1.1. The Athiaki catchment area

2Athiaki runs through the Limassol Basin in a geomorphological context dominated by steepped-hills (chap. 2, fig. 2-1). These hills were formed in different geological times (from Miocene to Upper Pleistocene) and are related to the uplift and simultaneous erosion of the Troodos Massif and Messinian marine regressions and Pliocene transgressions (Pantazis 1967). They are characterized mainly by the superimposition of calcareous formations of Lefkara (70–23My) and Pakhna (23–7My) and deposited on the Moni clayey formation (75–70My). Small streams run in the thalwegs between the hills. The tectonic activity of the region and the simultaneous uplift of the Troodos Massif contributed to the formation of alluvial and marine terraces during the Quaternary.

3Athiaki is one of the region’s mainstreams (Briois et al. 2005). Its activity built the landscape through a mixture of erosion and sedimentary deposits to create alluvial terraces. Debris fans have been identified in the upstream and median sector while alluvial terraces can be observed in different sectors along the catchment basin. These alluvial formations depend on the initial topography (lower hills, larger floodplain) and the type of geological formation, which is why alluvial terraces differ innumber or height along the catchment basin.

4In the middle sector of the catchment basin, the alluvial dynamic created three fill terraces on the left bank’s debris fan (fig. 3-2). The top of these terraces is situated 25 m higher than the actual streambed and is the closest point of alluvial formation to the Klimonas site. Here, alluvial deposits of various granulometric sizes alternate with deposits of fine sediments affected by pedogenetic processes characterized as paleosoils.

Fig. 3-2 – View from the west of the middle sector’s alluvial terraces in the Athiaki Basin.

1.2. The Klimonas slope terraces and their regolith

5The site itself is situated on a stepped-hills at 30–140 m altitude, in a context dominated by colluvium deposits from the Late Glacial Maximum and Holocene: the Havara (local name for calcrete; Pantazis 1967, Schirmer 1998) and Lefkara formations (Benech et al., 2017: fig. 15). A profile along the site’s slope shows the different deposits on which the site was settled (Vigne et al. 2011).

6Together with the study of this road profile (Vigne et al. 2011a), the numerous test excavations and stripping opened on the Klimonas slope (chap. 2) allowed us to map the regolith and geological substratum (Benech et al. 2017). All the recorded Limassol Basin Sedimentary formations are represented (fig. 3-3; see details and profile in appendix 3-1): i) Upper Cretaceous ultrabasic rocks and dark green clays (Mamonia MP_L and Moni formation, gh²’/gh²’’fromthe geological map of Cyprus, Pantazis 1967), ii) Late Cretaceous/Paleogene grey-green marls of the lower Lefkhara formation (hi^{1 marls}), iii) Eocene and Oligocene chalks and biocalcarenites of the upper Lefkhara formation (hi², hi³’), iv) Miocene Pakhna-Koronia carbonated formations (j³’’ and j⁴; fig. 3-4). The latter lies horizontally in discordance with the former, which is locally upright due to neotectonics movement, especially visible in the large fault that borders the site to the west. In addition, several strongly recarbonated colluvial deposits, calcretes (or havara) were observed in places, mainly dating from the final Pleistocene.

Fig. 3-3 – Map of the regolithic outcrops observed on the Klimonas slope. References to the corresponding geological bodies are from the geological map of Cyprus, after Pantazis(1967).

Fig. 3-4 – Geological map of the Klimonas region, eastern Limassol District and Limassol Forest in the north.

2.1. Middle course of the stream

2.1.1. Sampling strategy and analytical methods

7The geomorphological study of the Athiaki Stream was based on the study of its two most ancient terraces (Terraces 1 and 2) in the middle course of the fluvial basin. Six alluvial profiles were selected and studied from both sites. However, profiles AT1 and AT5 (from Terrace 1 and Terrace 2, respectively) warranted a more thorough analytical study (fig. 3-5 and 3-6).

Fig. 3-5 – Aerial photo of the Athiaki and Klimonas region with the georeferenced position of the six profiles under study.

Fig. 3-6 – Topographical section of the Klimonas-Athiaki Sector (A-B line on fig. 3-3).

8We sampled every layer recognized in the field for the four alluvial profiles ΑΤ2, AT3, AT4 and AT6 (appendix 3-3). For profiles AT1 and AT5, we collected 101 sediment samples of about 400 g, every 5 cm.

9All collected samples have been analysed using different sedimentological and geochemical techniques, such as granulometry, magnetic susceptibility, X-ray fluorescence, total and organic carbon analysis, soil micromorphology and radiocarbon dating. For detailed information on the techniques used see appendix 3-2.

2.1.2. Description of the AT1 and AT5 profiles from Terraces 1 and 2

10A detailed description of the profiles and radiometric data can be found in appendices 3-3 and 3-4, respectively. The analytical data has been treated using Principal Component Analyses (PCA, Past 4.03 software, appendix 3-5).

2.1.2.1. Alluvial Terrace 1: Profile AT1

11For the first terrace (T1), which is also the most ancient one, the study was focused on the AT1 profile (fig. 3-7). Here, the alluvial sedimentation of more than 3m was interrupted by six episodes affected by pedogenetic processes identified as paleosoils. Paleosoils are brown-grey in colour and characterized by different pedologic indicators, such as polyhedral structure, nodules of manganese and calcite on aggregates. Different alluvial lithofacies, such as Open Framework (OF), Fill Framework (FF), Uniform Suspension (US), Vertical grain sorting (GV) and Sand with concentration of gravels (SL), were identified in the deposits; these are used as indicators of different river transport dynamics (for a developed description of lithofacies see appendix 3-3; Devillers 2008, Collinson 2013). Paleosoil 1, at the base of the sequence, seemed to be very different from the rest of the paleosoils, according to its pedological features and thickness. However, some of these features, such as carbonate nodules, were also observed in paleosoil 4 and 5.

Fig. 3-7 – Drawing of profile AT1 from Terrace 1.

12Concerning profile AT1, radiocarbon dates have been processed on three paleosoils (tab. 3-1): the first, the third and the fifth. Organic matter from the first paleosoil gave a date of 22 166 – 21 411 cal BC (Stuiver and Reimer 1993). Considering the potential contamination of the organic matter from more recent deposits, as is usual for paleosoils, we propose to refer this event to the end of the Last Glacial Maximum. The third paleosoil was dated to 23 225 – 22 528 cal BC. This date is statistically more ancient than that of the first paleosoil (T = 29.2, Chi² = 3.84; Calib test). The low quantity of organic matter in the soil (organic carbon: 0.12%, carbon 5.27%) can slightly change the radiocarbon date (Evin 1994); however, we considered that this date inversion was not a significant stratigraphic issue. The third date, obtained from paleosoil 5, was 7 676 – 7 517 cal BC. Combined with the presence of flint flakes in this paleosoil, this 8^th millennium radiocarbon date appears as a terminus post quem for human presence in the Athiaki Valley.

Tab. 3-1 – Radiocarbon dates from the AT1 profile.

2.1.2.2. Alluvial Terrace 2: Profile AT5

13Profile AT5 was situated in the same sector as AT1 but on the second terrace (T2). At the bottom of the profile (fig. 3-8), the layers were brown and clayey. The paleosoil 1 contained manganese and calcite nodules. On the other hand, from layer 4 upwards, the sediments were loamy silty and grey/yellow. Concerning the alluvial deposits, we identified the floodplain facies of SL, SU, MS and riverbed facies (FF) deposits alternated with paleosoils.

Fig. 3-8 – Drawing of profile AT5 from Terrace 2.

14Radiocarbon dating was conducted on all the AT5 paleosoils (tab. 3-2). The first one dated back to 25 471 – 24 280 cal BC. This sets the formation during the Last Glacial Maximum, which is significantly more ancient than the date of the first paleosoil from the AT1 profile. The second paleosoil was dated at 26 328 – 25 744 cal BC, which was significantly more recent than the first one. This chronological inversion seems to be related to the bioturbation of paleosoil 1. The third paleosoil was dated to 18 694 – 18 197 cal BC, while the fourth was dated to 17 228 – 16 808 cal BC. Paleosoil number 5 gave a very different date, 3 956 – 3 767 cal BC, corresponding to the warm and humid Atlantic Period of the Middle Holocene.

Tab. 3-2 – Radiocarbon dates from the AT5 profile.

2.1.3. Reconstruction of the terraces’ geometry and chronology

15The graphic representation of selected geochemical and sedimentological markers (appendix 3-5) was used to understand the stream’s dynamic and the paleoenvironmental conditions.

16According to the radiocarbon dates, the terraces formation started at least from the Last Glacial Maximum (LGM) until the Holocene. Based on these dates, the geometry of terrace formations seems to be complex (fig. 3-9). At least three phases of incision-sedimentation contributed to the formation of the three terraces. Sediment accumulations between 29 000 and 26 000 cal BC were incised during the Last Glacial Maximum, with continuous activity during the Late Glacial Period and the beginning of the Holocene covering the first formation. At the beginning of the Holocene, due to strong and sudden warming, a second incision occurred. Following this, the Holocene sedimentation created a second terrace whose deposits extend at least until the Middle Holocene (around 4 000 cal BC). A third incision then took place creating the third terrace, which is not a part of our study. Paleosoil 9 (profile AT1 paleosoil 5: AT1-PL5) seems to be contemporary with the Klimonas occupation (c. 8 800 cal BC; Vigne et al. 2012, 2017b, Manning 2014, chap. 11) even if the date of the paleosoils are usually younger than their formation (Evin 1994).

Fig. 3-9 – A, Diagram of terraces’ geometry from the middle sector of Athiaki according to the paleosoils’ radiocarbon dates. B, Diagram presenting the hypothesis of alluvial terrace formation and the two incisions in Athiaki’s middle sector.

2.2. Mapping the Athiaki’s alluvial terrace from upstream to downstream

17A systematic survey of the lower course’s alluvial terrace, between the Ayios Tychonas village and Amathus Hill, was done during the 2016 and 2017 fieldwork seasons. The alluvial terraces were identified and mapped from a geomorphological and sedimentological perspective thanks to natural alluvial cross sections (fig. 3-10). While three terraces have been identified in the middle course, only two were observed in the downstream. Our field survey does not include the Amathus site.

Fig. 3-10 – Geomorphological map of the Athiaki Basin catchment and the profile of different alluvial terraces from different sectors.

18Terrace 1 is the most ancient. It was formed during the Last and Late Glacial Period and was observed mainly in the upstream of the Athiaki Basin. Terrace 2 and 3 were formed during the Holocene and were identified along the riverbed, from the upstream to the downstream sector. Terraces on the downstream are lower than in the upstream, where their height can be up to 25 metres.

2.3. Diachronic evolution of the catchment basin

2.3.1. Last Glacial Maximum Period

19According to the various kinds of analyses conducted on the Athiaki sediments (appendix 3-5) in relation to field observations, we were able to reconstruct the stream’s dynamic and the floodplain evolution. Athiaki probably constitutes the only site on the island with well-studied alluvial terraces combining both deposits from the Last Glacial Maximum until the Middle Holocene (Deckers 2002, Devillers 2008, Hourani 2008, Waters et al. 2010, Ghilardi et al. 2015). The deposits covered a period of at least 19 000 years and recorded different environmental episodes.

20The results obtained by this study showed that the Last Glacial Maximum period was characterized by regular flows which contributed to the aggradation of the floodplain (sedimentation rate 0.18 cm/year). Periods with weak flow deposed fine sediments and contributed to the formation of paleosoils (paleosoil 1). In some cases, these stability phases lasted for a long time (paleosoil 3, fig. 3-11, A).

Fig. 3-11 – Synthetic diagram of the alluvial dynamic and soil conditions for each formation period, according to the different data collected. A: Initial deposits in the study sector; B: Deposits after the first incision; C: Deposits after the second incision.

21Concerning Paleosoil 1, the high value of magnetic susceptibility (appendix 3-5: fig. 2) and the presence of iron and manganese showed very humid environmental conditions in the soil which had contributed to the alteration of organic matter (Maher and Thompson 1995, Mandel and Arthur Bettis III 2001). Periods of slow flow alternated with periods of strong dynamics able to transfer coarser sediments. During periods of strong and irregular alluvial dynamics, secondary flows on the floodplain contributed to the formation of SL facies (sand with concentrations of gravels), corresponding to secondary concentrated flows on the floodplain as a result of irregular and strong rainfall. Strong velocity in the riverbed is related to GV facies (Vertical sorting), characteristic of semi-arid conditions (Devillers 2005), and is evidenced by riverbed deposits of various dynamic flow. The different facies observed in the field also showed the change of the river’s route and the shift of the riverbed during the Last Glacial Period.

22The latest date obtained for this formation is 23 000 years BC. According to palaeoclimatic data from the Eastern Mediterranean, at about 24 000 years BC the climate was cold with few precipitations (Bar-Matthews et al. 1999, 2000, Fontugne et al. 1999) related to the Heinrich 2 event (Bar-Matthews et al. 1997, 1999, Hemming 2009; fig. 3-9 and 11, A), which seems to create the first incision in Athaiki’s floodplain.

2.4. Athiaki and the wider Eastern Mediterranean context

27Despite the small size of its catchment basin, the Athiaki Stream created alluvial terraces of more than 25m high. Comparing the environmental conditions recorded in Athiaki with palaeoclimate or palaeoenvironmental studies in the Eastern Mediterranean and Cyprus (Bar-Matthews et al. 1997, 1999, 2000, Fontugne et al. 1999, Devillers 2008, Hourani 2008, Kuzucuoğlu et al. 2010, Lespez et al. 2013, Berger et al. 2014, 2016, Kuzucuoğlu 2015), we observed that Athiaki recorded both local and regional events (fig. 3-12). In comparison with geomorphological studies in a Cypriot alluvial context, such as Maroni and Gialias, where the deposits are dated to the Holocene, the observations in the Athiaki Basin extent our observations back 15 000 years.

Fig. 3-12 – Comparison graph of environmental and alluvial conditions between different regions in the Eastern Mediterranean: Greece, Cyprus, Israel and Turkey.

28Based on oxygen isotopes from the Soreq Cave in Israel (Bar-Matthews et al. 1999, 2000), the humid periods recorded in the speleothem were represented mainly by the formation of paleosoils on the Athiaki floodplain. In about 24 000–23 000 cal BC, a humid period was recorded in lake sediments in Central Anatolia, as well as in the speleothems (Bar-Matthews et al. 1997, Kuzucuoğlu et al. 2010); during this period, at least three paleosoils (3, 4 and 5) were formed on the Athiaki floodplain. Paleosoil 3 was noted as being very different from the others, with local conditions that characterized a humid environment and regular alluvial flow. At the end of the Last Glacial Maximum, even if climatic conditions were described as cold and dry in the Eastern Mediterranean (Bar-Matthews et al. 1999, Fontugne et al. 1999, Kuzucuoğlu et al. 2010), two more paleosoils (6 and 7) were formed in the Athiaki floodplain, which may demonstrate different local environmental conditions characterized as humid with regular river flow and the presence of vegetation.

29At the transition of the Holocene, the sedimentation was regular, as is attested in other Mediterranean alluvial contexts (Lespez and Dalongeville 1998, Macklin et al. 2002), and -environmental conditions were warmer and more humid than before (Wells et al. 1987, Bar-Matthews et al. 1999, 2000, Rossignol-Strick 1999). During this period, a sapropel was formed in the Eastern Mediterranean (Rossignol-Strick 1999), due to wet conditions and rainfall that allowed the transportation of organic sediments to the sea by river activity. From a palaeoclimatic point of view, these settings may have favoured paleosoil formation in the semi-arid context of Cyprus. This is attested in Athiaki by the formation of paleosoils on the Athiaki floodplain (PL8 and 9). These conditions seem to be similar in different regions across the Eastern Mediterranean during the Holocene; for example, during the same period, paleosoils were formed in Central Anatolia (Kuzucuoğlu 2015, Berger et al. 2016), at Sidari in Corfu, Ionian Sea (Berger et al. 2014, 2016) and at Dikili Tash (Lespez et al. 2013) in northern Greece. Paleosoils have also been identified in Gialias (Devillers 2005, 2008) in central Cyprus and possibly in the Maroni Stream (Hourani 2008) on the southern coast of Cyprus.

30As mentioned above, the 6.2 ky BC cooling event could be the reason for the second incision identified on the Athiaki floodplain. This event had an important impact on the local topography, since after this incision event the floodplain was located several metres lower than the previous level. The aridity of this period also had an impact on the Neolithic populations on the island (Weninger et al. 2006, Daune-Le Brun et al. 2017). This environmental instability is reflected in changes in the sedimentation of the Maroni Stream and is by a change in vegetation cover, after anthracological study (Daune-Le Brun et al. 2017). At this time, the Neolithic village of Khirokitia had been restricted and moved, while the site of Kalavassos-Tenta was abandoned (Weninger et al. 2006).

4.1. Reconstructing the original topography of the Klimonas terraces

34If we put together the lowest and middle sections of the road profile (fig. 3-14) we observe a difference in altitude of more than one metre between them and the level of the most recent paleosoil, which contained a typical PPNA hearth pit (chap. 13) and can therefore be considered as corresponding to either the settlement or the abandonment of the PPNA village c. 8 800 cal BC. This suggests that the middle and lower terraces of the Klimonas slopes are not only historical cultivation terraces, but that they already existed during the occupation of Klimonas and that this topography was more contrasted than today.

Fig. 3-14 – Profile of Klimonas’ terraces where the lower part has been formed by colluvial deposits. The green line shows the topography during the occupation.

4.2. Environment of the PPNA village

35The geomorphological study indicated that, at the time of the PPNA occupation of Klimonas, the climatic conditions were humid, the river flow was regular and the water table near to the surface. These stable conditions contributed to paleosoil formation on the alluvial plain.

36Ten thousand years ago, the palaeotopography of the Athiaki Valley was different than today (fig. 3-15) as only the first terrace had been created, the river was situated about 15 m higher and the floodplain was wider. Fine elements such as silt, clay and the high water table provided suitable conditions for all agricultural activities.

Fig. 3-15 – Topographical profiles of the Athiaki valley during the Klimonas occupation and today.