Holocene landscape dynamics at the tell Arslantepe, Malatya, Turkey – Soil erosion, buried soils and settlement layers, slope and river activity in a middle Euphrates catchment

""Alluvial and colluvial sequences were studied around the prehistoric tell Arslantepe in 11 exposures and additional auger cores. The chronology is based on 11 optically stimulated luminescence (OSL) ages, four radiocarbon ages, and the

Please download to get full document.

View again

of 18
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.


Publish on:

Views: 11 | Pages: 18

Extension: PDF | Download: 0

  The Holocene2014, Vol. 24(10) 1351  –1368© The Author(s) 2014Reprints and permissions:sagepub.co.uk/journalsPermissions.navDOI: 10.1177/0959683614540949hol.sagepub.com Introduction The reconstruction of the paleoenvironmental history has a long history in Turkey (e.g. Bottema et al., 1990; Brice, 1978; Butzer, 1958; Kraft et al., 1977; Louis, 1938). In particular, during the  past decades, a growing number of papers were published on Late Quaternary sea-level changes and siltation history of estuaries (e.g. Beach and Luzadder-Beach, 2008; Brückner, 2005; Brück-ner et al., 2002, 2006; Eisma, 1978; Erin ҫ , 1978; Fouache et al., 2012; Kayan, 1988), on lake sediments or adjacent wetlands (e.g. Eastwood et al., 1999; England et al., 2008; Kazancı et al., 2004; Kuzucuoğlu et al., 2011; Roberts et al., 2001; Vermoere et al., 2000; Wick et al., 2003), dune systems (e.g. Erin ҫ , 1970; Kuzucuoğlu et al., 1998), glaciers (e.g. Ak  ҫ ar et al., 2007; Zahno et al., 2010), and in particular on alluvial sequences (e.g. Kaniewski et al., 2007, 2008; Kuzucuoğlu et al., 2004; Miller-Rosen, 1997; Wilkinson, 1999) from various sites in Turkey. Sur- prisingly, few data are published so far on Holocene slope deposits (e.g. Bay, 1999; Dusar et al., 2012; Wilkinson, 2005). This is hard to understand since they reflect the geomorphic response on a local scale and are therefore appropriate indicators of local human impact on the landscapes easily connectable with settlement his-tories of the surrounding area (e.g. Dreibrodt et al., 2010b; Van Andel et al., 1990). Additionally, slope deposits contain valuable information on Late Glacial and Holocene woody vegetation via their mega-charcoal content, with a high spatial resolution unmatched by pollen analysis (Nelle et al., 2010, 2013). Some fluvial geomorphologists question the potential of colluvial deposits because of the risk of reworking and thus a low preserva-tion potential of older deposits (Dusar et al., 2011). But with an increasing archaeological record, a demand for more long-term Holocene landscape dynamics at the tell Arslantepe, Malatya, Turkey – Soil erosion, buried soils and settlement layers, slope and river activity in a middle Euphrates catchment Stefan Dreibrodt, 1,2  Carolin Lubos, 3  Johanna Lomax, 4  Gyorgy Sipos, 5  Tim Schroedter  1,2  and Oliver Nelle 6   Abstract Alluvial and colluvial sequences were studied around the prehistoric tell Arslantepe in 11 exposures and additional auger cores. The chronology is based on 11 optically stimulated luminescence (OSL) ages, four radiocarbon ages, and the embedded artifacts. Sediments contained wood charcoals, providing information on former vegetation. Fluvial activity is documented during Late Glacial times (15.4 ± 2.5, 12.8 ± 3.1 kyr) and frequently after Roman times. Slope and soil erosion occurred in the early (10.6 ± 1.4, 8.2 ± 0.7 kyr) and mid–late Holocene (6.7 ± 0.9, 5.4 ± 0.7–4.7 ± 0.7, 2.6 ± 0.2–2.5 ± 0.2, 1.9 ± 0.2–1.8 ± 0.2 kyr, and during the last 1000 years). The early Holocene erosion phases pre-date the so far established onset of settlement at the tell. This either indicates an earlier onset of agricultural land use than assumed or climatic influence on erosion, such as the 10.3 and 8.2 kyr climate events known from Western Europe. The erosion phases at around 5.0 and 2.6 kyr could reflect geomorphic responses to societal collapse (Late Chalcolithic state, Neo-Hittite kingdom) at Arslantepe. Most intensive Holocene soil erosion and landscape degradation occurred after occupation of the region by the Roman Empire. This is paralleled by the onset of river activity. A part of the lower neo-Hittite town as well as an early Holocene Terra Rossa–like soil that had formed rapidly were found buried. So far, no indication for mid–late Holocene fluvial activity of the adjacent creeks until Roman times has been found. Our results illustrate the large potential of slope deposits for long term reconstructions of human induced landscape transformation in Anatolia. Keywords buried soils and sites, early Holocene, geoanthracology, historical soil erosion, slope-river connectivity, TurkeyReceived 8 November 2013; revised manuscript accepted 12 May 2014 1 University of Kiel, Germany 2 Kiel Graduate School Human Development in Landscapes, Germany 3 University of Frankfurt/Main, Germany 4  Justus-Liebig-University Giessen, Germany 5 University of Szeged, Hungary 6 Aix-Marseille Université, France Corresponding author: Stefan Dreibrodt, Institute for Ecosystem Research, University of Kiel, Olshausenstraße 40, 24098 Kiel, Germany.Email: sdreibrodt@ecology.uni-kiel.de 540949 HOL 0   0   10.1177/0959683614540949TheHoloceneDreibrodt et al. research-article   2014 Research paper   at Universitaetsbibliothek Kiel on June 10, 2015hol.sagepub.comDownloaded from   1352 The Holocene   24(10) data about the history of human impact on the environment at the local scale is growing.This paper presents results of a reconstruction of Holocene soil erosion and river activity based on sequences of slope and alluvial sediments deposited adjacent to the tell site of Arslantepe, Malatya. Regional setting The investigated site Arslantepe is situated at 38°21′N, 38°19′E, 912 m a.s.l., in a distance of c . 15 km from the upper Euphrates. It lies in an extensional fault-bounded basin filled with a sequence of alluvial and lacustrine sediments of Neogene age (Türkmen et al., 2007) (Figure 1). The immediate surroundings of the site are characterized by Neogene sequences of clay rich lake sedi-ments alternating with sandy fluvial deposits (Palmieri, 1978). Both types of sediments contain carbonates; 700 m to the north-east, the remnant of an andesitic volcano probably of Miocene age (Aksoy et al., 2005) forms the eastern flank of the small creek valley of the Orduzu Dere, which passes the tell of Arslantepe at its eastern rim (Marcolongo and Palmieri, 1983). According to Marcolongo and Palmieri (1983), Arslantepe is situated favorably within the landscape because of a large hydrogeological catch-ment, supplying the site reliably with running water. Today, large  parts of the area of the Malatya Plain are covered by sediments  providing fertile soils for extensive apricot plantations (Colak et al., 2010). So far, the genesis and age of the young sediments exposed at the surface (Pleistocene or Holocene) have rarely been studied (e.g. Türkmen et al., 2007). According to the meteorologi-cal office of Turkey, the annual mean temperature of Malatya is 13.7°C, and the mean annual sum of precipitation equals 386 mm, with pronounced monthly maxima during the winter and spring season (1–41 mm, 2–37 mm, 3–51 mm, 4–58 mm, 5–47 mm, 6–18 mm, 7–2 mm, 8–2 mm, 9–7 mm, 10–38 mm, 11–45 mm, 12– 41 mm) in the period 1960–2012 (Devlet Meteroloji Işleri Genel Müdürlüğü, 2013). Archaeological record The excavations at the tell Arslantepe started in 1932 by French archaeologists (Delaporte, 1933) and have been under systematic excavation by the team of the University of Rome ‘La Sapienza’ since 1961 (http://w3.uniromal.it/arslantepe,Frangipane, 2012a). The mound was occupied continuously from the 5th Millennium BC  to the end of the Neo-Hittite period (8th century BC ). Addition-ally, some smaller remnants of a village from Roman times and a  palace from Medieval times were found (Frangipane, 2004). Approximately 30 m of archaeo-sediments document the long and complex stratigraphy of the mound and testify the continuity of occupation as well as the cultural and political importance within the region. The lowermost excavated layers so far date to 4250 BC  (Chalcolithic layers 1–2). The Late Chalcolithic layer 5 (3350– 3000 BC ) documents the emergence of a first state. A ‘palace’, a complex of monumental public buildings in which economic, reli-gious, and administrative activities were performed, was found in that period (Frangipane and Palmieri, 1983). After the palace was completely destroyed by fire at c . 3000 BC , phases of alternating economic and political organizations, herders and agriculturalists, foreign and local, with cultural relations to the Trans-Caucasians followed at the site (Di Nocera, 1998). Between 2900 and 2800 BC  again northern-Mesopotamian influence is visible in the material culture, and houses built of mud-bricks reoccur. In the following Figure 1.  (a and b) Research area and locations of the investigated well profiles (b), and the investigated watershed (c). In Figure 1c, the mapped catchment and deposition areas are delineated. The hatched area gives the maximum size and the crossed area the probable extension of catchment area (see text) for the deposited sediments (gray). The index map gives the distribution of deposition areas and the age of the deposits in kyr BP.  at Universitaetsbibliothek Kiel on June 10, 2015hol.sagepub.comDownloaded from   Dreibrodt et al. 1353 centuries of the Early Bronze Age, layers of alternating architec-ture were documented (wooden houses–mud-brick houses), and the emergence of canalization systems in the settlement at the very end of Early Bronze Age (until 2000 BC ) indicates the renewed development of a small urban center. This trend continued through the Middle Bronze Age (2000–1700 BC ) and Late Bronze Age (1700–1200 BC ), where Assyrian influence is deduced from the material culture and large fortification systems (Di Nocera, 1998). During the Iron Age (1100–700 BC ), the city was the capital of a regional Hittite Emperor containing a large palace and the famous Lions-Gate (Delaporte, 1933). After the conquest and destruction of the city by the Assyrians in 712 BC , the importance of the site as part of an Assyrian province declined. When a military camp close to the Euphrates River became the local center during the 1st century AD , only a small village was situated on the tell during Roman times. Most of the remnants of this settlement date to the 4th and 5th century AD  (Di Nocera, 1998). The sequence of archaeological finds at Arslantepe is completed by the remnants of an Islamic palace at the top of the mound dating to the 12th– 13th century AD . Material and methods Field methods Sequences of alluvial and colluvial deposits were studied in 11 exposures (the larger ones were un-mantled large watering holes) of a depth between 0.5 and 6 m and a diameter of between 1 and >10 m (Figure 1b). An example for one of these large exposures is given in photographs of Well 9 (Figure 4). Additionally, a supple-mental net of 15 auger cores was drilled to a maximum depth of about 5 m in a small deposition area close to the tell to enable access to the sediment sequences between the exposures. The  position of the studied profiles was determined by an etrex-GPS device (FA. Garmin), leveled between the wells and auger cores in the investigated small watershed, and documented on a large scale topographical map (1:5000). The layers and horizons were discriminated by their visual (colors) and tactile (density, texture)  properties, as well as embedded objects (artifacts, charcoal, stones). The content of carbonates was estimated in the field using the HCl (10%) test. All properties of the layers and horizons were identified according to Munsell Scale (Munsell, 1992) and instructions for soil mapping (Boden, 2005) and were docu-mented as sketches in the field book, and as scaled drawings and  photographs. For practical reasons, colluvial and alluvial layers were abbreviated as M/aM (lat. Migrare/alluvial Migrare) accord-ing to the German Soil Survey Instructions (AG Boden, 2005). Samples for laboratory analysis and charcoal analysis were taken from the exposures according to the stratigraphy and documented in the profile sketches. Laboratory methods Dating of the layers was carried out with the help of OSL, radio-carbon dating of embedded charcoal, and archaeological dating of embedded artifacts. To avoid confusion, the radiocarbon ages in the results and discussion are given in cal. yr BP. The OSL data are given in years before present day (for conventional and cali- brated radiocarbon ages, see Table 1). OSL dating.  For sediments containing no organic remains or archaeologically datable material, OSL dating was applied. Using light proof stainless core cutters (20 cm length; 8 cm diameter), undisturbed sediment samples were taken from the  profile sequences. The coarse grain quartz fraction (100– 200 µm) was used to determine the equivalent dose (  D e ), using a single aliquot regenerative (SAR)–dose protocol after Murray and Wintle (2000, 2003). Three dose recovery tests yielded dose recovery ratios within 6% of unity, supporting the applica- bility of the chosen protocols. For the standard measurements, at least 20 aliquots were measured, of which several had to be rejected for reasons of unsatisfactory recycling ratio (<10%) or high test dose errors (>10%). Preheat–Cutheat Temperatures were set to either 220–200°C or 260–220°C, according to the results of the dose recovery test. To detect for possible incom- plete resetting of the luminescence signal during the last pro-cess of erosion, transportation, and deposition, the samples were measured using small (2-mm) aliquots (Olley et al., 1998). The  D e  distributions did not imply incomplete bleaching, for example, strong skewness. Therefore, a weighted mean (Cen-tral age model after Galbraith et al., 1999) was used for mean  D e  calculation. The dose rate was obtained by low-level high-resolution γ-spectrometry using the dose rate conversion Table 1.  Radiocarbon and OSL age data. ProfileSampleLaboratory NrRadiocarbon age ( 14 C yr BP)Calibrated age (cal. a) (2 σ ) BC / AD ∆ 13 C (‰)Well 1Archaeo-sedimentKIA 409382744 ± 72cal. BC  1076–1065, 1056–794−25.98 ± 0.13Well 1M 2KIA 409392337 ± 40cal. BC  537–356, 286–234−24.50 ± 0.36Well 6Base of alluvial sediment, 818.5 m a.s.l.,  Juniperus KIA 463591433 ± 24cal. AD  582–654−26.11 ± 0.09Well 8Archaeological findingKIA 463601770 ± 30cal. AD  137–264, 275–333−26.64 ± 0.17Profile, sample 238 U (ppm) 232 Th (ppm)K (%)D *   (Gy/kyr) D e  (Gy)Aliquots measured/usedOSL age (kyr)Well 1, aM11.00 ± 0.023.10 ± 0.100.47 ± 0.010.87 ± 0.0913.4 ± 1.71315.4 ± 2.5Well 1, aM21.90 ± 0.044.96 ± 0.150.95 ± 0.011.65 ± 0.1821.0 ± 4.51012.8 ± 3.1Well 1, M11.73 ± 0.036.22 ± 0.181.07 ± 0.021.63 ± 0.1811.0 ± 0.9176.7 ± 0.9Well 1, M31.82 ± 0.036.75 ± 0.191.03 ± 0.011.63 ± 0.134.02 ± 0.04162.47 ± 0.22Well 1, M41601.95 ± 0.044.75 ± 0.070.86 ± 0.021.72 ± 0.113.22 ± 0.22171.9 ± 0.2Well 1, M4 801.97 ± 0.034.56 ± 0.100.84 ± 0.021.71 ± 0.103.03 ± 0.19151.8 ± 0.2Well 5, Cw22.59 ± 0.1010.88 ± 0.331.82 ± 0.073.34 ± 0.2435.54 ± 3.7236/2010.6 ± 1.4Well 5, M11.94 ± 0.037.19 ± 0.301.19 ± 0.022.15 ± 0.1417.57 ± 0.92178.2 ± 0.7Well 5, M21.86 ± 0.075.72 ± 0.290.92 ± 0.041.84 ± 0.1610.80 ± 0.9036/215.4 ± 0.7Well 5, M31.74 ± 0.075.54 ± 0.280.92 ± 0.041.80 ± 0.168.49 ± 1.0136/244.7 ± 0.7Well 5, M41.95 ± 0.034.68 ± 0.020.90 ± 0.021.77 ± 0.114.53 ± 0.26262.6 ± 0.2OSL: optically stimulated luminescence.  at Universitaetsbibliothek Kiel on June 10, 2015hol.sagepub.comDownloaded from   1354 The Holocene   24(10) factors of Adamiec and Aitken (1998). A water content of 7 ± 5% is included in the dose rate. Radiocarbon dating.  Radiocarbon dating was applied on charcoal. Objects suspected to be redeposited by bioturbation were not con-sidered. The charcoal samples were dated at the Leibniz Labora-tory of the Kiel University using standard cleaning and analysis  procedures (Grootes et al., 2004; Nadeau et al., 1997, 1998). The conventional 14 C ages were calculated according to Stuiver and Polach (1977). The calendar ages were calculated using OxCal 4.1 (Data set: IntCal04, Reimer et al., 2004). Dating of artifacts.  Artifacts embedded within the sediments were dated according to the local typology by the archaeologists and interpreted to give maximum ages for the respective layers. Charcoal analysis.  Samples of 10 L were taken from Wells 6 and 8 for charcoal analysis. The processing was done with the floata-tion method, using a mesh-size of 300 µm. Additional charcoal samples were taken via dry sieving (mesh 2 mm) from the archaeo-sediment layer of Well 1. After drying, all charcoal  pieces > 1 mm (‘mega-charcoal’, Robin et al., 2013) were identi-fied using a stereo lens (Nikon SMZ1500, 7.5×–112.5× magnifi-cation) and an incident light microscope (Nikon ME600) at magnifications of 100×, 200× and 500×, on freshly broken sur-faces, according to Schweingruber (1990a, 1990b), Fahn et al. (1986), and the reference collection of the Palaeoecology Work-ing Group, Institute for Ecosystem Research, Kiel University. Sedimentological analysis The samples were air dried prior to careful dry sieving (mesh 2 mm) and weighing to separate the fine earth from stones. Thereby, frag-ments of artifacts, mollusc shells, bones, and charcoal were removed, and their content was determined gravimetrically. Grain size distribution was performed using the Laser-Spectrometry method by Mastersizer 2000 particle size analyzer (Malvern Instru-ments). This instrument has a measuring range from 0.02 to 2000 µm, which results in continuous spectra of 70 classes of grain sizes. These classes were grouped in order to represent the standard grain size fractions and to enable a better comparison with other sedimentological studies: clay < 2 µm (30 grain size classes), silt 2–63 µm (20 grain size classes), and sand 63–2000 µm (20 grain size classes). The organic matter and carbonate contents were esti-mated using LOI (2 h) at 550°C and 940°C of the samples dried  prior at 105°C overnight. Magnetic susceptibility (MS) was deter-mined on the <2 mm fraction for selected samples at low frequency (0.47 kHz) using the 10 mL volume measurement method by a Bar-tington Instruments MS2B susceptibility meter (resolution 2 × 10 −6  Si, measuring range 1–9999 × 10 −5  Si, systematic error 10%). Each sample was measured three times. Obtained MS values were calibrated using a 1% Fe 3 O 4  calibration sample. For the bur-ied soil in Well 5 and its parent material, bulk density was measured on volumetric samples (100 cm 3 , n  = 3). After drying at 105°C overnight, the bulk density was calculated from the dry weight and the volume of the samples. Quantification of soil erosion and sediment budgeting  The estimation of sediment budgets and minimum soil erosion rates for the Holocene slope deposits in a small watershed close to the tell is based on (1) the mapping of the volume of the sediments (mean thickness and distribution), (2) assumed density of 1.5 g × cm −3  of the slope deposits (conversion of sediment vol-umes into masses), (3) the age of the sediments (deposition time), and (4) the delineation of the superficial catchment area for each layer. The values represent minimum estimates because some material might have been transferred into the adjacent creek and  beyond.Since measured bulk density values are only available for the  buried soil, we assumed a similar value of 1.5 g × cm −3  for the slope deposits. This value has also been widely used for Holocene slope deposits in previous studies (e.g. Dreibrodt et al., 2010b).  No pronounced differences of bulk densities between the layers and horizons were detectable in the field. Results Chronostratigraphy  Wells 1–5Well 1 .  At the base of Well 1, the Tertiary lake sediment described by Palmieri (1978) was exposed in a depth of 4 m (Fig-ure 2). Above that, a gravel layer (mean diameter of clasts ~ 10 cm) in a depth of about 4–3.83 m was present. Sandy–gravelly alluvial deposits covered this gravel (3.83–3.33 m). In a depth of 3.33– 3.18 m, a second gravel layer (mean diameter of clasts ~ 4.5 cm) was deposited. This was overlain by fine grained sandy–silty (3.18–2.85 m) alluvial sediment (aM1), dated by OSL to 15.4 ± 2.5 kyr. Above that, a silty alluvial deposit (aM2) (2.85– 2.64 m) has an OSL age of 12.8 ± 3.1 kyr (Figure 2 and Table 1). It contained very few small charcoal particles. A first thin collu-vial layer (M1) covered the alluvial sediment. It contained higher amounts of stones and sand as well as few charcoal particles and some undeterminable pottery fragments (2.64–2.53 m) and had an OSL age of 6.7 ± 0.9 kyr. The colluvial layer was overlain by a layer rich in unrounded pottery, daub, bones, teeth, and charcoal. This extraordinary layer was identified as an archaeo-sediment and was subdivided into two sub-layers (Y1, 2.53–2.13 m; Y2, 2.13–1.93 m) based on their different stone contents. The embed-ded pottery was classified as prehistoric Iron Age. A charcoal embedded in the lower sub-layer dates to 3.1–2.8 cal. kyr BP. The archaeo-sediment was covered by a thin pebble layer (M2, depth: 1.93–1.85 m, mean diameter of stones ~ 2 cm). Above, a sequence of colluvial layers more sandy than the alluvial deposits was pres-ent from a depth of 1.85 m to the present surface. According to color and constituents, the colluvial sediment was divided into discrete layers. From 1.85 to 1.55 m, a layer (M3) containing some charcoal particles was exposed. It dates to 2.5 ± 0.2 kyr according to the OSL age. An embedded charcoal of an age of 2.5–2.3 cal. kyr BP confirms the luminescence age. After that, a thick, humic and sandy homogeneous layer (M4) was deposited (1.55–0.70 m). OSL dates in the lower and upper part of this layer reveal an age of 1.9 ± 0.2 and 1.8 ± 0.2 kyr. The sequence was completed by another colluvial layer (M5) making up the upper 0.70 m of the profile. This layer contains some pottery fragments. The youngest of them dates to Byzantine times, which gives a maximum age for its formation. Well 2 .  The sequence of Well 2 situated c . 40 m to the east of Well 1 resembles roughly the stratigraphy of Well 1 (Figure 2). At the base in a depth of 3.52 m, the Tertiary lake sediment was over-lain by sandy alluvial deposits (3.52–3.02 m). A gravel layer was deposited on top of this sand (depth: 3.02–2.72 m, mean diameter of gravel 5 cm). A gray silty–sandy alluvial layer was deposited above the gravel layer (depth 2.72–2.01 m). Above that, a 20 cm thick humic colluvial layer was present (2.01–1.81 m). It con-tained some small charcoal and stones. On top of this, the archaeo-sediment of grayish brown color with artifacts, charcoal, and stones was observed (depth: 1.81–1.10 m). Because of different concentrations of artifacts and stones, it was subdivided into two sub-layers. It was covered by a layer of pebbles of a mean diam-eter of ~2 cm. A thin colluvial layer rich in organic matter was  preserved (1.10–0.97 m) in the overlying position. An additional colluvial layer (depth: 0.97–0.50 m) buried the mentioned layers. at Universitaetsbibliothek Kiel on June 10, 2015hol.sagepub.comDownloaded from   Dreibrodt et al. 1355    F   i  g  u  r  e    2 .    S  c   h  e  m  a  t   i  c  v   i  e  w  o   f  t   h  e  s  e   d   i  m  e  n  t  s  e  q  u  e  n  c  e  s   i  n   W  e   l   l  s   1 ,   2 ,  a  n   d   5 .  at Universitaetsbibliothek Kiel on June 10, 2015hol.sagepub.comDownloaded from 
Related Search
Similar documents
View more...
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks