2% and reached a peak of c. 31% at 3640 cal BP. Three major spikes of fire
activity occurred within the zone but, overall, charcoal was rarer than in ACP–1.
• ACP–3 (138–0 cm; c. 2090–0 cal BP)
The gross morphology of sediments in ACP-3 were similar to those of ACP–2, although
laminations were not as well developed (Figure 5). Sedimentation rates could not be
measured in this zone but are inferred to have remained low based on a linear rate of
sedimentation from the last reliable 14 C age to the surface. The terrigenous elements show
increasing concentrations until c. 1000 cal BP, before plateauing and being less variable than
in ACP–2 (Figure 6). Poaceae and Amaranthaceae pollen continued to decline in abundance
 Plants 2024, 13, 1019 9 of 19

as Alnus, Asteraceae, and Brassicaceae showed increases. Zea was found in the basal
samples of the zone but was absent between c. 1780 and 260 cal BP (Figure 7). The samples
Plants 2024, 13, x FOR PEER REVIEW from this zone all had positive scores on DCA Axis 1 and were almost entirely segregated
10 of 21
from those of the other zones (Figure 8). Sporormiella rose in abundance throughout this
zone before showing a sharp spike to its peak abundance (37%) at c. 460 cal BP. Charcoal
was present in low and declining quantities throughout this zone.

Zone ACP1
Zone ACP2

Zone ACP3

Sporormiella >2%
Sporormiella <2%

Figure
Figure8. 8.
Results of the
Results Detrended
of the Correspondence
Detrended Analysis
Correspondence (DCA)
Analysis of fossil
(DCA) of pollen data from
fossil pollen dataLake
from
Acopia. Local pollen
Lake Acopia. Local zones
pollenare color
zones arecoded, and samples
color coded, that contained
and samples more or
that contained less or
more than
less2% Spo-
than 2%
rormiella spores (spores were excluded from the analysis) are marked by shape.
Sporormiella spores (spores were excluded from the analysis) are marked by shape.

• 4. Discussion
ACP–2 (542–139 cm; c. 4450–2090 cal BP)
Sediments
The Lake in ACP–2
Acopia wereprovides
record all finely alaminated clays and
detailed history gyttjas
of the rise with low (~10%)
of a complex car-
agrarian
bon
system in the high Andes of southern Peru. We did not capture lake formation, but the but
content (Figure 5). The rate of sedimentation increased markedly c. 3200 cal BP, start
then declined
of the record,to near was
which its lowest
in the level
midstofofthetherecord at c. 2300Dry
Mid-Holocene cal BP.
EventThe concentrations
(MHDE), documents of
terrigenous elements were volatile but showed a marked temporary
a shallow system with substantial deposition of CaCO3 (Figures 5 and 6). The lake begandecline at c. 2300 cal
BPto(Figure
deepen6). Thec.pollen
after 6500 calspectra
BP aswere broadly similar
the evaporitic to those
influence of the
that led precedingdeposition
to carbonate zone ex-
cept for higher
weakened. Atabundances
the onset of of thisAmaranthaceae,
record, a puna Asteraceae,
grassland isand Alnus, with
suggested by thepeaks of c. 41,of
dominance
22%, and 10%,
Poaceae respectively
and low (Figure
levels of fire 7). Samples
activity. from ACP–2
Large herbivores werewere
rare largely segregatedand
in the landscape, fromthis
allwas
other samples
a period by having
between negative
the demise of values on DCA Axis
the megafauna 1 and
[13] but positive
before values
the rise on Axis 2
of pastoralism,
(Figure 8) due topressures
when grazing Zea being firstrelatively
were recordedlight.
at c. 4400 cal BP,
Indeed, the with
lack ofnograzing
furtherindicated
record until c.
by the
3700 calof
rarity BP, whereafter(Figure
Sporormiella it was found
7) mayregularly
account (Figure
for this 7). Sporormiella
basal zone (ACP–1)was also consistently
being segregated
present at above
from other zones2%inand
thereached a peak
DCA results of c. 31%
(Figure 8). at 3640 cal BP. Three major spikes of fire
activity occurred within the zone but, overall, charcoal was rarer than in ACP–1.
• 4.1.ACP–3
How Did Climate
(138–0 cm;Change
c. 2090–0Influence
cal BP) Land Use around Lake Acopia?
Thegross
The deepening of the lake
morphology at c. 6500 cal
of sediments in BP occurred
ACP-3 wereduring
similarthetopeak
thoseofoftheACP–2,
MHDEalt-[25],
and does
hough not standwere
laminations out as being
not a significant
as well developed turning
(Figure point in other regionalrates
5). Sedimentation records [53–55].
could not
beConsequently, the zone
measured in this transition
but areto inferred
more organic
to have sediment
remained maylowreflect
basedlocal
on a climate effects
linear rate of
or that sufficient
sedimentation fromclays hadreliable
the last accumulated
14C ageto
to seal the basin
the surface. and
The organic material
terrigenous elements began
showto
accumulate
increasing as a permanently
concentrations untilinundated
c. 1000 cal system formed.
BP, before The long-term
plateauing trend less
and being toward wetter
variable
than in ACP–2 (Figure 6). Poaceae and Amaranthaceae pollen continued to declinecave
conditions following the MHDE (Figure 9) was consistent across regional lake and in
isotopic records
abundance [25,55,56].
as Alnus, In the
Asteraceae, andhighly-resolved Acopiaincreases.
Brassicaceae showed XRF data,Zea strong decadal-scale
was found in the
droughts
basal were
samples of frequent
the zoneevents
but was prior to 4000
absent cal BPc. (Figure
between 1780 and 8).260
Peaks
cal of
BPCa/Ti
(Figure(Figure
7). The9b)
represented
samples fromprolonged
this zone alldrought events [57]
had positive in which
scores on DCA lake evaporation
Axis 1 and were caused
almost carbonate
entirelyto
be deposited,
segregated from while
thoseinwash
of thefromotherterrigenous sources,
zones (Figure 8). characterized
Sporormiella roseby Ti,
infell markedly.
abundance
throughout this zone before showing a sharp spike to its peak abundance (37%) at c. 460
cal BP. Charcoal was present in low and declining quantities throughout this zone.

4. Discussion
The Lake Acopia record provides a detailed history of the rise of a complex agrarian
Plants 2024, 13, 1019 10 of 19

These peaks were also represented in the carbonate deposition measure by loss-on-ignition
(Figure 5). Charcoal fragments >180 µm were evident in almost all samples in this record
(Figure 9c). As regular fire in the Holocene Andes was so rare before human arrival [14], its
regular presence is strongly linked to human activity [13,58]. Consequently, the increased
abundance of charcoal strongly suggested a human presence at this location for the last
7140 years. Nevertheless, peaks of charcoal clearly aligned with peaks in Ca/Ti, and it was
likely that during times of extreme drought normal burning practices escaped to become
Plants 2024, 13, x FOR PEER REVIEW
larger wildfires. The pattern of fire events at Acopia, which reached a peak between12 of 21
6000
and 4000 cal BP, was also seen at Lake Pacucha [54] (Figure 1).

Drought events Complex Agrarian

- − 14

Wetter ->
a) Huagapo Cave

δ18O
1000 - - − 12
severity ->

b) Ca/Ti
Drought

0-

mm2/cm3
c) Charcoal - 10
50 0
d) Zea
e) Poac/Aster
0 40
f) Amaranthaceae
20
%
4 0
% g) Solanaceae
0
- 10
h) Brassicaceae %
0
25
% i) Spororm iella
- 0.6
Vegetation

0
change

j) DCA Axis 1 pollen data

- 0.4
7000 6000 5000 4000 3000 2000 1000 0
Age cal BP

Figure9.9.Evidence
Figure Evidenceof of climate
climate change,
change, cultivation,
cultivation, and and vegetation
vegetation trajectories
trajectories from Acopia,
from Lake Lake Acopia,
Peru,
Peru, relative
relative to a regional
to a regional climate climate
history. history.
Isotopic Isotopic datafrom
data (a) are (a) are from Huagapo
Huagapo Cave [56],Cave [56], data
all other all other
are
data are from Lake Acopia. Gold bars highlight local drought events defined by high
from Lake Acopia. Gold bars highlight local drought events defined by high Ca/Ti ratios. Bold lineCa/Ti ratios.
Bold line on b represents a 20-yr running mean, with full resolution data as paler line.
on b represents a 20-yr running mean, with full resolution data as paler line.

Zeamays
Zea mays(maize)
(maize)was wasfirst
firstobserved
observedatatAcopia
Acopiac.c.4400
4400calcalBPBP(Figure
(Figure9d)
9d)during
duringone one
of the dry events. This date is quite early for maize agriculture in southern
of the dry events. This date is quite early for maize agriculture in southern Peru [59], and Peru [59], and
ititmay
mayhave havebeen
beengrown
grownto toprovide
providecorn-beer
corn-beer(chicha)
(chicha)rather
ratherthan
thanas asaadietary
dietarystaple
staple[60].
[60].
Poaceae pollen was generally present at >30% of the pollen sum,
Poaceae pollen was generally present at >30% of the pollen sum, and Asteraceae was and Asteraceae was often
present
often at >10%
present (Figure
at >10% 7). The7).ratio
(Figure Theofratio
these
oftwo
thesetypes,
two however, has been
types, however, used
has been to used
indicate
to
major changes
indicate in moisture
major changes availability
in moisture [61]. At Acopia,
availability [61]. At the Poaceae/Asteraceae
Acopia, ratio (Fig-
the Poaceae/Asteraceae
ure 9e)
ratio tracked
(Figure 9e) the Ca/Tithe
tracked record,
Ca/Tisuggesting that droughts
record, suggesting induced grassland
that droughts expansion
induced grassland
at the expense
expansion at theofexpense
herbs and shrubs.
of herbs and shrubs.
InInanother
anotherpotential
potentialindication
indicationofofcropcropactivity,
activity,thethepollen
pollenofofAmaranthaceae
Amaranthaceaeshowed showed
a marked increase in abundance at c. 4400 cal BP (Figure 9f).
a marked increase in abundance at c. 4400 cal BP (Figure 9f). Although Amaranthaceae Although Amaranthaceae
canbe
can beweedy
weedyherbs
herbsof ofhigh
highAndean
Andeanlandscapes
landscapesand andwetlands,
wetlands,the thepollen
pollenwas
wasquinoa-type,
quinoa-type,
i.e.,ititwas
i.e., wasmorphologically
morphologicallyattributable
attributableto toaasubset
subsetofofAmaranthaceae
Amaranthaceaethat thatinclude
includethe the
pseudo-cerealquinoa
pseudo-cereal quinoa (Chenopodium
(Chenopodium quinoa).
quinoa). In the
In the modern
modern system,
system, where where
little little
quinoa quinoa
was
observed,
was observed,Amaranthaceae
Amaranthaceae inputsinputs
were about
were 2–4%,
about but
2–4%,its past abundance
but its betweenbetween
past abundance c. 4400
c. 4400 and 2800 cal BP was as high as 40% (Figure 9f). Other pollen records from the
Cuzco region of Peru inferred quinoa cultivation with pollen abundances of about 10–20%
[54,62,63]. It is very likely that quinoa cultivation peaked between 4300 and 3000 cal BP,
but declined after 2800 cal BP; this is a pattern that replicates that of other sites in southern
Peru where maize increases in apparent importance c. 2800 cal BP [63]. Another potential
Plants 2024, 13, 1019 11 of 19

and 2800 cal BP was as high as 40% (Figure 9f). Other pollen records from the Cuzco region
of Peru inferred quinoa cultivation with pollen abundances of about 10–20% [54,62,63]. It is
very likely that quinoa cultivation peaked between 4300 and 3000 cal BP, but declined after
2800 cal BP; this is a pattern that replicates that of other sites in southern Peru where maize
increases in apparent importance c. 2800 cal BP [63]. Another potential crop is represented
by Solanaceae. Potato (Solanum tuberosum) is a native crop of the high Andes that does
not produce much pollen [64], and the grains it does produce are only identifiable at the
family level. Finding Solanaceae pollen at >2% of the pollen sum is very rare (Table 2).
We tentatively suggest that the increase of Solanaceae pollen above the 0–1% level that
characterize the record prior to c. 4400 cal BP (Figure 9g) to levels of 3–4% after 4400 cal BP
could be indicative of potato cultivation.

Table 2. The timing of peak abundances of Brassicaceae and Solanaceae pollen in other Tropical
Andean fossil pollen records. Sites are sorted be decreasing elevation. Yellow highlighting denotes
the sites at an elevation where potatoes or maca could potentially be grown.

Max % of Peak Time Max % Peak Time Site Elevation

Source
Brassicaceae% Cal BP Solanaceae% Cal BP (masl)
Cerro Llamoca 11 600 2 400 4450 [65]
Caserococha 3 3000 3 2000 3900 [66]
Titicaca 3 4000 0.6 4720 3810 [53]
Acopia 11.3 424 3.5 4280 3750 this ms
Huaypo 3.7 340 0.4 3500 [63]
Natosa Peat bog 5 4000 <2 3482 [67]
Challacaba 7 300 <2 3400 [68]
Marcacocha 30 −20 <2 3350 [62]
Chochos 1 1.9 3285 [69]
Pacucha 6 270 1.5 3100 [54]

At c. 4400 cal BP, coincident with maize cultivation and the observed increase of
quinoa-type pollen, the abundance of Sporormiella spores rises above 2% of the pollen
sum for the first time (Figure 9h). Sporormiella is an obligate dung fungus, and in this
context its spores are an indicator of camelid presence. Prior studies have suggested that
when Sporormiella rises above 2% of the pollen sum, significant numbers of grazers are
locally present [70,71]. That there is no prior record of camelids around the lake, coupled
with the onset of maize cultivation, is taken to indicate the introduction of domesticated
animals to this landscape. These observations and the increase in quinoa could indicate the
synergistic inception of corralling and cultivation envisioned by Kuznar (1993) [23], but
perhaps it more likely reflects the arrival of a group already using those techniques in the
Acopia basin.
The initial occurrence of both maize and Sporormiella occurred in a time of low fire
activity and no indication of drought. A drought episode at c. 4200 cal BP, when there was
a spike of charcoal as well as other drought indicators, saw declines in all the indicators of
agrarian activity (Figure 9c,f). As humid conditions returned at c. 4000 cal BP, Sporormiella
increased in abundance, and maize pollen was found in most samples (Figure 9d,i). Sus-
tained Sporormiella occurrence was negatively related to Ca/Ti values, indicating that
pastoralism was sensitive to extreme drought. Quinoa-type pollen reached its peak of
41% at c. 3300 cal BP, coinciding with the only samples where maize pollen grains were
sufficiently abundant to be found in the original count of 300 grains rather than only en-
countered in extended counts. The overall trajectory of vegetation change was documented
in the results of a multivariate analysis of all fossil pollen types (but not spores) using
detrended correspondence analysis (DCA). Axis 1 scores of the DCA (Figure 9j) showed
little consistent pattern until c. 4400 cal BP. Thereafter, the vegetation composition followed
a new path, which increasingly diverged from the Mid-Holocene state and appeared to
accelerate after c. 2800 cal BP. These vegetational trajectories reflected the intensifying
impacts of both cultivation and pastoralism on the system. A pollen type that shows a
Plants 2024, 13, 1019 12 of 19

marked increase at c. 1300 cal BP is Brassicaceae (Figure 9h, Table 2). Brassicaceae pollen
cannot be identified below the family level, and they can be weeds of arable land. The
pattern of Brassicaceae pollen presence at <2% during phases of maize cultivation and
herding provides a baseline for this kind of weedy input. As Brassicaceae pollen is found
in multiple samples at >10%, we infer that it was being grown as a crop. Maca (Lepidium
meyeni) is a root crop in the Brassicaceae that is grown between 3800 and 4300 m asl. Maca
provides a protein-rich flour, and a drink that promotes stamina [72] and is reputed to assist
in animal and human fertility at high elevations [73]. Pearsall [74] found maca wild-type
macroremains in Panaulauca Cave in Central Peru (Figure 1) that dated to c. 3550 cal BP,
but modern-sized maca roots were represented by c. 950 cal BP. She also suggested a link
to camelid husbandry as it may have been an important source of iodine and iron in the
camelid diet [72]. Furthermore, the disturbance and fertilization by camelids might favor
maca [74]. Fossil pollen records of maca are scarce, but a spike of Brassicaceae to 30% in a
single sample from Marcacocha (Table 2) may have been be linked to maca cultivation [62].
We infer that around 4400 cal BP a complex agrarian system developed in which
domesticated camelids were raised in combination with quinoa, maize, and potato cultiva-
tion. The initial stages of these innovations appear to have been somewhat susceptible to
disruption by drought as they show downturns of representation during droughts signaled
by the Ca/Ti ratios (Figure 9). As late Holocene climates stabilized, and the management
of the system matured, crop representation became less variable. While each component of
this agropastoral system is already attested in other parts of the Andes by this date, the
Acopia data provide new insights into the scale at which these different components were
being used at this time, as well as their interconnectedness within the local subsistence
economy. That these components can be inferred as being intensely exploited together
at such an early date demonstrates the longevity of complex agrarian landscapes in the
Andes, well before the emergence of complex state societies.

4.2. Can We Detect Adaptations to Environmental Change in the Way the Land Was Used?
During transition out of the MHDE, significant changes were taking place in the
watershed that foreshadowed the need for adaptation. Between c. 5300 and 2300 cal BP,
the sediment chemistry indicated large variance in the abundance of elements associated
with erosion. Pulses of erosion were indicated by elevated values of inwash of terrigenous
inputs, e.g., Ti, Fe, and Rb/Sr, during wet events, and lower values as runoff reduced
during drier ones (Figure 10c–e). Despite the period from c. 2300 cal BP to the modern
period being even wetter (Figure 10a) than the period with high rates of erosion [55], the
variance in terrestrial inputs disappeared, and deposition slowed to its lowest level of the
record (Figure 10b–e).
The time period from c. 2290 cal BP to the modern period stands out in the Acopia
data because the mean levels of terrigenous input (Ti) and soil erosion (Rb/Sr) and the
overall rate of sediment accretion decreased to values similar to those of c. 5000 cal BP, i.e.,
before maize agriculture and pastoralism (Figure 10). Thus, despite evidence of intense
land use and increasing precipitation, the amount of erosion paralleled that of an earlier
time with a less disturbed landscape.
We infer that this pattern of reduced sedimentation at 154 cm depth, i.e., 2 cm above
the last reliable 14 C age, was due to the onset of terracing in the catchment of Lake Acopia.
While deposition of Sporormiella spores continued after terraces were constructed, the
representation of maize pollen was interrupted from c. 1700 cal BP until c. 300 cal BP. Such
interruption does not necessarily imply that maize was no longer being cultivated, however.
Maize pollen, which because of its size (>80 mm) was most likely carried by surface flow
into lakes, may have been trapped by newly constructed terraces before entering the
lake. Sporormiella, on the other hand, was directly deposited by animals defecating on
the shoreline while drinking [70]. Maca pollen may also have been directly defecated by
camelids if they were fed flowers and foliage.
Plants 2024, 13, x FOR PEER REVIEW 15 of 21
Plants 2024, 13, 1019 13 of 19

Complex Inferred
agrarian terracing
− 14 - a) Pumacocha

Wetter ->

δ18O
- 0.2
− 12 -

cm/yr
b) Sedimentation
rate - 0.1
800

400 c) Ti
cps

- 0.5
0
d) Rb/Sr
5- - 0
e) PCA 1
0-
XRF

-5 -
f) Zea 40
%
g) Spororm iella
20

0
7000 6000 5000 4000 3000 2000 1000 0
Age cal BP
Figure 10. Multiproxy
Figure 10. Multiproxy data
data from
fromLake
LakeAcopia
Acopiashowing
showingchanges
changesininland
landuse
useand
and erosion
erosion in in relation
relation to
to regional climate and inferred terracing. Isotopic data (a) from Lake Pumacocha [55]. Bold line on
regional climate and inferred terracing. Isotopic data (a) from Lake Pumacocha [55]. Bold line on c,
c, d, and e represents 20-yr running mean, with full resolution data as paler line. Red horizontal
d, and e represents 20-yr running mean, with full resolution data as paler line. Red horizontal lines
lines show mean value of Ti input during five periods, defined by Bayesian changepoint analysis
show mean
within value of Ti input during five periods, defined by Bayesian changepoint analysis within
the record.
the record.
The time period from c. 2290 cal BP to the modern period stands out in the Acopia
Support for terracing may also be seen in the Acopia 14 C record, which is compli-
data because the mean levels of terrigenous input (Ti) and soil erosion (Rb/Sr) and the
cated by reversals in the top 1.2 m (last 2000 years) of sediment (Table 1, Figure 4). Age
overall rate of sediment accretion decreased to values similar to those of c. 5000 cal BP,
reversals are not uncommon in series of 14 C dates from paleo sequences where people
i.e., before maize agriculture and pastoralism (Figure 10). Thus, despite evidence of in-
have induced landscape erosion [65,75,76]. In these systems, erosion creates an influx of
tense land use and increasing precipitation, the amount of erosion paralleled that of an
“old carbon” from the watershed, contaminating the signal of 14 C that accumulates in
earlier time with a less disturbed landscape.
lake sediments [77,78]. In Amazonia and the Andes, the dating of basement soils—-i.e.,
We infer that this pattern
C horizons—-commonly of reduced
produces 14 C agessedimentation at 154 cm
in excess of 15,000 yearsdepth,
[79],i.e.,
and2thus
cm above
these
the last reliable 14C age, was due to the onset of terracing in the catchment of Lake Acopia.
subsoils contain a substantial “old carbon” component. The two ages with problematic
While
dates, deposition
from samples of Sporormiella
obtained betweenspores 81 continued
and 120 after
cm atterraces
Acopia,were constructed,
are inferred to be the rep-
in error
resentation of maizebetween
(i.e., the difference pollen was interrupted
inferred age andfrom c. 1700 age)
measured cal BPbyuntil c. 300
c. 1530 andcal1500
BP. years
Such
interruption does not necessarily imply that maize was no longer
respectively, while the two overlying dates from 7 and 71 cm had even larger errors of being cultivated, how-
ever.
c. 5650 Maize pollen,
and 4170 which
years, because of(Table
respectively its size1).
(>80 mm) was
Broadly, the most
scale likely
of the carried
error mustby surface
reflect
flow into lakes, may have been trapped by newly constructed terraces
the proportion of ‘old’ carbon, or the input of older material. As all four of these prob- before entering the
lake.
lematic dates occur within the period of overall decreased terrigenous input, we offerthe
Sporormiella, on the other hand, was directly deposited by animals defecating on an
shoreline
explanation while drinking
based [70]. construction.
on terrace Maca pollen may also have been directly defecated by came-
lids ifAfter
they vegetation
were fed flowers and foliage.
was cleared to create areas for cultivation or to promote grazing,
Support for terracing may
the soil surface profile was disturbed also be (Figure
seen in the11),Acopia
and the14C record, which
landscape becameis complicated
more prone
by reversals in the top 1.2 m (last 2000 years) of sediment (Table
to erosion [80]. Erosion then increased the influx of “old carbon” from sediments, 1, Figure 4). Age reversals
which
are not uncommon
increases in seriessoils
in age as deeper of 14Canddates
clayfrom paleoare
subsoils sequences
eroded. where
Earthen people have
terraces induced
were built
landscape
by diggingerosion [65,75,76].
out a portion In from
of soil theseupslope,
systems,then erosion
usingcreates an influx
that material toof
fill“old carbon”
the terraced
from the watershed,
structure below [81].contaminating the signal
This process resulted inof 14C that accumulates in lake sediments
a sediment profile where red clays from
[77,78]. In Amazonia
eroded Andean and[82]
subsoils the Andes, the dating
were deposited of basement
with landscapesoils––i.e.,
disturbance. C horizons––com-
monly produces 14C ages in excess of 15,000 years [79], and thus these subsoils contain a
 After vegetation was cleared to create areas for cultivation or to promote grazing, the
soil surface profile was disturbed (Figure 11), and the landscape became more prone to
erosion [80]. Erosion then increased the influx of “old carbon” from sediments, which in-
creases in age as deeper soils and clay subsoils are eroded. Earthen terraces were built by
digging out a portion of soil from upslope, then using that material to fill the terraced
Plants 2024, 13, 1019 14 of 19
structure below [81]. This process resulted in a sediment profile where red clays from
eroded Andean subsoils [82] were deposited with landscape disturbance.

old
carbon

B
Original soil profile with
high elevation
vegetation. Carbon age
increases with depth.

C Dig
Soil profile loosened
Fill with cleared vegetation.
Dig Increased susceptibility
Fill to erosion.
Dig
D
Fill
Terrace construction
old profile on slope
carbon
exposed

Topsoil
Final soil profile of
Loosened topsoil / subsoil terrace. Decreased
horizon susceptibility to erosion.
Clay subsoil Old carbon mixed into
surface sediment and
Mixed topsoil and clay subsoil exposed.

Figure 11. Hypothesized influence of terracing upon Andean landscapes and old carbon input.
Figure 11. Hypothesized influence of terracing upon Andean landscapes and old carbon input.
Over time, terracing would generally have stabilized landscapes and reduced erosion,
but Over time,terrace
renewed terracing would generally
construction would havehavestabilized
producedlandscapes and reduced
pulses of erosion withinero-
the
sion, but renewed terrace construction would have produced pulses of erosion
landscape as they were built. Thus while all age reversals were probably caused by older within the
landscape as they
material being were built.
washed Thus
into the while
lake, theall age reversals
erosion patternswere probably
shown caused
by the XRF bysuggest
data older
material being washed
an additional processinto the lake,
impacting theage
the erosion patterns
reversals at 7 shown
and 71by theIncreased
cm. XRF data values
suggestof
an additional
Rb/Sr process
suggest moreimpacting the age reversals
intense weathering in the at 7 and 71 cm.
landscape Increased
associated values
with clay of Rb/Sras
input,
suggest more associated
Rb is mostly intense weathering
with fine in theand
silts landscape associated
clays [83,84]. Whenwith clay input,
compared with as Rb is
sediment
mostly associated with fine silts and clays [83,84]. When compared with sediment
stratigraphy, red clays began depositing again c. 1100 cal BP, coincident with the increase in
Rb/Sr and larger error in the 14 C ages (Figure 10d). Therefore, the increasing age reversals
in the top two ages of the core interpreted with Rb/Sr and red clay deposit could reflect a
later significant expansion of terracing around Lake Acopia.
We note that the beginning of this expansion coincides with the peak of the Wari cul-
ture [85], which is known to have exploited terracing systems throughout their empire [86].
Terrace expansion at Acopia then reaches a peak c. 550 cal BP during the period of Inkan ex-
pansion, again a time of pronounced landscape transformation and extensive terracing [87].
The trough in Rb/Sr at c. 420 cal BP occurs at the time of Spanish conquest [88], and may
be the consequence of depopulation [89] and temporary terrace abandonment.
The identification of widescale terracing in the Acopia lake records is significant method-
ologically and anthropologically. Methodologically, most archaeologically-determined terrace
chronologies rely on dating basal soils [90] or the soils beneath terrace walls [91], but the
inherently disturbed nature of the setting produces results with very large uncertainties.
Moreover, it is challenging to infer the scale of regional terracing from a handful of terrace
Plants 2024, 13, 1019 15 of 19

excavations, making it difficult to evaluate the scale or intensity of land use during any
one period. By investigating the environmental consequences of terraces (i.e., pronounced
changes in sedimentation), a more representative landscape-wide signal of terracing can be
measured and dated. At a broader scale, we have encountered many lakes in the Andes
where the sedimentary record exhibited significant 14 C reversals in the last 2000 years or
where the youngest age at the core surface was about 2000 years. We came to know this
as the ‘paleo Y2K problem’. The basins were frequently steep-sided and terraced, and we
infer that a similar analysis to that reported here would determine the onset of terracing in
those systems as well.
As for the anthropological significance of these results, the emergence of widescale ter-
racing around 2300 BP provides strong support for the handful of similarly early published
terrace dates. Terrace excavations from the Colca Valley, for example, have reported dates
as early as 3445 ± 85 BP [29], while the earliest dates from the Chichas-Soras Valley were
reported at 3537 ± 28 BP and 2562 ± 36 [91]. For many such early dates, there is skepticism
and uncertainty around whether they actually relate to terrace construction, or even to
human activity. Our results lend support to the idea that terracing was practiced in the
Andes well before the rise of centralized state societies.
We observe that that the apparent expansion of terracing c. 2300 cal BP coincided
with the strongest drought in almost 2000 years. This drought may have played a role in
motivating strategies to retain water on cultivated areas and maintain productivity in the
face of climatic change. Yet to rest on a monocausal climatic explanation for terracing at
Acopia would be to ignore the sequence and history of land-use practices that preceded
terracing. Erosion rates were already increasing well before 2300 cal BP as a result of long-
term exploitation of the landscape’s hillslopes for crop cultivation and camelid grazing—
practices that themselves waxed and waned in intensity with periods of aridity throughout
the MHDE. The decision of farmers around Acopia to invest in terraces 2300 years ago was
therefore informed as much by the climatic conditions at that time as it was by the landscape
and suite of crops, animals, and technologies they had inherited from their ancestors.

5. Conclusions
The data from Lake Acopia suggest that ancient people who lived near the lake
continually adapted their land-use practices to changing conditions—-and that they were
driven to do so not by a single factor but by the intersecting effects of climate change
and the unforeseen consequences of their own, prior land use strategies. Previous studies
have shown that Andean grasslands were already modified by the loss of megafauna and
increased fire activity in the early Holocene, before the start of this record. The onset of
pastoralism and mixed crop usage at Acopia c. 4400 cal BP was thus a continuation of much
longer-term patterns of human landscape modification in the southern Peruvian Andes.
Over time, crop cultivation, fire, and trampling caused top soil to become unstable; then,
farmers built terracing c. 2300 cal BP to mitigate this process. The acute trigger to initiate
the terracing effort may have been a drought, suggesting that local societies were sensitive
to short-term climate change, but the ultimate need for soil stabilization was the result
of a millennium of elevated erosion. Once present, terracing caused an abrupt decrease
in terrigenous elements, including surface-washed pollen entering the lake. Despite the
continuous use of the landscape and increasing precipitation after 2300 cal BP, the effects
of this major alteration to the watershed have lasted until the present. The appearance
of terracing coincides with progressively less accurate 14 C ages in the upper part of the
sediment core, and the generalizable pattern of a rapid reduction in sedimentary rate and
intense age reversals near the top of the core should be investigated for further evidence
of terracing.
Plants 2024, 13, 1019 16 of 19

Author Contributions: C.R.S. conducted pollen analyses; M.B.B., A.R.-D. and B.G.V. conducted
the fieldwork; R.D.B. and C.D. conducted charcoal analyses; L.B. conducted extended crop pollen
searches; L.C.P. conducted XRF analyses; D.P. conducted terracing surveys; M.B.B., D.P., A.R.-D.,
B.G.V., P.V. and L.C.P. wrote the manuscript; M.B.B. conceived the project. All authors have read and
agreed to the published version of the manuscript.
Funding: This research was funded by the National Geographic Society HJ-148R-17. Funding
for all other analyses was from NSF grant EAR 1624207 to MBB, and a Research Experience for
Undergraduates award to Florida Institute of Technology NSF REU 1359341.
Data Availability Statement: The datasets generated from this study are available through GITHUB
which include 14C, pollen, charcoal, loss-on-ignition (carbonate) and XRF (Ti, Si, and Ca).
Acknowledgments: We thank the government and people of Peru for allowing us to conduct
fieldwork in Peru. We are indebted to our funders.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Myers, N.; Mittermeier, R.A.; Mittermeier, C.G.; da Fonseca, G.A.B.; Kent, J. Biodiversity hotspots for conservation priorities.
Nature 2000, 403, 853–858. [CrossRef] [PubMed]
2. Rodríguez-Echeverry, J.; Leiton, M. State of the landscape and dynamics of loss and fragmentation of forest critically endangered
in the tropical Andes hotspot: Implications for conservation planning. J. Landsc. Ecol. 2021, 14, 73–91. [CrossRef]
3. Fernández, J.; Markgraf, V.; Panarello, H.O.; Albero, M.; Angiolini, F.E.; Valencio, S.; Arriaga, M. Late Pleistocene/Early Holocene
environments and climates, fauna, and human occupation in the Argentine Altiplano. Geoarchaeology 1991, 6, 251–272. [CrossRef]
4. Brunschön, C.; Behling, H. Reconstruction and visualization of upper forest line and vegetation changes in the Andean depression
region of southeastern Ecuador since the last glacial maximum—A multi-site synthesis. Rev. Palaeobot. Palynol. 2010, 163, 139–152.
[CrossRef]
5. Raczka, M.F.; Mosblech, N.A.; Giosan, L.; Folcik, A.M.; Valencia, B.G.; Kingston, M.; Baskin, S.; Bush, M.B. A human role in
Andean megafaunal extinction? Quat. Sci. Rev. 2019, 205, 154–165. [CrossRef]
6. Rozas-Davila, A.; Rodbell, D.; Bush, M.B. Pleistocene megafaunal extinction in the grasslands of Junin-Peru. Quat. Res. 2023, 50,
755–766. [CrossRef]
7. Piperno, D.R. Phytoliths: A Comprehensive Guide for Archaeologists and Paleoecologists; Alta Mira Press: Lanham, MD, USA, 2006;
p. 304.
8. Piperno, D.R. The Origins of Plant Cultivation and Domestication in the Neotropics. In Foraging Theory and the Transition to
Agriculture; Kennett, D., Winterhalder, B., Eds.; University of California Press: Berkeley, CA, USA, 2006; pp. 137–166.
9. McMichael, C.; Witteveen, N.; Scholz, S.; Zwier, M.; Prins, M.; Lougheed, B.; Mothes, P.; Gosling, W. 30,000 years of landscape
and vegetation dynamics in a mid-elevation Andean valley. Quat. Sci. Rev. 2021, 258, 106866. [CrossRef]
10. Borrero, L.A. The human colonization of the high Andes and Southern South America during the cold pulses of the late Pleistocene.
In Hunter-Gatherer Behavior; Routledge: Abingdon, UK, 2016; pp. 57–78.
11. Le Neün, M.; Dufour, E.; Zazzo, A.; Tombret, O.; Thil, F.; Wheeler, J.C.; Cucchi, T.; Goepfert, N. Holocene occupation of the
Andean highlands: A new radiocarbon chronology for the Telarmachay rockshelter (Central Andes, Peru). Quat. Sci. Rev. 2023,
312, 108146. [CrossRef]
12. Lynch, T.F. Quaternary climate, environment, and the human occupation of the south-central Andes. Geoarchaeology 1990, 5,
199–228. [CrossRef]
13. Bush, M.B.; Rozas-Davila, A.; Raczka, M.; Nascimento, M.; Valencia, B.; Sales, R.K.; McMichael, C.N.H.; Gosling, W.D. A
paleoecological perspective on the transformation of the tropical Andes by early human activity. Philos. Trans. R. Soc. Lond. B Biol.
Sci. 2022, 377, 20200497. [CrossRef]
14. Schiferl, J.; Kingston, M.; Åkesson, C.; Valencia, B.; Rozas-Davila, A.; McGee, D.; Woods, A.; Chen, C.; Hatfield, R.; Rodbell, D.
A neotropical perspective on the uniqueness of the Holocene among interglacials. Nat. Commun. 2023, 14, 7404. [CrossRef]
[PubMed]
15. Sarmiento, F.O. Breaking mountain paradigms: Ecological effects on human impacts in managed tropandean landscapes. AMBIO
J. Hum. Environ. 2000, 29, 423–431. [CrossRef]
16. Erickson, C.L. The Lake Titicaca basin: A precolumbian built landscape. In Imperfect Balance; Lentz, D.E., Ed.; Columbia University
Press: New York, NY, USA, 2000; pp. 311–356.
17. Rademaker, K.; Hodgins, G.; Moore, K.; Zarrillo, S.; Miller, C.; Bromley, G.R.; Leach, P.; Reid, D.A.; Álvarez, W.Y.; Sandweiss, D.H.
Paleoindian settlement of the high-altitude Peruvian Andes. Science 2014, 346, 466–469. [CrossRef] [PubMed]
18. Garvey, R.; Poe, K.; Pratt, L. Spatial and temporal trends in Peru’s radiocarbon record of middle Holocene foragers. Quat. Int.
2023. [CrossRef]
Plants 2024, 13, 1019 17 of 19

19. Dillehay, T.D.; Goodbred, S.; Pino, M.; Sánchez, V.F.V.; Tham, T.R.; Adovasio, J.; Collins, M.B.; Netherly, P.J.; Hastorf, C.A.; Chiou,
K.L.; et al. Simple technologies and diverse food strategies of the Late Pleistocene and Early Holocene at Huaca Prieta, Coastal
Peru. Sci. Adv. 2017, 3, e1602778. [CrossRef]
20. Piperno, D.R. The origins of plant cultivation and domestication in the New World tropics: Patterns, process, and new develop-
ments. Curr. Anthropol. 2011, 52, S453–S470. [CrossRef]
21. Pearsall, D.M. Plant Domestication and the Shift to Agriculture in the Andes. In The Handbook of South American Archaeology;
Silverman, H., Isbell, W.H., Eds.; Springer: New York, NY, USA, 2008; pp. 105–120.
22. Yacobaccio, H.D. The domestication of South American camelids: A review. Anim. Front. 2021, 11, 43–51. [CrossRef] [PubMed]
23. Kuznar, L.A. Mutualism between Chenopodium, Herd Animals, and Herders in the South Central Andes. Mt. Res. Dev. 1993, 13,
257–265. [CrossRef]
24. Kuentz, A.; Ledru, M.-P.; Thouret, J.-C. Environmental changes in the highlands of the western Andean Cordillera, southern Peru,
during the Holocene. Holocene 2012, 22, 1215–1226. [CrossRef]
25. Baker, P.A.; Fritz, S.C. Nature and causes of Quaternary climate variation of tropical South America. Quat. Sci. Rev. 2015, 124,
31–47. [CrossRef]
26. Riris, P.; Arroyo-Kalin, M. Widespread population decline in South America correlates with mid-Holocene climate change. Sci.
Rep. 2019, 9, 6850. [CrossRef] [PubMed]
27. Vining, B.R.; Steinman, B.A.; Abbott, M.B.; Woods, A. Paleoclimatic and archaeological evidence from Lake Suches for highland
Andean refugia during the arid middle-Holocene. Holocene 2019, 29, 328–344. [CrossRef]
28. Knapp, G. Andean Ecology: Adaptive Dynamics in Ecuador; Routledge: Abingdon, UK, 2019.
29. Brooks, S.O. Prehistoric Agricultural Terraces in the Rio Japo Basin, Colca Valley, Peru. Ph.D. Dissertation, The University of
Wisconsin-Madison, Madison, WI, USA, 1998; p. 1020.
30. Denevan, W.M. Cultivated Landscapes of Native Amazonia and the Andes; Oxford University Press: Oxford, UK, 2001.
31. Erickson, C.L. Neo-environmental determinism and agrarian ‘collapse’ in Andean prehistory. Antiquity 1999, 73, 634. [CrossRef]
32. Binford, M.W.; Kolata, A.L.; Brenner, M.; Janusek, J.; Seddon, M.T.; Abbott, M.B.; Curtis, J.H. Climate variation and the rise and
fall of an Andean civilization. Quat. Res. 1997, 47, 235–248. [CrossRef]
33. Sandweiss, D.H.; Solís, R.S.; Moseley, M.E.; Keefer, D.K.; Ortloff, C.R. Environmental change and economic development in
coastal Peru between 5800 and 3600 years ago. Proc. Natl. Acad. Sci. USA 2009, 106, 1359–1363. [CrossRef] [PubMed]
34. Meyfroidt, P.; Chowdhury, R.R.; de Bremond, A.; Ellis, E.C.; Erb, K.-H.; Filatova, T.; Garrett, R.; Grove, J.M.; Heinimann, A.;
Kuemmerle, T. Middle-range theories of land system change. Glob. Environ. Chang. 2018, 53, 52–67. [CrossRef]
35. Contreras, D.A. Correlation is not enough: Building better arguments in the archaeology of human-environment interactions. In
The Archaeology of Human-Environment Interactions; Contreras, D.A., Ed.; Routledge: Abingdon, UK, 2016; pp. 17–36.
36. Bruno, M.C.; Capriles, J.M.; Hastorf, C.A.; Fritz, S.C.; Weide, D.M.; Domic, A.I.; Baker, P.A. The rise and fall of Wiñaymarka:
Rethinking cultural and environmental interactions in the Southern Basin of Lake Titicaca. Hum. Ecol. 2021, 49, 131–145.
[CrossRef]
37. Luzar, J. The political ecology of a “forest transition”: Eucalyptus forestry in the Southern Peruvian Andes. Ethnobot. Res. Appl.
2007, 5, 85–93. [CrossRef]
38. Colinvaux, P.; de Oliveira, P.E.; Moreno, P.J.E. Amazon Pollen Manual and Atlas; Harwood Academic Publishers: Amsterdam, The
Netherlands, 1999; p. 332.
39. R Development Core Team. R: A Language and Environment for Statistical Computing; R Foundation for Statistical Computing:
Vienna, Austria, 2018. Available online: https://www.R-project.org/ (accessed on 5 May 2023).
40. Reimer, P.J.; Austin, W.E.; Bard, E.; Bayliss, A.; Blackwell, P.G.; Ramsey, C.B.; Butzin, M.; Cheng, H.; Edwards, R.L.; Friedrich, M.
The IntCal20 Northern Hemisphere radiocarbon age calibration curve (0–55 cal kBP). Radiocarbon 2020, 62, 725–757. [CrossRef]
41. Stuiver, M.; Reimer, P.; Reimer, R. CALIB 8.2 [WWW program]. 2021. Available online: http://calib.org (accessed on 8
August 2023).
42. Heiri, O.; Lotter, A.F.; Lemcke, G. Loss on ignition as a method for estimating organic and carbonate content in sediments:
Reproducibility and comparability of results. J. Paleolimnol. 2001, 25, 101–110. [CrossRef]
43. Faegri, K.; Iversen, J. Textbook of Pollen Analysis, 4th ed.; Wiley: Chichester, UK, 1989.
44. Markgraf, V.; D’Antoni, H.L. Pollen Flora of Argentina; The University of Arizona Press: Tucson, AZ, USA, 1978; p. 208.
45. Hooghiemstra, H. Vegetational and Climatic History of the High Plain of Bogota, Colombia. In Dissertaciones Botanicae 79; J.
Cramer: Vaduz, Liechtenstein, 1984; p. 368.
46. Bush, M.B.; Weng, C. Introducing a new (freeware) tool for palynology. J. Biogeogr. 2007, 34, 377–380. [CrossRef]
47. Whitney, B.S.; Rushton, E.A.; Carson, J.F.; Iriarte, J.; Mayle, F.E. An improved methodology for the recovery of Zea mays and other
large crop pollen, with implications for environmental archaeology in the Neotropics. Holocene 2012, 22, 1087–1096. [CrossRef]
48. Grimm, E.C. CONISS: A fortran 77 program for stratigraphically constrained cluster analysis by the method of incremental sum
of squares. Comput. Geosci. 1987, 13, 13–35. [CrossRef]
49. Rasband, W.S.; Rasband, W.S. ImageJ, U. S. National Institutes of Health, Bethesda, MD, USA. 1997–2016. Available online:
https://imagej.nih.gov/ij/ (accessed on 5 May 2023).
Plants 2024, 13, 1019 18 of 19

50. Richter, T.O.; Van der Gaast, S.; Koster, B.; Vaars, A.; Gieles, R.; de Stigter, H.C.; De Haas, H.; van Weering, T.C. The Avaatech XRF
Core Scanner: Technical description and applications to NE Atlantic sediments. Geol. Soc. Lond. Spec. Publ. 2006, 267, 39–50.
[CrossRef]
51. Killick, R.; Eckley, I. changepoint: An R package for changepoint analysis. J. Stat. Softw. 2014, 58, 1–19. [CrossRef]
52. Blaauw, M.; Christen, J.A. rbacon: Age-Depth Modelling Using Bayesian Statistics, R Package Version 2.4 (2018); Queen’s University
Belfast: Belfast, UK, 2018.
53. Paduano, G.M.; Bush, M.B.; Baker, P.A.; Fritz, S.C.; Seltzer, G.O. A vegetation and fire history of Lake Titicaca since the Last
Glacial Maximum. Palaeogeogr. Palaeoclimatol. Palaeoecol. 2003, 194, 259–279. [CrossRef]
54. Valencia, B.G.; Urrego, D.H.; Silman, M.R.; Bush, M.B. From ice age to modern: A record of landscape change in an Andean cloud
forest. J. Biogeogr. 2010, 37, 1637–1647. [CrossRef]
55. Bird, B.W.; Abbott, M.B.; Rodbell, D.T.; Vuille, M. Holocene tropical South American hydroclimate revealed from a decadally
resolved lake sediment δ18O record. Earth Planet. Sci. Lett. 2011, 310, 192–202. [CrossRef]
56. Kanner, L.C.; Burns, S.J.; Cheng, H.; Edwards, R.L.; Vuille, M. High-resolution variability of the South American summer monsoon
over the last seven millennia: Insights from a speleothem record from the central Peruvian Andes. Quat. Sci. Rev. 2013, 75, 1–10.
[CrossRef]
57. Haberzettl, T.; Corbella, H.; Fey, M.; Janssen, S.; Lücke, A.; Mayr, C.; Ohlendorf, C.; Schäbitz, F.; Schleser, G.H.; Wille, M.
Lateglacial and Holocene wet—Dry cycles in southern Patagonia: Chronology, sedimentology and geochemistry of a lacustrine
record from Laguna Potrok Aike, Argentina. Holocene 2007, 17, 297–310. [CrossRef]
58. White, S. Grass páramo as hunter-gatherer landscape. Holocene 2013, 23, 898–915. [CrossRef]
59. Perry, L.; Sandweiss, D.H.; Piperno, D.R.; Rademaker, K.; Malpass, M.A.; Umire, A.N.; de la Vera, P. Early maize agriculture and
interzonal interaction in southern Peru. Nature 2006, 440, 76. [CrossRef] [PubMed]
60. Washburn, E.; Nesbitt, J.; Burger, R.; Tomasto-Cagigao, E.; Oelze, V.M.; Fehren-Schmitz, L. Maize and dietary change in early
Peruvian civilization: Isotopic evidence from the Late Preceramic Period/Initial Period site of La Galgada, Peru. J. Archaeol. Sci.
Rep. 2020, 31, 102309. [CrossRef]
61. Coetzee, J.A. Pollen analytical studies in east and southern Africa. Palaeoecol. Afr. 1967, 3, 1–146.
62. Chepstow-Lusty, A.J.; Bennett, K.D.; Switsur, V.R.; Kendall, A. 4000 years of human impact and vegetation change in the central
Peruvian Andes—With events parallelling the Maya record? Antiquity 1996, 70, 824–833. [CrossRef]
63. Mosblech, N.A.S.; Chepstow-Lusty, A.; Valencia, B.G.; Bush, M.B. Anthropogenic control of late-Holocene landscapes in the
Cuzco region, Peru. Holocene 2012, 22, 1361–1372. [CrossRef]
64. Bush, M.B. Neotropical plant reproductive strategies and fossil pollen representation. Am. Nat. 1995, 145, 594–609. [CrossRef]
65. Schittek, K.; Forbriger, M.; Mächtle, B.; Schäbitz, F.; Wennrich, V.; Reindel, M.; Eitel, B. Holocene environmental changes in
the highlands of the southern Peruvian Andes (14 S) and their impact on pre-Columbian cultures. Clim. Past 2015, 11, 27–44.
[CrossRef]
66. Paduano, G. Vegetation and fire history of two tropical Andean lakes, Titicaca (Peru/Bolivia), and Caserochocha (Peru) with
special emphasis on the Younger Dryas chronozone. In Department of Biological Sciences; Florida Institute of Technology: Melbourne,
FL, USA, 2001.
67. Villota, A.; Leon-Yanez, S.; Behling, H. Vegetation and environmental dynamics in the Páramo of Jimbura region in the
southeastern Ecuadorian Andes during the late Quaternary. J. S. Am. Earth Sci. 2012, 40, 85–93. [CrossRef]
68. Williams, J.J.; Gosling, W.D.; Coe, A.L.; Brooks, S.J.; Gulliver, P. Four thousand years of environmental change and human activity
in the Cochabamba Basin, Bolivia. Quat. Res. 2011, 76, 58–68. [CrossRef]
69. Bush, M.B.; Hansen, B.C.S.; Rodbell, D.; Seltzer, G.O.; Young, K.R.; Leo, B.; Silman, M.R.; Abbott, M.B.; Gosling, W.D. A 17,000 year
history of Andean climatic and vegetation change from Laguna de Chochos, Peru. J. Quat. Sci. 2005, 20, 703–714. [CrossRef]
70. Davis, O.K.; Shafer, D.S. Sporormiella fungal spores, a palynological means of detecting herbivore density. Palaeogeogr. Palaeocli-
matol. Palaeoecol. 2006, 237, 40–50. [CrossRef]
71. Gill, J.L. Bison Grazing Intensity Predicts the Abundance of the Dung Fungus Proxy Sporormiella at Konza Prairie: Implications for the
Holocene Paleoecology of the Great Plains; Ecological Society of America: Portland, OR, USA, 2012.
72. Ochoa, C. Maca (Lepidium meyenii Walp.; Brassicaceae): A nutritious root crop of the central Andes. Econ. Bot. 2001, 55, 344.
[CrossRef]
73. Hermann, M.; Bernet, T. The Transition of Maca from Neglect to Market Prominence; Agricultural Biodiversity and Livelihoods,
Discussion Papers 1; Bioversity International: Rome, Italy, 2009.
74. Pearsall, D.M. Adaptation of prehistoric hunter-gatherers to the high Andes: The changing role of plant resources. In Foraging and
Farming; Routledge: Abingdon, UK, 2014; pp. 318–332.
75. Bush, M.B.; Correa-Metrio, A.; McMichael, C.H.; Sully, S.; Shadik, C.R.; Valencia, B.G.; Guilderson, T.; Steinitz-Kannan, M.;
Overpeck, J.T. A 6900-year history of landscape modification by humans in lowland Amazonia. Quat. Sci. Rev. 2016, 141, 52–64.
[CrossRef]
76. Rowe, H.D.; Guilderson, T.P.; Dunbar, R.B.; Southon, J.R.; Seltzer, G.O.; Mucciarone, D.A.; Fritz, S.C.; Baker, P.A. Late Quaternary
lake-level changes constrained by radiocarbon and stable isotope studies on sediment cores from Lake Titicaca, South America.
Glob. Planet. Chang. 2003, 38, 273–290. [CrossRef]
Plants 2024, 13, 1019 19 of 19

77. Edwards, K.J.; Whittington, G. Lake sediments, erosion and landscape change during the Holocene in Britain and Ireland. Catena
2001, 42, 143–173. [CrossRef]
78. Kiage, L.M.; Liu, K.-b. Late Quaternary paleoenvironmental changes in East Africa: A review of multiproxy evidence from
palynology, lake sediments, and associated records. Prog. Phys. Geogr. 2006, 30, 633–658. [CrossRef]
79. Bush, M.B.; Listopad, M.C.S.; Silman, M.R. A regional study of Holocene climate change and human occupation in Peruvian
Amazonia. J. Biogeogr. 2007, 34, 1342–1356. [CrossRef]
80. Moody, J.A.; Martin, D.A.; Cannon, S.H. Post-wildfire erosion response in two geologic terrains in the western USA. Geomorphology
2008, 95, 103–118. [CrossRef]
81. Dorren, L.; Rey, F. A review of the effect of terracing on erosion. In Briefing Papers of the 2nd SCAPE Workshop; Boix-Fayons, C.,
Imeson, A., Eds.; 2004; pp. 97–108. Available online: https://www.ecorisq.org/docs/Dorren_Rey.pdf (accessed on 18 March
2024).
82. Molina, A.; Govers, G.; Vanacker, V.; Poesen, J.; Zeelmaekers, E.; Cisneros, F. Runoff generation in a degraded Andean ecosystem:
Interaction of vegetation cover and land use. Catena 2007, 71, 357–370. [CrossRef]
83. Xu, H.; Liu, B.; Wu, F. Spatial and temporal variations of Rb/Sr ratios of the bulk surface sediments in Lake Qinghai. Geochem.
Trans. 2010, 11, 3. [CrossRef] [PubMed]
84. Jin, Z.; Cao, J.; Wu, J.; Wang, S. A Rb/Sr record of catchment weathering response to Holocene climate change in Inner Mongolia.
Earth Surf. Process. Landf. J. Br. Geomorphol. Res. Group 2006, 31, 285–291. [CrossRef]
85. Isbell, W.H. Wari and Tiwanaku: International Identities in the Central Andean Middle Horizon. In The Handbook of South
American Archaeology; Silverman, H., Isbell, W.H., Eds.; Springer: New York, NY, USA, 2008; pp. 731–759.
86. Williams, P.R. Agricultural innovation, intensification, and sociopolitical development: The case of highland irrigation agriculture
on the Pacific Andean watersheds. In Agricultural Strategies; Marcus, J., Stanish, C., Eds.; Cotsen Institute of Archaeology, Univ of
California: Los Angeles, CA, USA, 2006; pp. 309–333.
87. Covey, A.R. The Inca Empire. In Handbook of South American Archaeology; Silverman, H., Isbell, W.H., Eds.; Springer: New York,
NY, USA, 2008; pp. 809–830.
88. Hemming, J. The Conquest of the Incas; Harcourt: New York, NY, USA, 1970.
89. Cook, N.D. Born to Die: Disease and New World Conquest, 1492–1650; Cambridge University Press: Cambridge, UK, 1998; Volume 1.
90. Langlie, B.S. Building ecological resistance: Late intermediate period farming in the south-central highland Andes (CE 1100–1450).
J. Anthropol. Archaeol. 2018, 52, 167–179. [CrossRef]
91. Kemp, R.; Branch, N.; Silva, B.; Meddens, F.; Williams, A.; Kendall, A.; Vivanco, C. Pedosedimentary, cultural and environmental
significance of paleosols within pre-hispanic agricultural terraces in the southern Peruvian Andes. Quat. Int. 2006, 158, 13–22.
[CrossRef]

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.