Journal of Archaeological Science 39 (2012) 2272e2279
Contents lists available at SciVerse ScienceDirect
Journal of Archaeological Science
journal homepage: http://www.elsevier.com/locate/jas
Water availability and landuse during the Upper and Epipaleolithic
in southwestern Syria
Knut Bretzke a, *, Philipp Drechsler a, Nicholas J. Conard a, b
a
b
Eberhard Karls Universität Tübingen, Institut für Ur- und Frühgeschichte, Abt. Ältere Urgeschichte und Quartärökologie, Burgsteige 11, 72070 Tübingen, Germany
Senckenberg Center for Human Evolution and Paleoecology, Universität Tübingen, Schloss Hohentübingen, 72070 Tübingen, Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 31 October 2011
Received in revised form
27 February 2012
Accepted 28 February 2012
Paleolithic research often assumes that environmental conditions played a major role in shaping human
evolution. To study this relationship we present a spatially explicit approach based on the assumption
that the distribution of water within the landscape is an essential component of local environmental
conditions. Here, we analyze the relation of wetness and human landuse patterns from the Upper
Paleolithic (UP) and Epipaleolithic (EP) of Western Syria. In particular the spatially explicit character of
the approach enables the detection of a significant change in landuse patterns during the UP and EP
accompanied by a significant shift in the wetness characteristics of the preferentially used areas. These
results are discussed against the background of published data on climatic conditions in order to identify
both a possible time frame and triggers for this change in landuse. While we conclude an increased
influence of natural conditions on the spatial behavior for the UP, we suggest an additional influence of
cultural circumstances in shaping EP spatial behavior. For the region studied we argue that the bounded
pattern observed during the UP changes to a spatially flexible pattern during the Late Natufian.
Ó 2012 Elsevier Ltd. All rights reserved.
Keywords:
Spatial analysis
Human landuse
TWI
PAW
Paleolithic
1. Introduction
Paleolithic research often places a high level of importance on
the environment’s role in determining Paleolithic populations’
evolutionary and cultural development (Banks et al., 2008; Binford,
2001; Collard and Foley, 2002; Foley, 1995). A variety of studies has
been published using a broad range of approaches. The most
common approach analyzes archaeological site distributions with
respect to environmental variability (Bretzke, 2008; d’Errico and
Sanchez-Goni, 2003; Gamble et al., 2004). In addition, archaeologists increasingly realize the potential of GIS based models for
studying the relationship between humans and their inhabited
environment (Banks et al., 2008, 2006; Burke et al., 2008).
Regardless of the approach chosen, one shortcoming is the
disagreement between the spatial resolution of the archaeological
and climatic data. While archaeological finds can be located
precisely using GPS or high resolution maps, paleo-environmental
information taken from climate proxies or climatic simulations
usually considers average developments over large areas and
therefore yields low spatial resolution. Hence, available climatic
data often underestimates local variation, a factor which is particularly important for the distribution of vegetation that forms the
basis for animal and human life. Especially in arid regions, the
vegetation pattern is closely bound to the availability of water.
Therefore, knowledge about the local distribution of water in
the landscape is a critical prerequisite for the assessment of
humaneenvironment interactions.
Here, we describe a method that transforms low resolution
precipitation data into high resolution surface-wetness distribution
data at a regional scale. We assume water availability as a critical
factor shaping the spatial distribution of plants and animals, and
therefore human biotic resources. Then we compare the wetness
distribution with landuse patterns from the Upper Paleolithic (UP)
and Epipaleolithic (EP) of Western Syria to characterize the relationship between human activities and the wetness pattern in the
landscape as an indicator of the environment’s role in the organization of human spatial behavior.
2. Materials and methods
2.1. Modeling the spatial distribution of wetness
* Corresponding author. Eberhard Karls Universität Tübingen, Department of
Early Prehistory and Quaternary Ecology, Burgsteige 11, 72070 Tübingen, Germany.
Tel.: þ49 69 75421549; fax: þ49 7071 295714.
E-mail address: [email protected] (K. Bretzke).
0305-4403/$ e see front matter Ó 2012 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jas.2012.02.033
The availability of surface and subsurface water at a specific
point in the landscape is primarily a function of the amount and
distribution of precipitation and local topography. Information
K. Bretzke et al. / Journal of Archaeological Science 39 (2012) 2272e2279
about the amount and distribution of precipitation can be either
obtained from isohyet maps or calculated from climatic station
data. The most complete information about local topography exists
as digital elevation models (DEMs). The spatially continuous
information of DEMs about topography can be used to determine
the amount of water that potentially flows together at any given
point in a landscape.
This concept is realized in the ‘Topographic Wetness Index’
(TWI), a well-established secondary topographic attribute introduced by Beven and Kirkby (1979). It assumes that topography
controls the movement of water in sloped terrain and hence, the
spatial pattern of soil moisture (Schmidt and Persson, 2003). As
a consequence, TWI also accounts for water available on the
surface. While high TWI corresponds to low lying areas, low TWI
values occur in steep, diverging areas.
The TWI applies likewise for subsurface water movements
driven by terrain gradient (Barling et al., 1994; Sørensen et al.,
2005). One of the characteristics of arid and semi-arid regions is
that most precipitation does not infiltrate the sedimentary column
where it falls, but drains toward lower areas, even where the
surface inclination is slight (Schulz, 2000). The greatest concentration of water can be found in the flat areas between hills where
sediments and water collect. With these characteristics, the TWI
can be used to differentiate potentially moister areas with higher
herbal biomass from drier areas with less vegetation. The major
drawback for the application of the TWI with respect to the
distribution of surface water is the assumption of a homogenous
spatial precipitation pattern over the entire region of analysis.
While this assumption is valid for some regions of the world, other
regions show variable precipitation patterns on a regional scale
dependent upon topographic features. To overcome this weakness,
we add the spatial variability of precipitation. We achieved this by
multiplying a separate data layer containing information about the
spatial distribution of precipitation. The resulting formula defines
Plant Available Water (PAW) in a specific region:
PAW ¼ TWIln PRECpresent
(1)
where PRECpresent represents the present precipitation. The
consideration of a spatially explicit precipitation pattern produces
a more realistic picture of the distribution of moisture in the
landscape than the TWI alone.
2.2. Modeling past human landuse
Since we see the most direct interaction of humans with their
environment in the context of gathering and collecting, we
consider human activities away from the main living sites. But how
can we gain information from this area? One possibility we have
explored over the past years (Bretzke, 2008; Bretzke et al., 2011;
Conard et al., 2010) necessitates a new way of viewing cultural
remains in the landscape. If we change our perspective from
identifying sites to the distribution patterns of single artifacts, we
access data on human spatial activities across the entire landscape
(Dunnell and Dancey, 1983). Our investigations follow an off-site
approach, as introduced by Isaac (1981) and Foley (1981), and
assume that the accumulation of cultural remains in the landscape
is a function of the intensity of its use. Therefore, areas with high
artifact densities denote intensively used areas, while low artifact
densities indicate locales exploited to a lesser degree. This leads us
to conclude that by discerning the spatial distribution of artifact
densities, we can posit inferences regarding past human behavior
in particular settlement spaces.
Information about the patterning of artifacts in the landscape
derives from archaeological survey data. The systematic collection
2273
of artifacts and the estimation of artifact densities specific of
cultures at all surveyed localities provide a dataset suitable for this
analysis. Applying the “Inverse Distance Weighted” interpolation
technique (Conolly and Lake, 2006), the density information of
artifacts gained from spatially separated localities can be transferred into a spatially continuous dataset that better represents the
activity distribution (Bretzke, 2008).
2.3. Case study
We apply this approach to archaeological data from the Tübingen-Damascus Excavation and Survey Project (TDASP in German)
(Conard, 2006). The TDASP study area in the Damascus Province
about 50 km northwest of Damascus (Syria) is situated between the
Anti-Lebanon Mountain range and the Syrian Desert. An impressive
cliff line formed of resistant Oligocene limestone represents the
border between areas of higher elevation, up to 2300 m, and
lowland areas down to 800 m (Fig. 1). The study area is further
characterized by a steep gradient in precipitation. While substantial
precipitation falls along the western faces of the Anti-Lebanon
Mountains, precipitation decreases markedly toward the eastern
flanks and the hinterland. Enzel et al. (2008) argue that this overall
pattern has not changed considerably since the last glacial/interglacial period. Thus, we can use the present precipitation pattern to
create a simplified model of local environmental conditions. The
dataset on the spatial distribution pattern of present precipitation
in the study area was created by a thin-plate spline interpolation of
isohyets from a Syrian precipitation map (Syrian Arab Republic,
1977). Finally, data on the topography derives from a conveniently processed DEM that was based on maps from the shuttle
radar topography mission (United States Geological Survey, 2007)
(Fig. 2).
The TDASP study region is a highly differentiated landscape with
characteristics that likely encouraged Paleolithic settlement in the
region. Lithic raw material of good quality is readily available from
primary and secondary deposits in many areas of the study region
(Dodonov et al., 2007). Another advantageous characteristic is the
occurrence of perennial water sources along the cliff line. Today
active artesian springs can be found near the base of the cliff line in
vicinity to the modern villages of Ma’aloula, Jaba’deen and Yabroud
(Fig. 1). These springs are associated with deeply incised canyons,
which not only confirm the existence of water sources at these
locations for long periods, but also provided easy passage for
people and animals between the lowlands and the highlands.
The field work of TDASP builds on previous work in the Damascus Province by Rust (1950), Suzuki and Kobori (1970), Solecki
and Solecki (1987/1988), and Bakdach (2000). Before TDASP
began working in the Damascus Province, the only stratified
Paleolithic assemblages known were those from the sites of Yabrud
(Rust, 1950; Solecki and Solecki, 1966). During 12 years of research
in the Damascus Province, TDASP conducted excavations in Baaz
Rockshelter, a multi-period site with layers from early and late UP
and Late Natufian (Conard, 2000; Conard et al., 2006a,b; Deckers
et al., 2009), Kaus Kozah Cave, a site containing Middle and Epipaleolithic artifacts (Conard and Masri, 2006), Ain Dabbour, a site
with two Geometric Kebaran layers (Hillgruber, 2010), and Wadi
Mushkuna, a Middle Paleolithic site currently under excavation
(Fig. 1).
Since 1999 TDASP has conducted intensive surveys in the
Damascus Province that represent a significant record of Paleolithic
data in Southwestern Asia (Conard, 2006). Today, the collection
contains chipped stone artifacts from 598 areas (Fig. 3). The record
includes all major periods from the Lower Paleolithic to the
Neolithic (Table 1). Within our survey area of ca. 500 km2, we
collected artifacts using pedestrian transects through defined areas
2274
K. Bretzke et al. / Journal of Archaeological Science 39 (2012) 2272e2279
Fig. 1. Geographic setting of the study area and location of sites mentioned in the text. 1) Baaz Rockshelter, 2) Kaus Kozah Cave, 3) Ain Dabbour, 4) Yabroud, 5) Wadi Mushkuna.
in the landscape. Over the course of 8 seasons of survey, we were
able to sample approximately 15% of the surface, including samples
from all major landscape forms and topographic settings. Our
decision about the location of areas to sample was guided by the
aim of covering the entire area of study. We use the term localities
to identify areas where artifacts were collected. At each locality, the
survey team documented the presence or absence of every cultural
entity and the respective artifact density (Fig. 4). One surveyed
locality may contain artifacts from different periods.
To use one case study, we limit the presented data to the UP
(ca. 40e20 ka) and the EP (ca. 20e10 ka) and consider 181 localities
containing UP artifacts and 97 containing EP artifacts (Fig. 4). In 45
localities EP and UP artifacts occur together. 410 localities lack
artifacts from both periods and were used as negative evidence. A
sub-division of the entities UP and EP into specific archaeological
cultures is usually not possible due to the low chronological resolution of the survey finds and the scarcity of chrono-culturally
sensitive artifacts at the majority of our survey localities.
However, a distinction between the more general entities UP and
EP is possible. Guided by our knowledge from the stratified
assemblages of Baaz Rockshelter, Kaus Kozah Cave and Ain Dabbour
we classified the surface finds into UP or EP using primarily technological characteristics of cores and blades/bladelets (Hillgruber,
2010). Edge damage and patina are additional characteristics used
to support the assessments based on technological criteria.
The integrity of the artifact pattern identified in the landscape is
confirmed by geomorphological studies that suggest stable land
surface conditions and only minor dislocation of artifacts in the
landscape (Dodonov et al., 2007). Furthermore, in 52 localities EP
finds are accompanied by Middle Paleolithic artifacts but lack UP
finds. This observation suggests that the absence of UP is not due to
taphonomy but rather reflects actual differences in spatial behavior
between EP and UP.
3. Results
We successfully transformed low resolution precipitation data
into a high resolution model of wetness distribution using (1) and
readily available topographic data (Fig. 2). The calculated PAW
indices represent wetness in the landscape dependent on the
degree of local precipitation and topography (Fig. 5). Very high PAW
values indicate a high amount of local wetness often in the flat parts
Fig. 2. Depiction of data used to model the distribution of wetness. A) Digital elevation model (DEM), derived from SRTM data, dark areas indicate high values, B) Topographic
Wetness Index (TWI), calculated from A after Beven and Kirkby (1979), color code see A, C) local precipitation pattern, derived from precipitation data of the Syrian Arab Republic
(Syrian Arab Republic, 1977).
K. Bretzke et al. / Journal of Archaeological Science 39 (2012) 2272e2279
Fig. 3. Distribution of surveyed localities. A total of 598 localities were sampled to
collect data on the spatial behavior of prehistoric populations. One locality may contain
cultural remains from different chrono-cultural entities.
of the landscape, but also in runoff channels. Areas with increased
amounts of water are represented by medium PAW values, whereas
dry areas have low PAW values and occur in areas with relatively
steep slopes. A comparison of similar terrains from the eastern
and western part of the study area indicates higher PAW values in
the west. This results from the regional precipitation pattern
(Fig. 2C) where higher values in the northwest decrease toward the
southeast.
To analyze the relation between human groups and the distribution of wetness, we used data from 598 surveyed localities to
model UP and EP human spatial behavior. The results show that
a spatially compact cluster of UP activity areas gives way to
a pattern characterized by a wide distribution of loosely connected
activity areas during the succeeding EP (Fig. 6). Hence the modeling
of human spatial behavior suggests a profound change in the way
the inhabited landscape was used. The reasons for such a change
may be manifold, but changes in environmental conditions may
have played a major role.
We test this hypothesis by analyzing the relationship between
landuse patterns and the distribution of wetness as represented by
Table 1
TDASP survey record. Summary of sites based on chrono-cultural entity and artifact
density. These sites are present at 575 localities documented since 1999. No artifacts
were found at 23 surveyed localities, which leads to the total number of 598 survey
localities. A locality may contain sites from more than one cultural group. * each of
these sites contained only one handaxe.
Lower Paleolithic
Handaxes
Middle Paleolithic
(Levallois)
Middle Paleolithic
(non-Levallois)
Upper Paleolithic
Upper or Epipaleolithic
Epipaleolithic
Post-Epipaleolithic
Indeterminate
Total
Low
density
Medium
density
High
density
Total number
of sites
30 (37%)
16* (94%)
143 (40%)
36 (44%)
0
160 (45%)
16 (20%)
1 (6%)
51 (14%)
82
17
354
9 (38%)
12 (50%)
3 (13%)
24
65 (36%)
7 (88%)
55 (57%)
11 (41%)
60 (79%)
396
92 (51%)
1 (13%)
33 (34%)
10 (37%)
14 (18%)
358
24 (13%)
0
9 (9%)
6 (22%)
2 (3%)
112
181
8
97
27
76
886
2275
the PAW index. We overlay the wetness distribution raster (Fig. 5)
on the UP and EP landuse patterns (Fig. 6) and deduce the wetness
characteristics of the activity areas for both periods.
The wetness distribution is represented by raster data consisting
of cells of 90 by 90 m. The EP landuse pattern covers activity areas
of 3020 cells of the wetness raster. These cells have a mean PAW
value of 36.6 with a standard deviation of 5.4. In contrast, the UP
landuse pattern covers 4585 cells of the wetness raster. Here the
mean PAW value is 38.8 with a standard deviation of 4.7. The range
of the samples is blurred by few extreme outliers (UP: n ¼ 8; EP:
n ¼ 1). To improve the analytical value of the PAW samples, we
eliminated these outliers. This results in an UP range of 31.5 and an
EP range of 35.5 for all remaining values, indicating an increase of
about 13%. Hence we observe an increase in the diversity of the
wetness characteristics of the activity areas. The differences
between both samples were tested using a ManneWhitney rank
sum test. According to the test results a statistically significant
difference (p < 0.01) exists between the UP and EP. This marks
a pronounced shift in the way humans used the landscape, from the
preferential use of wet areas during the UP to the incorporation of
dry areas with low PAW values during the EP.
This change in the wetness characteristics can be visualized
using kernel density estimation (Baxter et al., 1997). We use here an
Epanechnikov kernel and a bandwidth of 2 for the PAW values of
either entity (Fig. 7). Greater kernel density values indicate
a greater frequency of values at respective PAW values. The results
show a significant shift toward smaller values from UP to EP. While
the peak PAW value in the UP is about 38, the PAW peaks at about
34 for the EP.
4. Discussion
Our results indicate a change in landuse patterns from UP to EP
that is accompanied by a significant change in the wetness characteristics of the used areas. Within the same landscape, activities
of UP peoples were restricted to wet areas, while the inhabitants of
the region during the EP expanded their landuse to include dry
areas. This leads us to conclude that a significant change in the
subsistence strategy occurred during the EP. The widespread,
unconnected activity areas deduced from the EP landuse model
suggest the evolution of new landuse strategies. In addition to an
increased PAW range for these activity areas, this observation
points to a habitat expansion and a rise in the resource diversity
potentially available to the EP population. Furthermore, significantly lower PAW values of the areas used during the EP, compared
to the UP, indicate a degree of independence from wet parts of the
landscape.
Our conclusions on changes in the subsistence strategy are
drawn from the computational models and surface finds. Despite
this new approach to the question of landuse and subsistence, our
results are in good agreement with results from more traditional
approaches using data from archaeological excavations. For
example, studies on cultural remains from Baaz Rockshelter indicate an increase in the diversity of both the prey spectrum
(Napierala et al., in press-a-b) and the lithic tool assemblages
between the UP and the EP layers there (Barth, 2006; Hillgruber,
2010). The good agreement between our survey results and those
from Baaz Rockshelter indicates that our approach using PAW as an
analytical tool can be implemented as an independent line of
evidence and as a possibility to add a new spatial perspective in the
analysis of the interaction between people and environments.
The TDASP survey documents that artifacts from the EP are
consistently found in areas of higher altitude (Conard et al., 2006d;
Dodonov et al., 2007). Moreover, they are often located on prominent and often highly exposed elevated points in the landscape.
2276
K. Bretzke et al. / Journal of Archaeological Science 39 (2012) 2272e2279
Fig. 4. Distribution of artifact densities used to model landuse. A) Distribution of artifact densities and negative evidence for the UP; circle size indicates density; white points
represent localities without UP artifacts (negative evidence). B) Distribution of artifact densities and negative evidence for the EP; symbols similar to A.
Areas of higher altitude are located in the northwestern part of the
TDASP study region and characterized by higher precipitation.
Despite higher precipitation, the wetness is less concentrated in
this part of the landscape due to the lower potential for the accumulation of runoff. Our results show that EP populations were able
to use areas with less wetness and were thus potentially able to use
areas situated upstream and at higher elevations more systematically than UP populations. Hence it can be concluded that the
increased degree of independence from wet areas is an adaptation
that allowed the successful expansion of EP spatial activities into
elevated areas.
Elevated areas in the TDASP study region are mainly located on
the eastern flank of the Anti-Lebanon Mountains, a region where
wild stands of cereals potentially grew (Conard et al., in press;
Deckers et al., 2009). Based on the conclusion that EP populations
were able to access elevated areas, we can conclude that EP
populations were also able to access this important resource
(see Figs. 4B and 6B). Mortars and pestles found in EP contexts at
Fig. 5. Distribution of wetness based on the estimation of PAW. Dark areas indicate an
increased amount of accumulated surface and subsurface water. White areas are
relative dry.
Baaz Rockshelter (Conard, 2000), Kaus Kozah Cave (Conard and
Masri, 2006) and several survey localities (Conard et al., 2006c)
also suggest that the EP inhabitants of the region occasionally
processed cereals, in addition to pigments, nuts and perhaps other
materials (Hillgruber, 2010). Although there are no direct indications for the regular use of cereals, the evidence from our landuse
models points to the likelihood that EP populations in Western
Syria may have accessed wild stands of cereals.
It remains uncertain during which period of the EP the change in
landuse occurred. The EP can be subdivided into the Early
(ca. 22,000e16,000 cal BC), the Middle (ca. 16,000e13,100 cal BC)
and Late EP (13,100e9600 cal BC) (Goring-Morris et al., 2009).
While the Early EP contains the Kebaran, the Geometric Kebaran
belongs to the Middle EP. The Late EP is comprised of the Early and
Late Natufian. Hillgruber’s (2010) study of EP lithics from Western
Syria points to a paucity of evidence for Early Natufian groups in
this region. This observation narrows the options for when the shift
in landuse took place and suggests that this change in landuse
occurred either during the Early/Middle EP or near the start of the
Late Natufian.
The frequent occurrence of backed artifacts and microliths
during the Kebaran and Geometric Kebaran (Hillgruber, 2010)
marks a noticeable technological difference between the Early/
Middle EP and the Late UP, where such pieces occur less frequently.
Moreover, these artifacts indicate new hunting technologies, which
may have dictated changes in landuse. Despite these differences in
the material culture, archaeozoological studies suggest that, like
their UP predecessors, Kebaran and Geometric Kebaran populations
were mobile hunter gatherers relying on a similar spectrum of prey
(Banning et al., 2001; Bar-Oz and Dayan, 1999; Stiner, 2005; Stutz
et al., 2009). Given the similarities in the economic behavior
between the Late UP and the Early/Middle EP, we suspect that the
main shift in landuse may postdate the Early/Middle EP.
Although the chronological resolution of the survey finds is too
low to be certain, we hypothesize that the most significant shift in
landuse in TDASP study region occurred with the start of the Late
Natufian. An argument for the hypothesized shift in landuse with
the advent of the Late Natufian may be reflected in the changing
material culture. The archaeological record of the Late Natufian in
our study region is characterized by a variety of differences in
material cultural compared to the UP and Early/Middle EP. The Late
Natufian archaeological record contains evidence of residential
structures (Conard, 2000), pestles and mortars (Conard et al.,
2006a,b), and a greater diversity of hunted game (Napierala et al.,
in press-a-b). Since these data suggest multiple differences in
K. Bretzke et al. / Journal of Archaeological Science 39 (2012) 2272e2279
2277
Fig. 6. Landuse model for the Upper Paleolithic (A) and the Epipaleolithic (B) based on data from the Tübingen-Damascus Excavation and Survey Project. Black areas indicate areas
of increased artifact density.
lifeways and subsistence, the Late Natufian seems to be a better
candidate for changes in landuse as seen in our results. This is in
keeping with many publications on subsistence during the EP,
which suggest that a significant change occurred during the
Natufian (Baadsgaard et al., 2010; Bar-Yosef, 1998; Gilead, 1988,
1991; Munro, 2003; Stiner, 2001; Stutz et al., 2009). Usually this
change is demonstrated for the Early Natufian, but since the Early
Natufian is poorly documented in the TDASP study region
(Hillgruber, 2010), the Late Natufian is the most plausible candidate
in our regional context. We plan to test this hypothesis in the future
by conducting detailed studies of the lithic assemblages from the
survey localities with the goal of establishing improved chronological control of the history of occupation and landuse in the
TDASP survey area.
After identifying the probably period in which landuse changed,
we can go one step further and ask the next question. What triggered this development toward the use of drier, often elevated,
areas? The Late Natufian is often cited as an example of human
response to environmental change, which suggests a natural
trigger. Indeed, there is ample evidence that the environmental
conditions changed dramatically during the Younger Dryas (YD),
a dry spell chronologically associated with the Late Natufian
(Bar-Matthews et al., 1999; Frumkin et al., 1999). In the southern
Levant the YD has often been seen as a trigger for the decline of
large scale villages in a semi-sedentary settlement system during
the Early Natufian toward smaller and more mobile groups during
the Late Natufian (Bar-Yosef, 1987; Bar-Yosef and Valla, 1990).
However, new data challenges this model (Deckers et al., 2009;
Henry, in press; Lev-Yadun and Weinstein-Evron, 2005; Napierala
kernel density
0.10
EP
UP
0.08
0.06
0.04
0.02
0
30
35
40
45
50
PAW
Fig. 7. Kernel density estimates of the PAW values deduced from the Upper Paleolithic
(UP) and Epipaleolithic (EP) landuse models. The shift to lower PAW values during the
EP indicates an altered landscape use.
et al., in press-a) by suggesting that the environmental change in
some parts of the Levant was of small magnitude, if present at all.
Our data from the TDASP study region do not document a sharp
environmental decline during the YD. Trout remains from the Late
Natufian assemblage of layer III in Baaz Rockshelter (Napierala
et al., in press-b) suggest a sufficient supply of water to a perennial river system in the vicinity of the site. Furthermore, Deckers
et al. (2009) present evidence suggesting that woodland vegetation was not fully established during the preceding Bölling/Alleröd
interstadial. Instead it is probable that trees and shrubs were only
thinly scattered over the study region. A scenario also concluded
for the YD in our study area. The archaeobotanical data suggest
a degree of continuity in the environmental conditions between
the Early EP and the Late EP and argues against a noticeable
impact of the YD, and thus against a natural trigger for the initiation of changes in EP landuse. Indeed, nearly constant environmental condition between the Early EP and the Late EP on the one
hand, and a significant change in the landuse on the other, suggest
an increased importance of the social-economic behavior, as
opposed to natural environmental causality, in shaping patterns of
landuse.
The landuse pattern deduced for the UP stands in stark
contrast to the EP pattern. We interpret the strong relationship
between wet areas and UP activities bounded within one part of
the landscape as a response of the UP populations to decreased
precipitation (Bar-Matthews and Ayalon, 2003). If we consider
the importance of water for the development of the biosphere,
we can conclude that under a regime of decreased precipitation,
in addition to the areas in the immediate vicinity of perennial
springs, only those areas where surface and subsurface water
accumulates would offer adequate biomass to allow for an
archaeologically visible human presence. Such areas serve as
magnets, leading to a concentration of activities within
a restricted part of the landscape, as visible in the distribution of
artifacts from the UP. Although arguing here for natural causes
being mainly responsible for the shape of the UP landuse pattern,
we do not exclude the possibility that the landuse pattern could
also be influenced by the social environment, as discussed by
Bretzke (2008) in more detail.
The discussion above highlights the potential of a spatially
explicit approach to the study of human behavior with respect to
environmental characteristics. The strength of this approach is its
quantitative character that provides the possibility of adding
a spatial perspective to the study of the interaction between
humans and their environment.
2278
K. Bretzke et al. / Journal of Archaeological Science 39 (2012) 2272e2279
5. Conclusion
This paper demonstrates the potential of spatially explicit
studies of human landuse. Based on readily available data on
topography and precipitation we introduced an approach to
transform low resolution precipitation data into high resolution
surface-wetness data to model local variability in ecological
conditions. By comparing the wetness distribution pattern with
landuse patterns of UP and EP human groups from the Western
Syria, we identified significant differences in the use of the landscape between both periods. While landuse during the UP was
characterized by a contraction of activities into relatively small,
wetter parts of the landscape, EP landuse featured the exploitation
of a broader range of areas with different environmental characteristics including dry and elevated parts of the landscape.
Furthermore, the spatially explicit and quantitative character of our
approach allows us to investigate the response of human landuse to
changing ecological conditions. We conclude from our survey
results and the discussion of the climatic background that the UP
pattern is largely influenced by natural conditions. In contrast, the
significant change in landuse, which likely took place during the
Late Natufian, cannot entirely be explained by changes in the
natural environment. The available climatic data lack evidence for
a significant change in the local environmental conditions during
the Late Natufian. Hence we assume an additional trigger probably
resulting from changes in the social-economic environment.
Acknowledgments
We are grateful to the Syrian General Directorate of Antiquities
and Museums and especially to our colleagues B. Jamous,
M. Maqdisi, M. Masri, M. Hamoud and S. Muhesen for providing
access to the study area. We thank A.W. Kandel for discussion and
proofreading. We are also deeply indebted to A. Dodonov, who
made great contributions to the geoarchaeology of the TDASP
project. This work was funded by the Heidelberger Akademie der
Wissenschaften, the Deutsche Forschungsgemeinschaft grant Co
226/20-1, and the Department of Early Prehistory and Quaternary
Ecology of the University of Tübingen.
Reference
Baadsgaard, A., Chazan, M., Scott Cummings, L., Janetski, J.C., 2010. Natufian strategy
shifts: evidences from Wadi Mataha 2, Petra, Jordan. Eurasian Prehist. 7, 9e30.
Bakdach, J., 2000. Die paläolithischen Freilandfundstellen auf dem Hochplateau um
Jabrud und in der Nabek-Ebene. Damaszener Mitteilungen 12, 1e16.
Banks, W.E., d’Errico, F., Dibble, H.L., Krishtalka, L., West, D., Olszewski, D.I.,
Peterson, A.T., Anderson, D.G., Gillam, J.C., Montet-White, A., Crucifix, M.,
Marean, C.W., Sánchez-Goñi, M.-F., Wolfarth, B., Vanhaeren, M., 2006. Ecocultural niche modeling: new tools for reconstructing the geography and
ecology of past human populations. PaleoAnthropology 4, 68e83.
Banks, W.E., d’Errico, F., Peterson, A.T., Vanhaeren, M., Kageyama, M., Sepulchree, P.,
Ramsteine, G., Jostf, A., Lunt, D., 2008. Human ecological niches and ranges
during the LGM in Europe derived from an application of eco-cultural niche
modeling. J. Archaeol. Sci. 35, 481e491.
Banning, E.B., Parslow, C., Betts, M., Lohr, M., Maher, L., 2001. Middle Epipalaeolithic
sites in Wadi Ziqlab, Northern Jordan. Paléorient 27, 5e19.
Barling, R.D., Moore, I.D., Grayson, R.B., 1994. A quasi-dynamic wetness index for
characterizing the spatial distribution of zones of surface saturation and soil
water content. Water. Resour. Res. 30, 1029e1044.
Bar-Matthews, M., Ayalon, A., 2003. Climatic conditions in the Eastern Mediterranian during the last glacial (60-10 kyr bp) and their relations to the Upper
Paleolithic in the Levant: oxygene and carbon isotope systematics of cave
deposits. In: Goring-Morris, A.N., Belfer-Cohen, A. (Eds.), More than Meets the
Eye: Studies on Upper Palaeolithic Diversity in the Near East. Oxbow, Oxford,
pp. 13e18.
Bar-Matthews, M., Ayalon, A., Kaufman, A., Wasserburg, G.J., 1999. The eastern
Mediterranean paleoclimate as a reflection of regional events: Soreq cave,
Israel. Earth. Planet. Sci. Lett. 166, 85e95.
Bar-Oz, G., Dayan, T., 1999. The Epipalaeolithic faunal sequence in Israel: a view
from Neve David. J. Archaeol. Sci. 26, 67e82.
Barth, M.M., 2006. The lithic artifacts of Baaz Rockshelter. In: Conard, N.J. (Ed.),
Tübingen Damascus Excavation and Survey Project 1999e2005. Kerns Verlag,
Tübingen, pp. 25e110.
Bar-Yosef, O., 1987. The late Pleistocene in the Levant. In: Soffer, O. (Ed.), The
Pleistocene Old World: Regional Perspectives. Plenum Press, New York,
pp. 219e236.
Bar-Yosef, O., 1998. The Natufian culture in the Levant, threshold to the origins of
agriculture. Evol. Anthropol. 6, 159e177.
Bar-Yosef, O., Valla, F., 1990. The Natufian culture and the origin of the Neolithic in
the Levant. Curr. Anthropol. 31, 433e436.
Baxter, M.J., Beardah, C.C., Wright, R.V.S., 1997. Some archaeological applications of
kernel density estimates. J. Archaeol. Sci. 24, 347e354.
Beven, K.J., Kirkby, M., 1979. A physically-based, variable contributing area model of
basin hydrology. Hydrol. Sci. Bull. 24, 43e69.
Binford, L.R., 2001. Constructing Frames of Reference: An Analytical Method for
Archaeological Theory Building Using Hunter-gatherer and Environmental Data
Sets. University of California Press, Berkeley.
Bretzke, K., 2008. Oberflächenfunde als Quelle siedlungsarchäologischer Untersuchungen. Eine Fallstudie zum Jungpaläolithikum der Zentrallevante. Mitt. Ges.
Urgesch. Blaubeuren 17, 69e96.
Bretzke, K., Kandel, A.W., Conard, N.J., 2011. Deducing landuse patterns from
archaeological survey data. Chron. Archéol. Syrie 5, 11e18.
Burke, A., Ebert, D., Cardille, J., Dauth, D., 2008. Paleoethology as a tool for the
development of archaeological models of land-use: the Crimean Middle
Palaeolithic. J. Archaeol. Sci. 35, 894e904.
Collard, I.F., Foley, R.A., 2002. Latitudinal patterns and environmental determinants
of recent human cultural diversity: do humans follow biogeographical rules?
Evol. Ecol. Res. 4, 371e383.
Conard, N.J., 2000. An overview of the recent excavations at Baaz Rockshelter,
Damascus Province, Syria. In: Aslan, R., Blum, S., Kastl, G., Schweizer, F.,
Thumm, D. (Eds.), Mauer Schau - Festschrift für Manfred Korfmann. Verlag
Bernhard, Remshalden-Grunbach, pp. 623e640.
Conard, N.J., 2006. Tübingen - Damascus Excavation and Survey Project 1999e2005.
Kerns, Tübingen.
Conard, N.J., Masri, M., 2006. The 2004 excavation at Kaus Kozah cave. In:
Conard, N.J. (Ed.), Tübingen Damascus Excavation and Survey Project
1999e2005. Kerns Verlag, Tübingen, pp. 189e192.
Conard, N.J., Kandel, A.W., Abdurachman, A., 2006a. The 2000 excavation at Baaz
Rockshelter. In: Conard, N.J. (Ed.), Tübingen Damascus Excavation and Survey
Project 1999-2005. Kerns Verlag, Tübingen, pp. 171e175.
Conard, N.J., Kandel, A.W., Dyab, A., 2006b. The 1999 excavation at Baaz Rockshelter.
In: Conard, N.J. (Ed.), Tübingen Damascus Excavation and Survey Project
1999e2005. Kerns Verlag, Tübingen, pp. 161e166.
Conard, N.J., Kandel, A.W., Masri, M., 2006c. Results of the 2005 TDASP survey in the
Damascus Province, Syria. In: Conard, N.J. (Ed.), Tübingen Damascus Excavation
and Survey Project 1999e2005. Kerns Verlag, Tübingen, pp. 337e346.
Conard, N.J., Kandel, A.W., Masri, M., 2006d. The 2003 Ma’aloula Paleolithic survey.
In: Conard, N.J. (Ed.), Tübingen Damascus Excavation and Survey Project
1999e2005. Kerns Verlag, Tübingen, pp. 311e318.
Conard, N.J., Bretzke, K., Masri, M., Napierala, H., Kandel, A.W., 2010. Modeling
middle Paleolithic landuse in the Damascus Province, Syria. In: Conard, N.J.,
Delagnes, A. (Eds.), 2010. Settlement Dynamics of the Middle Paleolithic and
Middle Stone Age, Vol. III. Kerns Verlag, Tübingen, pp. 123e144.
Conard, N.J., Bretzke, K., Deckers, K., Hillgruber, K.F., Kandel, A.W., Masri, M.,
Napierala, H., Riehl, S., Stahlschmidt, M., in press. Natufian lifeways in the
eastern foothills of the Anti-Lebanon Mountains. In: Bar-Yosef, O., Valla, F.
(Eds.), The Natufian Culture of the Levant II, International Monographs in
Prehistory, Ann Arbor.
Conolly, J., Lake, M., 2006. Geographical Information Systems in Archaeology.
Cambridge University Press, Cambridge.
d’Errico, F., Sanchez-Goni, M.-F., 2003. Neanderthal extinction and the millennial
scale climatic variability of the OIS 3. Quaternary Sci. Rev. 22, 769e788.
Deckers, K., Riehl, S., Jenkins, E., Rosen, A., Dodonov, A., Simakova, A.N., Conard, N.J.,
2009. Vegetation development and human occupation in the Damascus region
of southwestern Syria from the Late Pleistocene to Holocene. Veg. Hist.
Archaeobot 18, 329e340.
Dodonov, A.E., Kandel, A.W., Simakova, A.N., al-Masri, M., Conard, N.J., 2007.
Geomorphology, site distribution, and Paleolithic settlement dynamics of the
Ma’aloula Region, Damascus Province, Syria. Geoarchaeology 22, 589e606.
Dunnell, R.C., Dancey, W.S., 1983. The siteless survey: a regional scale data collection
strategy. J. Archaeol. Method Th. 6, 267e287.
Enzel, Y., Arnit, R., Dayan, U., Crouvi, O., Kahana, R., Ziv, B., Sharon, D., 2008. The
climatic and physiographic controls of the eastern Mediterranean over the late
Pleistocene climates in the southern Levant and its neighboring deserts. Global.
Planet. Change. 60, 165e192.
Foley, R., 1981. Off-site archaeology: an alternative for the short-sited. In: Hodder, I.,
Isaac, G., Hammond, N. (Eds.), Patterns of the Past: Studies in Honour of David
Clark. Cambridge University Press, Cambridge, pp. 157e183.
Foley, R., 1995. The adaptive legacy of human evolution: a search for the environment of evolutionary adaptedness. Evol. Anthropol. 4, 194e203.
Frumkin, A., Ford, D.C., Schwarcz, H.P., 1999. Continental paleoclimatic record of the
last 170,000 years in Jerusalem. Quaternary Res. 51, 317e327.
Gamble, C., Davis, W., Pettitt, P., Richards, M., 2004. Climate change and evolving
human diversity in Europe during the last glacial. Philos. T. Roy. Soc. B. B359,
243e254.
K. Bretzke et al. / Journal of Archaeological Science 39 (2012) 2272e2279
Gilead, I., 1988. The Upper Palaeolithic to Epi-Palaeolithic transition in the Levant.
Paléorient 14, 177e182.
Gilead, I., 1991. The Upper Paleolithic period in the Levant. J. World. Prehist. 5,
105e154.
Goring-Morris, N., Hovers, E., Belfer-Cohen, A., 2009. The dynamics of Pleistocene
and Early Holocene settlement patterns and human adaptations in the Levant:
an overview. In: Shea, J.J., Lieberman, D.E. (Eds.), Transitions in Prehistory:
Essays in Honor of Ofer Bar-Yosef. Oxbow Books, Oxford.
Henry, D.O., in press. The Natufian and the Younger Dryas. In: Bar-Yosef, O., Valla, F.
(Eds.), The Natufian Culture of the Levant II, International Monographs in
Prehistory, Ann Arbor.
Hillgruber, K.F., 2010. The Last Hunter-Gatherers: The Epipaleolithic in Southwestern Syria. Dep. of Early Prehistory and Quaternary Ecology, University of
Tuebingen, Tuebingen.
Isaac, G.L., 1981. Stone age visiting cards: approaches to the study of early land-use
pattern. In: Hodder, I., Isaac, G., Hammond, N. (Eds.), Patterns of the Past: Studies
in Honour of David Clark. Cambridge University Press, Cambridge, pp. 131e155.
Lev-Yadun, S., Weinstein-Evron, M., 2005. Modeling the influence of wood use
by the Natufians of El-Wad in the forest of Mount Carmel. J. Isr. Prehist. Soc. Mitekufat Haeven 35, 285e298.
Munro, N.D., 2003. Small game, the Younger Dryas, and the transition to agriculture
in the southern Levant. Mitt. Ges. Urgesch. Blaubeuren 12, 47e71.
Napierala, H., Kandel, A.W., Conard, N.J., in press-a. Small game and the shifting
subsistence patterns from the Upper Paleolithic to the Natufian at Baaz Rockshelter, Syria. In: Maskour, M., Beech, M. (Eds.), Archaeozoology of the Near East
IX, Al Ain, U.A.E.
Napierala, H., Van Neer, W., Kandel, A.W., Peters, J., Uerpmann, H.-P., Conard, N.J., in
press-b. Fish in the desert? The Younger Dryas and its influence on the paleoenvironment at Baaz Rockshelter, Syria. In: Bar-Yosef, O., Valla, F. (Eds.), The
2279
Natufian Culture of the Levant II, International Monographs in Prehistory, Ann
Arbor.
Rust, A., 1950. Die Höhlenfunde von Jabrud (Syrien). Karl Wachtholz Verlag,
Neumünster.
Schmidt, F., Persson, A., 2003. Comparison of DEM data capture and topographic
wetness indices. Precis. Agric. 4, 179e192.
Schulz, J., 2000. Handbuch der Ökozonen. Verlag Eugen Ulmer, Stuttgart.
Solecki, R., Solecki, R., 1966. New data from Yabroud. Les Annales Archéologiques
Arabes Syriennes: Revue d’Archéologie et d’Histoire 14, 121e153.
Solecki, R., Solecki, R., 1987/1988. Archaeological researches at Yabroud, Syria and
vicinity, summer 1987. Ann. Arch. Syr. 37/38, 9e49.
Sørensen, R., Zinko, U., Seibert, J., 2005. On the calculation of the topographic
wetness index: evaluation of different methods based on field observations.
Hydrol. Earth Syst. Sci. Discuss. 2, 1807e1834.
Stiner, M.C., 2001. Thirty years on the “broad spectrum revolution” and Paleolithic
demography. Proc. Natl. Acad. Sci. U S A 98, 6993e6996.
Stiner, M.C., 2005. Paleocommunities and trends in species as prey. In: Stiner, M.C.
(Ed.), The Fauna of Hayonim Cave, Israel. A 200.000-Year Record of Paleolithic
Diet, Demography, and Society. Peabody Museum of Archaeology and Ethnography, Cambridge, pp. 127e148.
Stutz, A.J., Munro, N.D., Bar-Oz, G., 2009. Increasing the resolution of the broad
spectrum revolution in the southern Levantine Epipaleolithic (19-12 ka).
J. Hum. Evol. 56, 294e306.
Suzuki, H., Kobori, I., 1970. Report of the Reconnaissance Survey on Palaeolithic Sites
in Lebanon and Syria. The University Museum, Tokyo.
Syrian Arab Republic, 1977. Climatic Atlas of Syria, Climate Division. Meteorological
Department. Ministry of Defence, Damascus.
United States Geological Survey, 2007. Shuttle radar topographic mission. http://
srtm.usgs.gov/index.php (accessed 06.06.08.).