Climate, settlement patterns and olive horticulture in the southern Levant during the Early Bronze and Intermediate Bronze Ages (c. 3600–1950 BC)

Abstract

We report results of palynological investigation of a core of sediments extracted from the bottom of the Sea of Galilee. The core was sampled at high resolution for both palynological analysis (a sample was taken c. every 40 years) and radiocarbon dating. The article focuses on the Early Bronze and Intermediate Bronze Ages, c. 3600–1950 BC. The results enable reconstruction of the vegetation and thus climate in the lake's fluvial and alluvial catchment, which includes large parts of northern Israel and Lebanon and south-western Syria. The study sheds light on topics such as changes in olive cultivation through time and regions, processes of urbanization and collapse and settlement expansion and retraction in the arid zones.

Keywords:

• Early Bronze Age
• Intermediate Bronze Age
• southern Levant
• olive domestication
• climate change
• urbanization
• urban collapse
• Egypt and the Levant

Acknowledgments

This study was funded by the European Research Council under the European Community's Seventh Framework Program (FP7/2007–2013)/ERC grant agreement no. 229418. We wish to thank Michael Kitin for pollen sample preparation and Ahuva Almogi-Labin for her help in collecting recent pollen samples. Tal Langgut, Mark Cavanagh and Itamar Ben-Ezra are acknowledged for their help in drawing the different figures. Thanks are also due to Mordechai Stein (Israel Geological Survey) and Thomas Litt and his team from the Bonn Palynological Laboratory (University of Bonn) for their part in the extraction of the Sea of Galilee core. We would like to thank the Editor, Philip Graham, and the two anonymous reviewers for their constructive comments.

Notes

1 For a possible longer period of activity, covering the entire period discussed here, see Avner and Carmi 2001; Sebbane et al. 1993.

2 An exception is the well-dated high-resolution Soreq Cave (Israel) isotopic record that covers the 5000–2000 BC time interval (Bar-Matthews and Ayalon, 2011).

3 For example, in the late Holocene palynological investigation conducted by Baruch (1986) in the Sea of Galilee, the time interval of 3300–1600 BC is represented by only three pollen samples. Usually the pollen sampling resolution is restricted due to limited financial resources (pollen identification is time consuming).

4 Olive orchards produce a large crop one year and small crops in alternate years (=Biennial bearing; e.g. Lavee 2007). Olive is mainly a wind-pollinated species and therefore it releases large amounts of pollen to the atmosphere in the spring (March–May) to compensate for the low pollination efficiency that characterizes wind-pollinated trees (Baruch 1993).

5 At least 500 terrestrial pollen grains were counted per sample. Pollen grains were identified to the lowest possible systematic level. A reference collection of Israel's pollen flora (The Steinhardt Museum of Natural History, Tel Aviv University), as well as pollen atlases (e.g., Beug 2004; Reille 1995; 1998; 1999), were used for identification. A detailed pollen diagram of the Sea of Galilee is presented by Langgut et al. 2015.

6 Within this group, only trees common to the Mediterranean vegetation territory were included; the majority among them is wind-pollinated trees. The most dominant are evergreen and deciduous oaks (Quercus calliprinos and Quercus ithaburensis, respectively). Other Mediterranean trees appear in lower percentages: Phillyrea, Pistacia (pistachios), Pinus halepensis (Aleppo pine) and Ceratonia siliqua (carob tree).

7 The most dominant plants within this group are (in declining order): Poaceae (wild grasses), Cerealia (cereals), Asteraceae (daisy family), Chenopodiaceae (goosefoot family), Artemisia (sagebrush) and Brassicaceae (cabbage family).

8 Trees and shrubs common to the bank of the lake vegetation (hydrophilic plants), as well as aquatic plants, were excluded from the total pollen sum (Arboreal Pollen + Non Arboreal Pollen = 100%). The palynological diagram (Fig. 2) was plotted using the POLPAL program (Walanus and Nalepka 1999).

9 In addition, some of the trees are feral and hybrids between domesticated and wild.

10 Samples for recent pollen investigation were collected from the depth of 0–1 cm of the topmost lake sediments (see more details in the Supplementary Material).

11 Usually, in age-depth models in lakes the variability used is derived mainly from the organic material that is being dated. In our age-depth model, we chose to include radiocarbon dates which mostly resulted from short lived terrestrial organic material. Bulk organic material or plants remains of aquatic origin (or from unknown origin) were not included in order to prevent the need for a correction due to the reservoir effects.

12 The width of the different strata from which the new organic material was extracted was up to 5 cm (Table 1).

13 The new ages have a relatively wide range of uncertainty since each sample included several tiny twigs, to differ from the singe twig, larger in size, that were used in the previous samples (Table 1).

14 Other palynological investigations in the northern Levant which show an increase in Olea values during the Holocene — Orontes Valley and the Ghab area (Niklewski and van Zeist 1970; Yasuda et al. 2000) — proved to be difficult in establishing a robust age model (e.g., Cappers et al. 1998; Meadows 2005).

15 The conflicting data concerning the region of olive domestication (southern vs. northern Levant) presented here, may derive from sampling issues within the Besnard et al. (2013) study, in which samples from Israel were collected only from one location (Mount Carmel; Besnard et al. 2013: supplementary information Table S1). Further, according to the authors, owing to the highly fragmented and human-disturbed Mediterranean habitat, oleaster populations were mainly collected from present-day orchards; yet, it could not be excluded that some of the sampled trees/populations were feral (i.e. issued from cultivars). In any event, further genetic analyses are required in order to solve this regional discrepancy.

16 On the impact of olive cultivation upon perceptions of the landscape, and attitudes to territory, see Philip 2003.

17 See questions concerning origin of these vessels raised by Porat and Goren 2002.

18 Grape pollen (Vitis) is under represented in the palynological spectrum due to very low pollen dispersal efficiency, and therefore can hardly be traced in the Sea of Galilee pollen record.

19 ‘Olive oil’ as a West Semitic loanword (zayit) into Egyptian ( ddt) does not appear until the 13th century BC (Late Bronze Age). Many words for types of oil appear in the earliest Egyptian texts of the Naqada III to Old Kingdom (i.e. the Early Bronze Age), but many of these remain undeciphered and it is unclear which might refer to olive oil. However, Ahituv (1996) has argued well that the general word for oil bαq was probably used for olive oil through much of the Pharaonic period.

20 Even though commercial relations between Egypt and the southern Levant most probably triggered the establishment of specialized olive oil production in the region during the EB I, it does not seem to have been crucial for its continuation on a lower scale (local consumption) during the EB II–III (Salavert 2008).

21 For the diverse economy of the EB II, including the manufacture of specialized products such as Northern Canaanite Metallic Ware and the establishment of new networks of exchange, see the latest syntheses, Greenberg 2011; 2014; in press; Milevski 2011. Further, Berger (2013) showed that the EB II inhabitants of Tel Bet Yerah added some new fruits to their diet in comparison to the EB I inhabitants (Berger 2013: fig. 14).

22 However, it should be noted that some chronological uncertainties should be taken into consideration with this particular study (Yasuda 1997) due to relatively poor radiocarbon dating.

23 The differences between the pollen and the isotopic records lie only within the fluctuations of the general trends, since both records point to relatively humid climate conditions during the period under discussion. The slight discrepancies may derive from differences in sampling resolution (the Soreq Cave record was sampled in higher resolution), and/or differences within the dating methods (¹⁴C vs. Uranium-Thorium) and/or can be related to the slower responses of the vegetation in comparison to the more sensitive isotopic proxy.

24 The peak in Olea frequencies during the Iron Age I has also been noted in the high resolution Dead Sea (Ze'elim) palynological record (Langgut et al. 2014b: fig. 2).

25 The Soreq Cave speleothems isotopic record points to decrease in precipitation ∼2200–1900 BC, which is mainly evident in the increasing values of the δ¹³C with a low at 2000 BC. However, the estimated annual rainfall during this aridity did not reach below 500 mm (the mean annual rainfall today; Bar-Matthews and Ayalon 2004: figs 10b and 12). Declining Dead Sea levels during the period of 2250–1900 BC also indicate increased arid conditions (Migowski et al. 2006: fig. 3). The isotopic composition of tamarisk wood from the Mount Sedom Cave on the south-western margins of the Dead Sea shows a succession of droughts at ∼2200–1930 BC, with a prominent but short-lived dry event at approximately 2020 BC, followed by a longer event at approximately 1930 BC that ultimately killed the tree (Frumkin 2009), meaning that the dry event could have lasted even longer.

26 We were unable to compare our results with the late Holocene palynological study conducted by Baruch (1986) in Sea of Galilee, due to its low sampling resolution.