Professional Documents
Culture Documents
Méndez Et Al., 2015
Méndez Et Al., 2015
Quaternary International
journal homepage: www.elsevier.com/locate/quaint
Mid Holocene radiocarbon ages in the Subtropical Andes (~29 e35 S),
climatic change and implications for human space organization
sar Me
Ce ndez a, *, Adolfo Gil b, c, Gustavo Neme b, Amalia Nuevo Delaunay d,
Valeria Cortegoso c, Consuelo Huidobro e, Víctor Dura n c, Antonio Maldonado f
a
Departamento de Antropología, Facultad de Ciencias Sociales, Universidad de Chile, Santiago, Chile
b
IANIGLA-CONICET, Museo de Historia Natural de San Rafael, Mendoza, Argentina
c
CONICET-Facultad de Filosofía y Letras, Universidad Nacional de Cuyo, Mendoza, Argentina
d
Departamento de Antropología, Facultad de Ciencias Sociales, Universidad Alberto Hurtado, Santiago, Chile
e
UMR 7041 Ethnologie pr ehistorique, Universit
e Paris 1 Panth
eon Sorbonne, Paris, France
f
Centro de Estudios Avanzados en Zonas Aridas, Universidad de La Serena, La Serena, Chile
a r t i c l e i n f o a b s t r a c t
Article history: This article discusses the distribution of radiocarbon age signatures obtained from archeological sites
Available online 22 July 2014 between 29 and 35 S in Central Chile and Midwest Argentina. The goal of this analysis is to establish
bases from which to interpret regional trends in the distribution of the archaeological record that
Keywords: connect these areas, which have been traditionally considered to be geographically decoupled. We
Radiocarbon dates propose a standardized methodology for selecting ages that provide a reliable human signature. Varia-
Climate change
tions in date frequencies in a regional scale are discussed with the use of summed probability distri-
Human paleoecology
butions. Radiocarbon voids at the regional level previously identified in Midwest Argentina are explored.
Hunteregatherers
Mid-Holocene
Regional chronological information is compared to the available paleoenvironmental records, thereby
Subtropical Andes emphasizing the possible role of climate pulses in the spatial organization of human populations. Sig-
nificant arid conditions between 7800 and 5700 cal BP are coincidental with a focused occupation of the
Andes Mountains, an area which may have offered stable resources and thus was more effectively
occupied than other environmental bands.
© 2014 Elsevier Ltd and INQUA. All rights reserved.
http://dx.doi.org/10.1016/j.quaint.2014.06.059
1040-6182/© 2014 Elsevier Ltd and INQUA. All rights reserved.
16 C. Mendez et al. / Quaternary International 356 (2015) 15e26
spatial scale through the use of series of regionally categorized 14C 2009). These antecedents allowed us to generate two preliminary
ages (Araujo et al., 2005-6). hypotheses:
The area of the Andes between 29 and 35 S provides a key
spatial range to evaluate the relationship between the environ- 1. The hiatuses or decreases in the 14C signature of certain areas
mental changes and human responses during the mid-Holocene will coincide with time periods of higher aridity detected
(Fig. 1). In this area, sufficient resolution has been achieved with through regional environmental records.
regard to the environmental changes, their magnitude, and tem- 2. The hiatuses or decreases in the 14C signature of certain regions
poral development. Furthermore, it corresponds to an area will denote positive signatures in neighboring areas with
inhabited by hunteregatherers during the mid-Holocene, whose different ecological characteristics, which would have worked as
most visible response to environmental change was the reconfi- alternative zones for the hunteregatherers during the same
guration of their mobility patterns (Zarate et al., 2005; Me
ndez and period.
Jackson, 2006; Neme and Gil, 2008, 2009). Moreover, the area has
significant ecological differences, thereby representing varied en- To test these hypotheses, we expanded the spatial range to
vironments offering different resources, and thus potential for construct the 14C-age database including all the Andean areas be-
settlement complementarity, as suggested by previous works tween 29 and 35 S and between the Pacific coast and 67 W. This
(Gambier, 1993; Jackson, 2002; Cortegoso et al., 2012). We propose area provides a comparable dataset from archaeological sites
that the quantification and comparison of the set of archaeological studied west and east of the Andes, Chile, and Argentina, respec-
14
C ages can be used to evaluate the settlement preferences of tively. In this paper, we explore the behavior of 14C patterns, discuss
human groups in disparate areas. potential biases, chiefly research decisions and reservoir effect, and
Neme and Gil (2008, 2009) previously discussed 14C date dis- suggest preliminary interpretations linking changes in climate to
tributions to understand the human response to the environmental the spatial organization of hunteregatherers as expressed through
changes that arose from the extreme aridity of the mid-Holocene. the archaeological record.
Their study in the eastern Andes and foothills and plains to the
east between 32 and 38 S showed a marked decrease in dates 2. Regional setting
between 7000 and 6000 BP; and although it was not expressed
uniformly, it was more pronounced in the extra-Andean plains than The study area includes a region between ~29 and 35 S and
in the intermountain valleys and high Andes (Neme and Gil, 2008, from the Pacific coast (72 e71300 ) to 67 W. Prevailing air masses,
14
Fig. 1. Distribution of archaeological sites with C ages between 9000 and 3000 BP in the study area; sites in Chile (circles), sites in Argentina (squares).
C. Mendez et al. / Quaternary International 356 (2015) 15e26 17
the proximity to the sea, and altitude are the three main constraints 3. Mid-Holocene paleoclimate
on the climate and environmental characteristic from the region.
The air masses depend on the actions of two anticyclones, one Holocene paleoclimatic records for the study area are mainly
located to the west (Pacific Anticyclone), seasonally blocking the located on the western Andean slope (Fig. 2). The Pleistoce-
westerlies and allowing mostly winter rains, and the other located neeHolocene transition in this area is characterized by changes in
to the east (Atlantic Anticyclone) permitting summer rains hydrology and precipitation regimes which lead to drier conditions
(Garreaud et al., 2009). Between them, the Andean mountain raises (Lamy et al., 1999; Valero-Garce s et al., 2005) through a step-wise
with heights exceeding 6000 masl, thus producing a strong impact transition mode (Kim et al., 2002; Maldonado et al., 2010).
over the regional climate, vegetation, and faunal distributions. Although resolutions vary, several marine and land records (espe-
In a broad sense, we subdivide this region in three major cially those pollen from sedimentary cores) indicate arid conditions
environments: between 9000 and 5500 cal BP (e.g., Heusser, 1990; Veit, 1996;
Villa-Martínez and Villagra n, 1997; Lamy et al., 1999; Jenny et al.,
a) The Pacific coast and western valleys comprises an area char- 2002; Kim et al., 2002; Valero-Garce s et al., 2005; Maldonado
acterized by dry summers and relatively humid winters, which and Villagra n, 2006). This dry phase was interpreted as the result
is a product of the interaction between the Pacific anticyclone of the southern migration of the Westerlies Wind Belt area of in-
and the northern margin of the Westerlies Wind Belt (Garreaud fluence, which was expressed as the appearance and/or increase of
et al., 2009). As such, this area represents a latitudinal gradient the xerophytic vegetation south of its original area of occurrence
of precipitation increase (~80e740 mm per year) and the (Heusser, 1990; Villagra n and Varela, 1990; Maldonado and
average annual temperature decrease (~16 e14 C) (Romero, Villagran, 2006). In addition, a drier pulse during the arid phase
1985; Luebert and Pliscoff, 2006). Climate variability at inter- was recorded between 7800 and 5700 cal BP in the coastal Palo
annual and multi-decadal scales is controlled by El Nin ~o Colorado swamp forest gully located at 32 S (Maldonado and
Southern Oscillation (ENSO), with relatively wet years during El Villagran, 2006). This observation coincides with the inferences
Nin~ o events and the opposite trend during La Nin ~ a years based on the 18O alkenone record of marine sediments at 32 45' S
(Rutllant and Fuenzalida, 1991; Garreaud et al., 2009). Differ- showing the occurrence of maximum sea surface temperatures
ences in rainfall regimes generate Asteraceae-dominated between 7500 and 6000 cal BP and a subsequent decrease toward
shrubland in the north of the area, whereas the sclerophyll modern conditions since that time (Kim et al., 2002). Furthermore,
forests dominate the south (Luebert and Pliscoff, 2006). Faunal this period is broadly coincidental to the minimum activity of ENSO
communities include mostly rodents and other small-sized as expressed by offshore sediment records at 13 30' S (Rein et al.,
mammals. However, the coastal band provides a year-round 2005).
assemblage including fish, invertebrates, seabirds, and edible In the eastern Andes, records framed within the period of in-
plant resources (Quintanilla, 1983). terest are scarce, and they include pollen profiles and geomor-
b) The Andean ecosystem extends all along the area and is char- phological analyses of glacial advances (Za rate, 2002). These
acterized by relatively wetter environment, though strongly records suggest glacier re-advances between 6000 and 6500 cal BP
conditioned by the altitude. Winter precipitation controlled by due to increased winter precipitation at high altitudes. The only
the Pacific anticyclone dominates, with ranges between 300 and extra-Andean record is that of the pollen profile of the Gruta del
800 mm per year, the latter at the highest parts (Capitanelli, Indio site. This record indicates an increase in annual plants, as
1972, 2005; Abraham et al., 2009). The flora belongs mostly to would be expected with a decrease in summer precipitation or an
the Altoandina ecosystem, especially up to 2500 masl, with xeric increase in summer temperature (Markgraf, 1983). However, the
vegetation of the Patagonia ecosystem below this altitude and existence of discordant stratigraphic data and a temporal hiatus in
Punen ~ a ecosystem in the north (Cabrera, 1976; Roig et al., 2000; this profile weakens the validity of this inference for the entire
Luebert and Pliscoff, 2006; Martínez Carretero, 2006; Abraham period between 8850 and 4380 cal BP.
et al., 2009). Most of this environment is highly productive with The coastal swamp forests at ~32 S are sensitive to regional
regards to large mammals (dominated by Lama guanicoe); but winter precipitation (Maldonado et al., 2010; Maldonado and
less productive in seeds, roots, fruits, and other human edible Moreiras, 2013). These records indicate a gradual increase of hu-
plant resources; with productivity higher in the south than in midity starting at 5700 cal BP, with a significant rise between 4200
the north. and 3800 cal BP (Maldonado and Villagra n, 2002, 2006). Other
c) Semi-desert foothills and plains mostly dominated by the fossil pollen records suggest similar precipitation conditions in the
Atlantic anticyclone extend to the east of the Andes. Precipita- study area (Jenny et al., 2002), which are further in agreement with
tion ranges between 250 and 500 mm per year and allows the sedimentary changes observed in marine cores near at 32 45' S
existence of the Patagonia ecosystem in the south and west and since 4000 cal BP (Lamy et al., 1999). This information is consistent
the Monte ecosystem in the north and east (Cabrera, 1976; Roig with the supposition of glacial advances in the east Andes and the
et al., 2000; Abraham et al., 2009). Plant productivity is higher pollen records of locations at ~35 S (Navarro et al., 2012), which
with regards to human edible plants than in the Andes, but at suggest that the current climate patterns were established between
the same time, faunal productivity decreases, especially in terms 4500 and 4000 cal BP, in agreement with strong ENSO activity (Rein
of big game. et al., 2005).
Other minor traits have impacts on local scale environmental 4. Material and methods
conditions, such as the coastal mountain range in central Chile,
precordillera in northern eastern slope, the volcanic field of La All of the available 14C ages for the study area between 9000 and
Payunia in the southern eastern slope, and the general decrease in 3000 BP (~10000e3100 cal BP) were included in the database.
the altitude of the Andes from north to south. However, there is: (i) Uncalibrated radiocarbon dates were considered as minimum
a west-east precipitation gradient, (ii) a north-south temperature analytical units based on the archaeological site of origin. These
gradient, and (iii) an altitude Andean gradient. Combined, these were paired with geographical and ecological information, labo-
have a strong impact over the plant and faunal distributions, their ratory codes, standard errors, d13C, dated material, dating methods,
productivity, seasonality, and complementarity. primary references, and a basic type of site definition. The current
18 C. Mendez et al. / Quaternary International 356 (2015) 15e26
Fig. 2. Summary of major paleoclimatic records for the Subtropical Andes and adjacent areas, including a summed probability plot for all occupational events discussed in this
study. The major dry pulse (7800e5700 cal BP) is highlighted.
analysis excluded incomplete 14C data, archaeological sites with 5. Assessment of potential biases
unreliable contextual information, and dates with sigma 200
(Pettitt et al., 2003). Although the database included thermolumi- Several potential biases may greatly affect the temporal trends
nescence dates (obtained from rocks exposed to fire), these data (e.g. Williams, 2012), thus caution is required while analyzing
were excluded because this method has low reliability for ages them. One of the most significant biases in the study area is the
>3000 BP (Dincauze, 2000). Although the database considered the “aging” of the 14C signatures due to the use of samples that led to
ages on mollusk shells, analyses isolated these measurements potential reservoir effect (i.e. shell samples). Their exclusion im-
because there is no available correction for reservoir effect in this proves the precision of the analyses but countered the information
region. provided by data gathered at various archaeological sites, which
To critically evaluate our research question, we minimized the might significantly contribute to the understanding of the temporal
data biases that may arise from considering large amount of dates trends. Only when we are able to properly measure this error will it
from rich-dated archaeological sites, condition that might artifi- be possible to arrive at more substantial conclusions.
cially enhance the 14C signature of one or more specific chrono- Taphonomic bias is among the main factors influencing the re-
logical episodes and thereby introduce an over-represented human sults of 14C time series analysis because the strength of a later
signature. To control for this sample size effect, we defined “occu- archaeological signature, less affected by destructive processes,
pational events” at the minimum spatial unit of the archaeological may mimic population growth (Surovell and Brantingham, 2007).
site (based on the literature). These events were calculated by However, as suggested by Williams (2012), taphonomic correction
averaging the dates at sites with multiple 14C measurements. should not be routinely applied without first considering if time-
Specifically, dates were sorted in ascending order and two or more dependent taphonomic loss is a valid assumption. In this article,
dates were averaged when their differences were not significant we include a region sufficiently diverse enough to include several
(a ¼ 0.05) according to a T-test (Ward and Wilson, 1978). The contrasting landforms which are not expected to be constrained by
number of occupational events that resulted from these operations same time-dependent taphonomic losses and therefore minimizing
must be considered a “minimum number of events” (Me ndez, the possibility of unique trends affecting the disappearance of
2013). We decided not to average the dates on mollusk shells datable material and/or sites containing such. Additionally, we
with those on other types of material due to the potential biases explicitly do not address population growth as a research question,
introduced by a reservoir effect. but rather we focus on the presence/paucity of the radiocarbon
The analysis consisted of plotting frequency curves as well as signal in order to assess trends in the temporal distribution of the
quantifying the number of dates and occupational events in 500 archaeological record.
calibrated-year intervals. In addition, the summed probability The most significant research bias was the difference between
distributions were obtained using the grouped and ungrouped the greater frequencies of Chilean sites included in the database,
data according to the different ecological bands. The use of versus the greater number of dates per site obtained in Argentina
summed probability distributions is an accepted method for the (Table 1, Appendices A, B, C). This is chiefly related to a greater
analysis of 14C data series obtained at the regional level proportion of researchers involved in the production of radio-
(Williams, 2012). All calibrations were performed with Calib 7.0.0 carbon dates in the former region, most probably due to the
program (Stuiver et al., 2013) with the ShCal13 curve (Hogg et al., proximity of home institutions to the study area. On the other hand,
2013). archaeological data from western Argentina comes from more
C. Mendez et al. / Quaternary International 356 (2015) 15e26 19
intensely studied sites. However, comparing data from the two sample (42.93%) was obtained from sites yielding two to five 14C
countries, although the number of dates is different, the quantity of dates, although the percentage of sites with one single 14C age re-
averaged events is slightly more similar. This result leads us to mains high (17.93%). In addition, this result emphasizes the fact that
suppose that the chronological behavior of both slopes is more the majority of sites with more 14C ages within the interval under
comparable; and that dates, especially when averaged into occu- study are located in Argentina.
pational events, are defendable indirect demographic The use of charred organic material (primarily charcoal; 65%)
measurements. dominated (Table 2), thereby providing reliability to the dataset.
Mollusk shells (15.76%) are a frequent choice in sites near the Pacific
because they are predominantly found as byproducts derived from
Table 1
14
shellfish processing. In turn, charcoal in these locations is infre-
Regional distribution of C dates in the Mid-Holocene database (includes only
accepted 14C dates).
quent due to strong winds. The current inexistence of reservoir
effect determinations in this area limits the possibility of defining
14
Slopes Administrative regions C age frequency
the offsets on radiocarbon dated shells and caution against their
East of the Andes San Juan 26 14.13% use for regional comparisons. Other organic materials, such as
(Argentina) Mendoza 56 30.43% bones, are predominantly related to the dates on bioarchaeological
Subtotal 82 44.57%
remains and are mostly represented in the sample obtained west of
West of the Andes Coquimbo 56 30.43%
(Chile) Valparaíso 13 7.07% the Andes (11.41%).
Metropolitana 21 11.41%
O'Higgins 12 6.56%
Subtotal 102 55.74%
Table 2
Total 184 100.00% 14
Dated materials in the Mid-Holocene database (includes only accepted C dates).
Another research bias is the uneven spatial coverage of the East of the Andes Bone 5 2.72%
archaeological sample. This remains a limitation based on the his- Charred organic material 74 40.22%
Dung 1 0.54%
tory of investigation and it is well illustrated by data on the Pacific Human tendon 1 0.54%
coast. Here, the chronological signature is biased in favor of the Los Wood 1 0.54%
Vilos area which includes 62.3% (N ¼ 36 ages) of the coastal dates. Subtotal 82 44.57%
In other words, a linear 32-km strip representing 4.77% of the West of the Andes Bone 21 11.41%
Charred organic material 46 25.00%
670 km between 29 and 35 S comprises more than half of the
Human tooth 4 2.17%
coastal sample. This minimal proportion can hardly represent the Pigment (organic material) 1 0.54%
archaeological 14C trends of the entire environmental band. This Sediments 1 0.54%
situation may be considered analogous to that in the Argentinean Shell 29 15.76%
segment of the study area where four rich-dated sites (El Desecho, Subtotal 102 55.43%
Total 184 100.00%
Arq 18, Arroyo Malo 3, and Gruta del Indio) account for 57.32% of
the dates. These sites are located on the north and south extremes
of the study area and do not represent a random geographical
distribution. A review of dating methodologies suggests that conventional
14
In light of these problems, the behavior of the 14C signatures C has been given priority (49.74%, Table 3). Accelerator mass
must be understood as preliminary and in need of further testing spectrometry (AMS) 14C dating is mainly restricted to recent
with additional lines of evidence. The analysis of the 14C data series studies, and as such, it has a weaker representation throughout the
should be viewed as a way to represent datasets and compare them, dataset (31.6%). However, the frequencies of the use of different
but not as a conclusive test. However, once the potential biases of methods are similar on both slopes of the Andes. Thermolumi-
the overrepresentation of the 14C signatures and the taphonomic nescence dates are restricted to Chile and correspond to experi-
constraints for most dated sites have been addressed, it is possible mental analyses on rocks exposed to high temperatures (Roma n
to suggest general trends regarding the chronological behavior of and Jackson, 1998). They were excluded based on their high
the studied area. sigma values.
Table 3
6. Results Dating techniques in the Mid-Holocene database (includes all dates).
the primary source of chronological information in the eastern geographical units (Fig. 5 and Table 5). In order to discuss temporal
slope (Table 4). A chi-square test indicates that the samples in the trends with more precision, dates on shells were treated inde-
eastern and western slopes are statistically different across the type pendently. Although this only affects the area west of the Andes, its
of site (p < 0.0001). effect is worth noting because the dates in mollusk shells comprise
41% of the average human occupational events for this area.
Table 4
Date frequencies partitioned according to general type of site in the studied sample
(includes only accepted 14C dates).
Table 5
Slopes Type of site Date frequencies Frequency of occupational events at each geographical unit considering dated
materials.
East of the Andes Open-air 16 8.70%
Cave/rockshelter 66 35.87% Intervals in cal BP Coast and western Andes Foothills and Total
Subtotal 82 42.57% valleys mountains eastern plains
West of the Andes Open-air 85 41.30%
Cave/rockshelter 26 14.13% Various Shell Various Various
Subtotal 111 55.43% materials materials materials
Total 184 100.00% 3000e3500 6 2 5 1 14
3501e4000 4 1 5 1 11
14 4001e4500 3 2 6 3 14
6.2. Spatial trends in the C signatures of the mid-Holocene
4501e5000 3 0 8 1 12
5001e5500 4 1 5 2 12
Dates are unevenly distributed across the environments 5501e6000 5 3 5 1 14
comprised in the study area (Fig. 3). Two major environmental 6001e6500 2 5 8 15
bands dominate the frequencies of dates and events: the Pacific 6501e7000 3 1 3 1 8
7001e7500 0 1 3 1 5
coast and the eastern part of the Andes Mountains. The coast and
7501e8000 1 3 7 11
valleys western slope represent a geographical unit because major 8001e8500 1 0 2 3 6
barriers restricting movement do not exist. The distribution of the 8501e9000 3 2 1 1 7
sites along the coastal axis is uneven and strongly biased in favor of 9001e9500 1 3 5 3 12
Los Vilos area (31500 S; e.g., Jackson et al., 2004; Me ndez and 9501e10000 1 2 6 2 11
>10000 0 0 0 1 1
Jackson, 2004, 2006). In this area, the Punta Chungo sites (LV.
Total 37 26 69 21 153
046a and LV. 046b) stand out, with eight 14C dates that equally
represent eight human occupational events (Jackson and Me ndez,
2005), as well as the Los Rieles site (Fig. 4) with seven 14C dates
that represent six occupational events (Jackson et al., 2012). First, 14C data preceding 10,000 cal BP for all three geographical
Although the larger portion of the sites during this interval is units is available. The information corresponding to the first 14C
located along the coast, valleys yield important 14C datasets, such as signature in these areas has been previously systematized (Neme
the Cuchipuy site, with six 14C dates which represent four occu- and Gil, 2008, 2009; Me ndez, 2013). We chose this maximum
pational events (Kaltwasser et al., 1983). date limit to frame the beginning of the maximum Holocene aridity
The Andes corresponds to the major orographic unit in the study episode in order to generate an artificial starting point for our
area. Mountain settlements pose particular challenges that make the frequencies. Likewise, the minimum limit (~3100 cal BP) of the
humans who inhabited them seem more represented at cave/rock- sample also introduced an artificial term and broadly corresponds
shelter sites. In these naturally confined areas, site redundancy to the last presence of hunteregatherer adaptations throughout the
prevails; thus there are more prone to yield datable events. High study area.
dating frequency increases the probability of defining averaged For the interval between ~10,000 and 9000 cal BP, seven occu-
occupational events (see Fig. 3, eastern Andes). Although the num- pational events were observed in the coast and western valleys,
ber of dates is greater in Argentina (N ¼ 57) than in Chile (N ¼ 33), the eleven in the Andes, and six in the eastern foothills and plains. If we
difference between the minimum number of occupational events at exclude the dates on shells, then the number of events in the first
both slopes of the Andes is less marked (39 and 30 events, respec- environmental band is reduced to two. The recorded values in each
tively); thus, they are fairly comparable with regard to their chro- unit define the different trends for the interval between 9000 and
nological signatures. The larger portion of the 14C ages and events 8000 cal BP. While the numbers of occupational events at the
from the western slope originate from the sites in the Cajo n del western and eastern slopes are relatively maintained, a marked
Maipo (e.g., Cornejo et al., 1998; Cornejo and Sanhueza, 2003, 2011). decrease in frequencies is observed for the Andes Mountains
The archaeological sites are more infrequent on the eastern slope, because only three occupational events were defined.
but the chronological precision level is higher. The ARQ18 site in San The millennium between 8000 and 7000 cal BP comprises the
Juan stands out (Fig. 4), with thirteen 14C ages, representing nine beginning of the arid regional episode (7800e5700 cal BP). The
human occupational events (Cortegoso et al., 2012; Cortegoso, coast and valleys west of the Andes show a marked lack of 14C
2013), as does the Arroyo Malo 3 and El Desecho sites in Mendoza, signature because only one event was recorded when the shell
the former yielding eleven 14C ages and seven occupational events dates (which defined four events) are excluded. This decrease is
(Dieguez and Neme, 2003) and the latter with fifteen 14C ages and also visible at the eastern foothills and plains because only one
seven occupational events (Gil et al., 2008). occupational event was counted. On the other hand, the Andes
Eastern foothills and plains comprise the third geographical unit Mountains yield ten occupational events, thereby experiencing a
and yield the lowest 14C age frequencies in the studied sample, thus notable rise during this same period. During the following mil-
representing only a minor proportion of the occupational events. lennium (7000e6000 cal BP) when arid conditions prevailed, the
mountain environments continue to exhibit high 14C signatures
6.3. Chronological frequencies during the mid-Holocene (eleven occupational events). In the areas to the east of the Andes
only one occupation event was recorded, while to the coast and
The occupational events, defined by averaged 14C dates, allow us western valleys five events (excluding shell dates) suggest an
to describe the temporal trends across the three defined increase.
C. Mendez et al. / Quaternary International 356 (2015) 15e26 21
Between 6000 and 5000 cal BP, occupational events show this interval shows peaks and troughs enhanced in the larger
relative increases across two of the three geographical units and are dataset. Thus, we opted to exclude the dates using shells due to
relatively maintained in the Andes Mountains. Nine events were their potential inaccuracy.
recorded at the coast and valleys (after excluding shell dates), ten Fig. 7 shows a comprehensive selection of the most reliable
events at the Andes, and three events at the eastern foothills and occupational events through summed probability plots divided by
plains. This millennium shows a gradual return towards more hu- geographical units underscoring their potential complementarity.
mid environmental conditions. The events in the Andes and the The curve for the Pacific coast and western valleys depicts a variable
14
eastern units continued to increase (fourteen and four events, C signature prior to 5900 cal BP, with periods of lowest 14C sig-
respectively) between 5000 and 4000 cal BP, whereas the coast natures or troughs between ~7850e7300 and 6650-6300 cal BP.
showed a slight decrease with six non-shell averaged events. Although still variable after 5900 cal BP, the signature was relatively
The millennium between 4000 and 3000 cal BP revealed the continuous.
marked beginning of humid conditions and precipitation regimes By separating the samples of both slopes of the Andes, differ-
similar those in the present. Occupational events increased in the ences stand out. These can be better observed when considering
western slope units (N ¼ 10, excluding shell dated events), and the troughs in 14C signature. For the western Andes, periods with
decreased in the Andes (N ¼ 10) and to the eastern units (N ¼ 2). lowest chronological signature were located at ~9100e8200,
The values of this last interval must be considered with caution 7550e7200, and 6600e6200 cal BP, the first one being the most
because the cutoff of our sample at ~3100 cal BP reduced this last pronounced and continuous in the whole data set. On the other
interval to 900 calibrated years. hand, the eastern Andes show an intermittent, yet continued 14C
signature, yielding only one significant trough at ~7000e6500 cal
6.4. Chronological trends based on summed probability distribution BP. Periods with strong 14C signature and even peaks in one slope
models fully coincide with segments with low or no 14C signature. This
alternation may be the reason for the resulting averaged
Summed probability distributions, both aggregated and divided morphology of the curve that incorporates all chronological data
by the geographical units, allow us to discuss methodological issues obtained in Andean sites, with only one relative decrease in the 14C
and the temporal behavior across the period under study. The signature between ~9100 and 8200 cal BP.
complete dataset results in an averaged curve which conceals the The foothills and plains to the east showed a curve with more
differential behavior of the archaeological record (peaks and intermittence, depicting two pronounced peaks (at ~8400e8100
troughs) observed when the sample is divided into geographical and 4500e4100 cal BP) and five periods of troughs. The first
units (Fig. 6). Equally, the averaged curve for the Andes Mountains three 14C signature decreases occurred before 8400 cal BP and are
disguises the variability acquired when partitioning the sample small. However, the more pronounced troughs occurred between
between both sides. Although the shell-based dates affected the 14C ~7700e7300 and 6550e6000 cal BP. This variability is due to the
age frequencies and events on the coast when using 500 and 1000- scarce 14C data that comprise the curve.
year intervals, the results are different when employing summed
probability distributions. Significant differences between the curve 7. Discussion
morphologies that included the shell samples and those that
excluded them were not observed, except for the interval between Considering the potential biases outlined earlier, results show
~7600 and 6300 cal BP. Except for one marked peak at ~7500 cal BP, that the inclusion/exclusion of shell dates does not change
14
Fig. 3. Distribution of mid Holocene C ages and occupational events at each environmental band.
22 C. Mendez et al. / Quaternary International 356 (2015) 15e26
Fig. 4. Selected sites and environments, A. ARQ-18 (eastern Andes), B. Los Rieles (Pacific coast), C. Agua de la Cueva (eastern Andean foothills), D. Techo Negro (western Andes).
significantly chronological patterns as expressed in the summed population growth or taphonomic bias (Surovell and Brantingham,
probability plots in this case study. Although a slight raise in the 2007). The main sources of bias derive from research decisions and
chronological signal may be observed in the later segment of this are related to the intensity of dating sites and the geographical
dataset, it does not represent an exponential growth either by coverage at the different environmental bands. In the future, more
intense dating programs on a site by site basis must be performed
in the western slope. Dating sites with a wider distribution should
be of chief importance in the eastern slope and in the Pacific coast.
Quantification of 14C ages using 500 and 1000-year intervals
restricted the identification of periods without chronological sig-
natures. This initial analysis revealed variations in the date fre-
quencies and differential behaviors at each geographical unit
during the millennia of maximum aridity (8000e6000 cal BP). We
observed a marked decrease of the 14C signatures in the coast and
western valleys and in the eastern foothills and plains when
considering occupational events, whereas a notable and continuous
increase during this period occurred in the Andes Mountains. The
areas to the west and east of the Andes differed in the frequency,
magnitude, and timing of these changes. This quantification
allowed us to identify the first trends to be contrasted with the
analysis of the summed probability distributions.
We pose a preliminary interpretation considering the analysis of
different geographical units using the chronological distribution
analysis and emphasizing temporal trends. From these curves, we
observed the intervals with troughs in 14C signatures and compared
them with the regional environmental trends. We assume that
these troughs indicate the absence of dated deposits and therefore
indirectly indicate a lack of anthropogenic information in time and
space. In the coast and western valleys, the period with the most
significant troughs was between ~7850 and 6300 cal BP. Other
previous decreases were isolated and non-continuous. The Andes
show a more sustained 14C signature with less pronounced in-
termittences, and a trough between ~9100 and 8200 cal BP. How-
ever, the most significant difference results when dividing the 14C
information across both slopes within the mountain range.
Although 14C signature absences were observed on one slope, the
other showed a concentration of dated occupational events,
Fig. 5. Descriptive graphs showing 14C age and occupational event frequencies at each thereby underscoring their complementarity. Finally, the most
environmental band. The major dry pulse (7800e5700 cal BP) is highlighted. sustained trough in the eastern foothills and plains occurred
C. Mendez et al. / Quaternary International 356 (2015) 15e26 23
Fig. 6. Summed probability plots of occupational events for each environmental band and subunits. The major dry pulse (7800e5700 cal BP) is highlighted.
between ~7700 and 6000 cal BP, despite other minor earlier de- climate phase. The decrease in the occupational events in the
creases in the chronological signature. Andes as a whole occurred before the signature decrease in
Two scenario interpretations are possible: the other two extra-Andean areas and predates the major dry
pulse.
(1) The chronological signatures of the Pacific coast and western (2) The chronological signatures of the Pacific coast and western
valleys and the eastern foothills and plains showed troughs valleys, the west Andes, and the eastern foothills and plains
that were in relative synchronicity and with the major dry showed 14C signature troughs that are in relative
Fig. 7. Summed probability plots of occupational events for each environmental band. The major dry pulse (7800e5700 cal BP) is highlighted. Solid blocks represent periods with
null 14C signature.
24 C. Mendez et al. / Quaternary International 356 (2015) 15e26
synchronicity (~7800e7300 and 6600e6400/6000 cal BP) devoid of radiocarbon signals. A program devoted to determining
and with the major dry climate pulse. However, a previous the reservoir effect is crucial in assessing if by including corrected
and sustained trough was observed in the west Andes. The shell-dated material from a significant amount of dated sites,
chronology of the occupational events across the three chronological trends will vary or remain. This work preliminarily
described geographical units differed from that observed at concludes that an effect existed in the reconfiguration of the spaces
the eastern Andes, which showed chronological increases occupied by humans during the mid-Holocene in the Andean re-
and decreases across during different episodes compared gion between 29 and 35 S. The drought conditions recognized
with the signatures of the other units, both in terms of peaks along the coast and valleys of the western slope between 7800 and
and troughs. The only period with a marked decrease in 14C 5700 cal BP may have had an effect on preferential use of spaces as
signature in the eastern Andes fully coincides with peaks in shown by 14C signature patterns. The Andes Mountains also may
all other units. have offered alternative and more stable resources, and thus they
were effectively occupied in a relatively inverse pattern with
In summary, the beginning and end of the phase with troughs in regards to the other environmental bands. However, the different
the 14C signature of the geographical units located to the west and 14
C signatures between the western and eastern Andes may be
east of the Andes are noteworthy for their synchronicity. In addi- equally interpreted to represent differing human presence on both
tion, the synchronicity between the decrease in human activity and slopes during the evaluated period.
the phase of maximum aridity during the mid-Holocene
(7800e5700 cal BP) emerges as a suggestive result. Finally, the Acknowledgments
chronological signature of the coast and western valleys and foot-
hills and eastern plains, being asynchronous with regard to dated Many research projects contributed to the construction of the
human activity in the Andes, is also significant. Specifically, a sig- database; in particular, the following funding agencies: FONDECYT
nificant trough was localized in the western Andes between ~9100 1140824, DI SOC 13/2 2007 Universidad de Chile, CONICET, ANPCYT
and 8200 cal BP, and the two millennia between 8000 and 6000 cal PICT 2012-1015, and UNCuyo. In addition, we thank Donald Jack-
BP showed intermittent and alternating chronological patterns son, Daniel Pavlovic, Andre s Troncoso, Blanca Tagle, and Luis Cor-
between both slopes. nejo who either provided unpublished data, or kindly responded
our queries. Comments provided by Daniel H. Sandweiss and one
8. Conclusions anonymous reviewer were of great help for improving the original
manuscript, as was the support provided by guest editor Luciano
Analysis of radiocarbon date series is a powerful quantitative Prates.
tool that can establish human presence at a regional level and
indicate the magnitude of its spatiotemporal signature. As such, it is
Appendix A. Supplementary data
an important (though not infallible) method of establishing indirect
demographic parameters once research and taphonomic biases are
Supplementary data related to this article can be found at http://
addressed. Because hunteregatherers populations were highly
dx.doi.org/10.1016/j.quaint.2014.06.059.
mobile societies, spatial reorganization is considered an alternative
to the negative environmental fluctuations that affect the avail-
References
ability of resources (i.e., Fagan, 1999). Spatial reorganization has
been previously observed with regard to Holocene environmental Abraham, E., Del Valle, H.F., Roig, F., Torres, L., Ares, J.O., Coronato, F., Godagnone, R.,
fluctuations in South America (Nún ~ ez et al., 2002; Barrientos and 2009. Overview of the geography of the Monte Desert biome (Argentina).
Perez, 2005; Araujo et al., 2005-6; Me ndez and Jackson, 2006; Journal of Arid Environments 73, 144e153.
Anderson, G., Sandweiss, D., Maasch, K.A. (Eds.), 2007. Climate Change and Cultural
Grosjean et al., 2007; Neme and Gil, 2009). In these cases, the ef-
Dynamics: a Global Perspective on Holocene Transitions. Academic Press, San
fects of the changes were interpreted as restrictions in the avail- Diego.
ability of resources and limitations with regard to inhabiting Araujo, A.G.M., Pilo , L.B., Neves, W.A., Atui, J.P.V., 2005-6. Human occupation and
paleoenvironments in South America: expanding the notion of an “Archaic
certain regions. However, it is true that the spatial organization of
Gap”. Revista do Museu de Arqueologia e Etnologia 15e16, 3e35.
human groups depends on the degree of knowledge of environ- Barrientos, G., Perez, S.I., 2005. Was there a population replacement during the Late
mental conditions. The onset of the Mid-Holocene (~9000 cal BP) Mid-Holocene in the southeastern Pampas of Argentina? Archaeological evi-
postdates by some four thousand years the initial occupations dence and paleoecological basis. Quaternary International 132, 95e105.
Bueno, L., Politis, G., Prates, L., Steele, J., 2013. A late Pleistocene/early Holocene
recorded in the region and ever since a chronological signal has archaeological 14C database for Central and South America: palae-
been recorded (Neme and Gil, 2008, 2009; Me ndez, 2013). As such, oenvironmental contexts and demographic interpretations. Quaternary Inter-
it is fair to consider that most of the study area was already national 301.
Cabrera, A., 1976. Regiones fitogeogra ficas Argentinas. Enciclopedia Argentina de
inhabited by the start of the driest phase. Agricultura y Jardinería, vol. II(1). Acme, Buenos Aires.
Positive relationships between adverse climate pulses and de- Capitanelli, R., 1972. Geomorfología y clima de la provincia de Mendoza. Boletín de
creases in 14C signatures should be evaluated as potential indicators la Sociedad Argentina de Bot anica 13, 15e48.
Capitanelli, R., 2005. Climatología de Mendoza. Editorial de la Facultad de Filosofía y
of reconfigurations in mobility patterns and use of space. However, Letras, Universidad Nacional de Cuyo, Mendoza.
collation between trends derived from different datasets should Cornejo, l., Saavedra, M., Vera, H., 1998. Periodificacio n del Arcaico en Chile Central:
not be necessarily understood as causation (Sandweiss and Quilter, una propuesta. Boletín de la Sociedad Chilena de Arqueología 25, 36e39.
Cornejo, L., Sanhueza, L., 2003. Coexistencia de cazadores recolectores y horti-
2012). Casual hypotheses should be evaluated and judged with cultores tempranos en la cordillera andina de Chile Central. Latin American
regard to their magnitude, either by increasing archaeological Antiquity 14 (4), 389e407.
sampling, 14C dating, including complimentary datasets, and/or Cornejo, L., Sanhueza, L., 2011. North and south: hunter-gatherer communities in the
Andes mountains in Central Chile. Latin American Antiquity 22 (4), 487e504.
expanding the spatial scale of analysis. For instance, the coastal
Cortegoso, V., 2013. Valle de las Taguas, ARQ-18: estratigrafía, secuencia temporal y
environments have been signaled as potential buffer zones in times ocupaciones humanas. In: Cortegoso, V., Dur an, V., Gasco, A. (Eds.), Arqueología
of resource shortages because they yield predictable resources and de ambientes de altura de Mendoza y San Juan (Argentina). EDIUNC, Uni-
are occasionally high in productivity (Waselkov, 1987; Erlandson versidad Nacional de Cuyo, Mendoza.
Cortegoso, V., Dura n, V., Castro, S., Winocur, D., 2012. Disponibilidad de recursos
and Fitzpatrick, 2006). However, our results show that when líticos y explotacion humana de la divisoria andina. Valle del río de Las Taguas,
excluding shell dates, at least during some phases, these areas were San Juan, Argentina. Chungara. Revista de Antropología Chilena 44 (1), 59e72.
C. Mendez et al. / Quaternary International 356 (2015) 15e26 25
Dieguez, S., Neme, G., 2003. Geochronology of the archaeological site Arroyo Malo 3 protegidas de la Provincia de Mendoza, Anales de Arqueología y Etnología,
and the first human occupations in the North Patagonia Early Holocene. In: Mendoza, pp. 217e235.
Miotti, L., Salemme, M., Flegenheimer, N. (Eds.), Ancient Evidence for Paleo Mendez, C., 2013. Terminal Pleistocene/Early Holocene 14C dates form archaeo-
South Americans: from where the South Winds Blows. Center for the Study of logical sites in Chile: discussing critical chronological issues for the initial
the First Americans, College Station, pp. 87e92. peopling of the region. Quaternary International 301, 60e73.
Dincauze, D., 2000. Environmental Archaeology. Principles and Practice. Cambridge Mendez, C., Jackson, D., 2004. Ocupaciones humanas del Holoceno tardío en Los
University Press, Cambridge. Vilos (IV Regio n, Chile): origen y características conductuales de la poblacio n
Erlandson, J., Fitzpatrick, S., 2006. Oceans, islands, and coasts: current perspectives local de cazadores recolectores de litoral. Chungara. Revista de Antropología
on the role of the sea in human prehistory. Journal of Island & Coastal Chilena 36 (2), 279e293.
Archaeology 1, 5e32. Mendez, C., Jackson, D., 2006. Causalidad o concurrencia, relaciones entre cambios
Fagan, B., 1999. Floods, Famines, and Emperors: El Nin ~ o and the Fate of Civilizations. ambientales y sociales en los cazadores recolectores durante la transicio n entre
Pimlico, London. el Holoceno medio y tardío (costa del Semi arido de Chile). Chungara. Revista de
Faught, M., 2008. Archaeological roots of human diversity in the New World: a Antropología Chilena 38 (2), 169e180.
compilation of accurate and precise radiocarbon ages from earliest sites. Navarro, D., Rojo, L.D., De Francesco, C.G., Hassan, G.S., 2012. Paleoecología y
American Antiquity 73 (4), 670e698. reconstrucciones paleoambientales en Mendoza durante el Holoceno. In:
Gajewski, K., Munoz, S., Peros, M., Viau, A., Morlan, R., Betts, M., 2011. The Canadian Neme, G., Gil, A. (Eds.), Paleoecología humana en el sur de Mendoza: per-
archaeological radiocarbon database (CARD): archaeological 14C dates in spectivas arqueolo gicas. Sociedad Argentina de Antropología, Buenos Aires,
North America and their paleoenvironmental context. Radiocarbon 53 (2), pp. 17e56.
371e394. Neme, G., Gil, A., 2008. Biogeografía humana en los Andes meridionales: tendencias
Gambier, M., 1993. Prehistoria de San Juan. Ansilta Editora, San Juan. arqueolo gicas en el sur de Mendoza. Chungara. Revista de Antropología Chilena
Gamble, C., Davies, W., Pettitt, P., Richards, M., 2004. Climate change and evolving 40 (1), 5e18.
human diversity in Europe during the last glacial. Philosophical Transactions of Neme, G., Gil, A., 2009. Human occupation and increasing Mid-Holocene aridity.
the Royal Society B 359, 243e254. Southern Andean perspectives. Current Anthropology 50 (1), 149e163.
Garreaud, R., Vuille, M., Compagnucci, R., Marengo, J., 2009. Present-day south Nún ~ ez, L., Grosjean, M., Cartajena, I., 2002. Human occupations and climate change
American climate. Palaeogeography, Palaeoclimatology, Palaeoecology 281, in the Puna de Atacama, Chile. Science 298, 821e824.
180e195. Pettitt, P., Davies, W., Gamble, C., Richards, y M., 2003. Palaeolithic radiocarbon
Gil, A., Guerci, A., Neme, G., 2008. Human occupation, environmental structure & chronology: quantifying our confidence beyond two half-lives. Journal of
the archaeofaunal record during the mid-Holocene in West-Central Argentina. Archaeological Science 30, 1685e1693.
Before Farming 3, 67e80. Quintanilla, V., 1983. Biogeografía. Geografía de Chile. Instituto Geogra fico Militar,
Grosjean, M., Santoro, C.M., Thompson, L.G., Nún ~ ez, L., Standen, V., 2007. Mid-Ho- Santiago.
locene climate and cultural change in the South-Central Andes. In: Rein, B., Lückge, A., Reinhardt, L., Sirocko, F., Wolf, A., Dullo, W.C., 2005. El Nin ~o
Anderson, G., Sandweiss, D., Maasch, K.A. (Eds.), Climate Change and Cultural variability off Peru during the last 20,000 years. Paleoceanography 20, PA4003.
Dynamics: a Global Perspective on Holocene Transitions. Academic Press, San http://dx.doi.org/10.1029/2004PA001099.
Diego, pp. 51e115. Rick, J.W., 1987. Dates as Data: an examination of the Peruvian Preceramic radio-
Heusser, C., 1990. Ice age vegetation and climate of subtropical Chile. Palae- carbon record. American Antiquity 52, 55e73.
ogeography, Palaeoclimatology, Palaeoecology 80, 107e127. Roig, F.A., Martínez Carretero, E., Me ndez, E., 2000. Mapa de vegetacio n de la Pro-
Hogg, A., Hua, Q., Blackwell, P.G., Niu, M., Buck, C.E., Guilderson, T.P., Heaton, T.J., vincia de Mendoza. In: Abraham, E.M., Rodríguez Martínez, F. (Eds.), Argentina
Palmer, J.G., Reimer, P.J., Reimer, R.W., Turney, C.S.M., Zimmerman, S.R.H., 2013. Recursos y problemas ambientales de la zona a rida. Primera parte, Tomo II.
SHCAL13 southern Hemisphere calibration, 0e50,000 years CAL BP. Radio- PAN-IADIZA, Universidad de Granada, Mendoza, Atlas b asico.
carbon 55 (4), 1889e1903. Roma n, A., Jackson, D., 1998. Dataciones por termoluminiscencia de rocas de
Jackson, D., 2002. Cazadores y recolectores del Holoceno Medio del Norte Semia rido fogones de asentamientos arcaicos, Provincia del Choapa. Chungara 30 (1),
de Chile. Unpublished master's thesis. Universidad de Chile, Santiago. 41e48.
Jackson, D., Ba ez, P., Arata, J., 2004. Composicio n de conchales, estrategia de sub- Romero, H., 1985. Geografía de los climas. Geografía de Chile. Instituto Geogra fico
sistencia y cambios paleoambientales en un asentamiento Arcaico, Norte Chico Militar, Santiago.
de Chile. Boletín Sociedad Chilena de Arqueología 37, 37e48. Rutllant, J., Fuenzalida, H., 1991. Synoptic aspects of the Central Chile rainfall vari-
Jackson, D., Me ndez, C., 2005. Reocupando el espacio: historia de un asentamiento ability associated with the Southern Oscillation. International Journal of
multicomponente, sus relaciones inter-sitios y los cambios paleoambientales de Climatology 11, 63e76.
la costa del Choapa. Werken 6, 97e112. Sandweiss, D., Quilter, J., 2012. Collation, correlation, and causation in the prehis-
Jackson, D., Me ndez, C., Aspillaga, E., 2012. Human remains directly dated to the tory of coastal Peru. In: Cooper, J., Sheets, P. (Eds.), Surviving Sudden Environ-
PleistoceneeHolocene transition support a maritime diet among the first set- mental Change. Understanding Hazards, Mitigating Impacts, Avoiding Disasters.
tlers of the Pacific coast of South America. Journal of Island and Coastal University Press of Colorado, Boulder, pp. 117e139.
Archaeology 7, 363e377. Steele, J., Politis, G., 2009. AMS 14C dating of early human occupation in southern
Jenny, B., Valero-Garce s, B., Villa-Martínez, R., Urrutia, R., Geyh, M., Viet, H., 2002. South America. Journal of Archaeological Science 36, 419e439.
Early to mid-Holocene aridity in Central Chile and the southern westerlies: the Stuiver, M., Reimer, P.J., Reimer, R., 2013. CALIB 7.0.0. Www Program and Docu-
Laguna Aculeo record (34 S). Quaternary Research 58, 160e170. mentation. http://calib.qub.ac.uk/calib/.
Kaltwasser, J., Medina, A., Munizaga, J., 1983. Estudio de once fechas de RC14, Surovell, T., Brantingham, P.J., 2007. A note on the use of temporal frequency dis-
relacionadas con el hombre de Cuchipuy. Boletín de Prehistoria de Chile 9, tributions in studies of prehistoric demography. Journal of Archaeological Sci-
9e14. ence 34, 1868e1877.
Kim, J., Schneider, R., Hebbeln, D., Muller, P., Wefer, G., 2002. Last deglacial sea- Valero-Garce s, B., Jenny, B., Rondanelli, M., Delgado-Huertas, A., Burns, S., Veit, H.,
surface temperature evolution in the Southeast Pacific compared to climate Moreno, A., 2005. Palaeohydrology of Laguna de Tagua Tagua (34 300 S) and
changes on the South American continent. Quaternary Science Reviews 21, moisture fluctuations in Central Chile for the last 46 000 yr. Journal of Qua-
2085e2097. ternary Science 20 (7e8), 625e641.
Lamy, F., Hebbeln, D., Wefer, G., 1999. High-resolution marine record of climatic Veit, H., 1996. Southern Westerlies during the Holocene deduced from geomor-
change in mid-latitude Chile during the last 28,000 years based on terrigenous phological and pedological studies in the Norte Chico, Northern Chile
sediment parameters. Quaternary Research 51, 83e93. (27e33 S). Palaeogeography, Palaeoclimatology, Palaeoecology 123, 107e119.
Luebert, F., Pliscoff, P., 2006. Sinopsis bioclima tica y vegetacional de Chile. Editorial Villagran, C., Varela, J., 1990. Palynological evidence for increased aridity on the
Universitaria, Santiago. Central Chilean Coast during the Holocene. Quaternary Research 34, 198e207.
Maldonado, A., Me ndez, C., Ugalde, P., Jackson, D., Seguel, R., Latorre, C., 2010. Early Villa-Martínez, R., Villagr an, C., 1997. Historia vegetacional de bosques pantanosos
Holocene climate change and human occupation along the semiarid coast of de la costa de Chile central durante el Holoceno medio y tardío. Revista Chilena
north-central Chile. Journal of Quaternary Science 25 (6), 985e988. de Historia Natural 70, 391e401.
Maldonado, A., Moreiras, S.M., 2013. Comment on: major hydrological regime Villa-Martínez, R., Villagra n, C., Jenny, B., 2003. The last 7500 cal yr B.P. of westerly
change along the semiarid western coast of South America during the early rainfall in Central Chile inferred from a high-resolution pollen record from
Holocene by Ortega et al. Quaternary Research 78 (3), 513e527. Quaternary Laguna Aculeo (34 S). Quaternary Research 60, 284e293.
Research 80, 138e139. Ward, G.K., Wilson, S.R., 1978. Procedures for combining radiocarbon age de-
Maldonado, A., Villagr an, C., 2002. Paleoenvironmental changes in the semiarid terminations: a critique. Archaeometry 20, 19e31.
coast of Chile (~32 S) during the last 6200 cal years inferred from a swamp- Waselkov, G., 1987. Shellfish gathering and shell midden archaeology. In:
forest pollen record. Quaternary Research 58, 130e138. Schiffer, M.B. (Ed.), Advances in Archaeological Method and Theory 10. Aca-
Maldonado, A., Villagr an, C., 2006. Climate variability over the last 9900 cal yr BP demic Press, Orlando, pp. 257e310.
from a swamp forest pollen record along the semiarid coast of Chile. Quater- Williams, A., 2012. The use of summed radiocarbon probability distributions in
nary Research 66, 246e258. archaeology: a review of methods. Journal of Archaeological Science 39 (3),
Markgraf, V., 1983. Late and Postglacial vegetational and paleoclimatic changes in 578e589.
Subantarctic, Temperate, and Arid environments in Argentina. Palynology 7, Williams, A., Santoro, C., Smith, M., Latorre, C., 2008. The impact of ENSO in the
43e70. Atacama desert and Australian arid zone: exploratory time-series analysis of
Martínez Carretero, E., 2006. Reserva natural y provincia fitogeogr afica. In: archaeological records. Chungara, Revista de Antropología Chilena 40,
Duran, V., Cortegoso, V. (Eds.), Arqueología y ambiente de a reas naturales 245e259.
26 C. Mendez et al. / Quaternary International 356 (2015) 15e26
Williams, A., Ulm, S., Goodwin, I.D., Smith, M., 2010. Hunter-gatherer response to rate, M., 2002. Los ambientes del Tardiglacial y Holoceno en Mendoza. In: Gil, A.,
Za
late Holocene climatic variability in northern and central Australia. Journal of Neme, G. (Eds.), Entre montan ~ as y desiertos: arqueología del sur de Mendoza.
Quaternary Science 25 (6), 831e838. Sociedad Argentina de Antropología, Buenos Aires, pp. 103e118.
Yacobaccio, H.D., 2013. Towards a human ecology for the Middle Holocene in the rate, M., Neme, G., Gil, A., 2005. Mid-Holocene paleoenvironments and human
Za
Southern Puna. Quaternary International 307, 24e30. occupation in southern South America. Quaternary International 132, 1e3.