LATE PLEISTOCENE

Chapter 1 (Pages 11-20)

Pre-History and Archaeology Glossary

Excerpts and Definitions and Addendums:

During much of the period of the last glaciation the inland basins of Palestine and Transjordan were filled with pluvial lakes. The largest of these was the Lisan Lake which flooded much of the Rift Valley at present occupied by the Sea of Galilee, the River Jordan and the Dead Sea. This lake came into existence after about 70,000 BP and was maintained at its highest level from about 50,000 to 20,000 BP during the Lisan Stage (See Page 26 and Figure 16 in *1 Below). It was then about 220 kilometres long although no more than 17 kilometres wide and its surface was at about 180 metres below mean sea level, some 200 metres above the present surface of the Dead Sea (See Page 25 ibid).

It is believed that the Lisan lake was created during a period of increased precipitation or at least at a time when there was more run-off of surface water and less evaporation in the Rift valley (See Page 26 ibid). The ecological equilibrium in this region is easily disturbed so even small changes in climate can have a great effect on the environment. It has been calculated that a rainfall increase of as little as 200 metres in the Rift Valley catchment would be sufficient to create the Lisan Lake (See Page 46 in *2 Below) without any change in other variables. We know from other evidence that the temperature fell by at least 5 degrees C during the last glaciation as Butzer has pointed out (See Page 393 in *3 Below) would have increased effective precipitation by reducing evaporation. It may be that there was little more actual rainfall but combined with the drop in temperature this was enough to fill the Lisan basin.

Although the lake was so large it was always too salty for fish and molluscs to live in it (See Page 81 in *1) and thus useless to man as a source of food. Because of its length it would also have hindered rapid communication between the Judean uplands and Transjordan.

During the last glaciation another very large lake existed in the Jafr basin to the east of Maan and smaller ones in the Azrak depression and perhaps elsewhere in Transjordan. The Jafr lake was between 1000 and 1800 square kilometres in area. It was a freshwater lake and its sediments were rich in molluscs. An upper layer of the lake deposits has been dated to 27,700 ± 87O BP Hv-1719.

Further north the Damascus basin was also flooded during the Wurm [glacial stage] and there was a small lake in the Barada gorge. Large bodies of fresh-water such as the Jafr and Damascus basin lakes would have created highly favourable environments for man. They contained an abundance of fish and molluscs while their marshy shores would have attracted game. Their surface area was sufficiently great for evaporatioan to create greater humidity in the region than now. This would have made moisture available for plants through both increased precipitation amd dew.

About 18,000 BC the Lisan Lake shrank until its surface lay at about 37 metres below mean sea level (See Page 26 ibid); this happened very quickly, perhaps within a millennium. The water level of the lake dropped so much during this "Dead Sea Stage that the Lisan split into four relict lakes: the Sea of Galilee, a lake at Beth-Shan and two lakes in the Dead Sea basin. It is believed that this rapid transformation was partly caused by tectonic subsidence but either a decrease in precipitation or increase in evaporation or a combination of the two must also have taken place. The water level remained low for several millennia but then rose again apparently because of an increase in available moisture. This phase of higher water level is dated by a single C-14 date of 7900 +/- 150 BC (See Page 28 ibid).

The Jafr lake gradually dried up sometime after 26,000 BC, a process that is probably associated with the demise of the Lisan Lake. There followed a very arid stage of uncertain length. This was succeeded by a phase of effective precipitation during which mudflats were formed in the Jafr basin. The sequence in the Jafr basin matches that in the Rift Valley so although the mudflat phase is not accurately dated it can probably be correlated with the phase of Dead Sea higher water level about 8000 BC.

More arid conditions during the late glacial also caused the Damascus basin lake to diminish in area. This happened some time after 20,000 BC and the end of the regression phase itself has a single C-14 determination of 17,040 +/- 520 BC Hv-4471. There was considerable aeolian erosion during this phase as in the Jafr basin.

Recent work in Nubia indicates that the climatic pattern in Egypt was closely related to conditions in the southern Levant during the Wurm. The climate there was semi-arid but cooler for much of this period with more available moisture than today (See Pages 397 and 404 in *3 Below). The principal phase of increased moisture took place during the earlier Wurm between 50,000 and 25,000 BP when substantial beds of gravels and silts were laid down by the increased discharge of the Nile and its tributaries (See Page 395ff ibid). This phase was contemporary with the Lisan Stage in the Jordan basin. Thereafter an arid phase ensued as in the southern Levant which lasted until about 15,000 BC. This was followed by two moister phases extending well into the Holocene. The first continued until about 10,000 BC and is associated with the deposition of the Darau member of the Jebel Silsila formation (See Page 396 ibid). The second lasted from about 9200 until 6000 BC during which period the Arminna member was formed. These phases coincided with the high level of the Dead Sea that occurred in the Levant at the time of the transition from Pleistocene to Holocene. The coincidence is not exact as on present evidence the two Egyptian phases spanned a longer period than the Levantine phase. After 6000 BC the Egyptian climate became arid and apart from another moist interval in the 4th millennium BC it has remained so until today.

The principal source of information about the vegetation during this period comes from palynological studies of cores drilled in lacustrine and riverine sediments. These studies are also an important additional source of information about climatic conditions even if the data are open to conflicting interpretations.

A deep core, K-Jam, has been drilled in the bed of Lake Huleh into sediments which are believed to date from the present back to the Riss-Wurm interglacial (See Page 266 in *4). A pollen sequence has been prepared from this core which shows fluctuations in vegetation that reflect the same climatic changes as the pluvial lake sediments. Samples taken from a depth of between 50 and 35 metres had relatively high values of arboreal pollen (See Page 267 ibid), principally tabor oak (quercus ithaburensis), indicating that the Galilee hills to the west were clothed with deciduous oak forest. Gramineae and Cyperaceae were poorly represented because Lake Huleh was more extensive then but there was abundant open field vegetation around the lake. This phase apparently can be equated with the Lisan Stage when there was more available moisture than today.

The vegetation changed during the next phase from 35 to 25 metres in the core. The total percentage of arboreal pollen decreased, reflecting principally a sharp decrease in oak, but the values of olive (Olea europaea), pistachio (Pistacia species), Aleppo pine (Pinus halepensis) and cypress (Cupressus species) pollen slightly increased (ibid). As these are all Mediterranean species it appears that the former oak forest in the hills was replaced by open maquis vegetation. Both Gramineae and Artemisia pollen increased; these and other vegetational changes suggest that Lake Huleh diminished in size. Horowitz attributes these developments to the onset of a warmer humid climate (ibid), a conclusion which his evidence appears to contradict. The replacement of oak forest by maquis and the shrinking of Lake Huleh suggest rather that there was less available moisture then. A sample from the 30 metre level has been dated by C-14 to 16,850 ± 195 B.C. Hv-1725 (See Page 264 ibid) so this phase appears to be contemporary with the contraction of the pluvial lakes. We know from the evidence of the deep-sea cores that the temperature was near its glacial minimum at this date, another fact which conflicts with Horowitz's interpretation.

The phase above from 25 to 18 metres in the core was marked by a rise in arboreal pollen mostly from oak: the percentage of Mediterranean tree pollen diminished (See Diagram on Page 260 ibid). Pollen of open field species increased with the oak but pollen of marsh plants declined as the lake expanded. Horowitz believes that these changes were caused by cooler moister conditions (See Page 267 ibid) but this is partly contradicted by other evidence which suggests that the post-glacial rise in temperature had alreadly begun. This phase in the pollen core corresponds to the rise in level of the Dead Sea in the lower Jordan valley.

The pollen evidence suggests that northern Sinai carried a denser vegetation during the Wurm than today (See Page 221 in *5 Below) with scrub cover at least and even some trees in the Negev highlands. There were some variations in the pollen record reflecting fluctuations in the climate during the Wurm. One phase of somewhat richer vegetation in the Negev during the Natufian can be detected from pollen analysis of samples from the site of Rosh Zin (See Page 47 in *6 Below). Much of the pollen was from Chenopodiaceae but there was a little arboreal pollen from evergreen oak (Quercus calliprinos) and olive, species that could not survive in the area today.

Pollen analysis of samples from stations on the seaward side of the Mountains of Lebanon has also shown that the vegetation there varied during the Wurm under the influence of a changing climate. The Nahr Ibrahim samples indicate that at one stage during the earlier Wurm the slopes of the mountains were covered with forest almost down to present sea-level. The forest was composed partly of deciduous species such as lime and hazel with cedar, pine and ivy also present. Conditions needed to be both cooler and moister then than now for these deciduous species to flourish and, as we have seen, there is supporting evidence for such a climatic change from other sources.

Very few late Wurm pollen samples have been analysed from Lebanon but there are some indications that the vegetation thinned out in response to a decrease in available moisture. Then almost at the end of the Wurm perhaps about 12,000 or 11,000 BC tree pollen increased during a moister interlude. This phase may be correlated with a similar fluctuation in Palestine about this time.

Pollen analysis of a core from Sahl Aadra north-east of Damascus gives further support to this pattern of vegetation change in Palestine and Lebanon. The core was drilled through sediments at the edge of the Damascus basin pluvial lake and it spans the later Wurm after perhaps 22,000 BC. Zones 1 and 2 at the bottcm of the core, dated by two determinations of 21,555 +/- 350 BC Hv-4468 and 20,060 ± 350 BC Hv-4469, had high herbaceous and Gramineae pollen values with some Chenopodiaceae and a maximum of 14% arboreal pollen. The latter was principally composed of cedar and pine with a little oak, walnut and some o1ive. There were traces of silver line (Tilia tomentosa), hornbeam (Carpinus orientalis/ostrya) and wing nut (Pterocarya) which are not found so far south today because they cannot tolerate such a warm dry climate. The arboreal species would have occupied the hills around the Damascus basin and the Anti-Lebanon Mountains to the west. The profile suggests that the climate during this phase was cooler than today with more available moisture.

After a high water phase in zone 3 the lake retreated in zone 4 as the climate grew more arid; arboreal pollen fell to less than 8% and the herbaceous species, Gramineae and Chenopodiaceae dominated the profile. Cedar and pine pollen decreased sharply although oak and walnut were still present in small numbers.

Another significant change in the vegetation pattern took place in zone 5. Gramineae dominated the profile at first but silver lime was present once more. Then the arboreal pollen, consisting mostly of cedar, increased to over 35%. Marsh plants were once more very abundant which suggests that the Damascus basin lake expanded during this phase. Conditions were now moister than before. The full development of this stage took place sometime after 17,000 BC.

Central and northern Syria appear to have experienced a climate and vegetation different from the rest of the Levant during the Wurm. There is no clear evidence that substantial bodies of water like the Lisan and Jafr lakes formed in the inland drainage basins north of Damascus which at once suggests that there was little more available moisture in the region than now, despite the lower temperature. This view is supported by the pollen evidence from a core drilled in the Ghab section of the Orontes valley 15 kilometres south of Jisr esh-Shaghur. The core spans much of the period of the last glaciation and the earlier Holocene. Throughout this time the Ghab contained a lake surrounded by marshes which survived well into the Holocene (See Page 751 in *7). This lake was created by tectonic movement in the Orontes Valley and its continuance was determined more by the local geology than by climatic conditions. The pollen which collected in the lake records the past vegetation of Jebel Alawiye (Ansariye) and Jebel Zawiye to the east as well as the Ghab itself.

The samples from the earliest section of the core, zones S, T and U, indicate a spread of forest cover on both Jebel Alawiye and Jebel Zawiye which reached a peak in sub-zone T1 (See Page 747 ibid). The climate would have been cooler than at present throughout this period and with much the same amount of available moisture as today in zones T and U although drier in zone S. The core is dated by three C-14 determinations only; extrapolation from these indicates that this earliest section may date from before 45,000 BP (See Page 751ff and Table 2 ibid).

Thereafter towards the end of the glaciation from zone V through to Y the forest decreased, giving way to steppe. The loss of tree cover was most extreme in sub-zone Y5 when the only forest left was on the seaward side of Jebel Alawiye; Jebel Zawiye was steppic as it had been continuously since zone V (See Page 750 and Figure 4 ibid). There were variations in vegetation during this long period as the climate fluctuated but for most of the time it was drier and often cooler than now. Only towards the end of this long dry period does there seem to be a close correlation with the southern Levant where in response to drier conditions after perhaps 18,000 BC the Lisan and Jafr lakes shrank.

The vegetation in the Ghab region changed markedly in zone Z (See Page 750ff ibid). In response to an increase in both temperature and available moisture the forest expanded over the whole of Jebel Alawiye and clothed Jebel Zawiye with light tree cover. It is suggested that the moisture increased briefly beyond present day levels allowing the vegetation to reach a slightly richer climax than today in sub-zone Z2. The vegetation changes recorded in zone Z can probably be related to changes in vegetation and lake levels that took place in Lebanon and Palestine at the end of the Wurm. It is not possible to date zone Z exactly but Niklewski and van Zeist suggest on the evidence of the latest of the Ghab C-14 dates that it may have begun about 9400 BC although they admit a possible margin of error of several thousand years (See Page 752 ibid); on their evidence it is more likely that zone Z began earlier rather than later. This would permit a closer correlation with changes elsewhere in the Levant even though these are admittedly almost equally uncertainly dated.

Seeds preserved in the Mesolithic deposits at Tell Abu Hureyra throw light on the environment further east at this time (See Page 70ff in *8 Below). The vegetation was steppic but quite rich in species compared with the present. It is possible that some trees such as backberry and turpentine were also growing in the area; if so the rainfall must have been slightly higher then. Such an environment would be quite similar to that of today if man had not disturbed the vegetation. An intermediate zone of open forest would have lain between the forested coastal mountains and the park-like steppe around Abu Hureyra.

The pollen record of the Ghab core is almost the only detailed evidence we have for vegetation and climatic change in northern Syria during the later Wurm but the pattern it reveals is supported by other pollen cores from Greece and the Iranian Zagros. Wijmstra has made a detailed study of the upper section of a deep core drilled at Tenaghi Philippon in Macedonia. Zones V, P and X of this core were characterised by high counts of Artemisia and Chenopodiaceae and low values for arboreal pollen (See Page 525 in *9 Below). These zones cover much of the period of the last glaciation from about 50,000 BP to 14,600 BP (See Pages 523 and 527 and Figure 2 ibid). Wijmstra suggests that the climate was both cooler and drier then than now (See Page 526 ibid) which accords with the evidence of the Ghab core from Syria. The proportion of arboreal pollen increased through zone Y reaching modern levels about 8000 BC.

The Tenaghi Philippon data are supported by another pollen core from Ioannina in Epirus. Although this was drilled at a higher elevation than the Tenaghi Philippon core it nevertheless shows the same vegetational and climatic trends. Zones IIC and III were characterised by high values of herbaceous pollen and low arboreal pollen caused by cool dry conditions (See Page 28 in *10 Below). These zones cover the period from approximately 40,000 to before 10,000 BP. At the beginning of zone IV Artemisia pollen decreased and oak pollen increased as the forest expanded into a previously steppic area, a phase which corresponds to zone Y at Tenaghi Philippon.

Several cores have been drilled in late Quaternary deposits in the floors of valleys in the Zagros Mountains. From pollen analysis of these cores it is possible to reconstruct the pattern of vegetation in the region from the late Pleistocene into the Holocene. In zone A1 of the 63-J core from Lake Zeribar there were very high proportions of Chenopodiaceae and Artemisia but almost no arboreal pollen (See Figure 3 in *11). This zone is dated to between 20,000 and 12,000 BC and it is suggested that the climate then was cooler and drier than today. Analysis of two other cores from Lalabad Springs and Lake Nilofar in the Kermanshah Valley shows that similar conditions prevailed further to the south-east (See Page 309 ibid). Chenopodiaceae and Artemisia values remained high in zones A2 and B of the Zeribar core but plantain (Plantago) pollen now formed 10% of the diagram and arboreal pollen, mostly oak, became significant for the first time. Van Zeist suggests that this indicates the spread of savanna vegetation caused by an increase in both temperature and precipitation (See Page 310 ibid). A gradual expansion of trees continued until about 40OO BC; the percentage of arboreal pollen then rose rapidly as an oak forest developed in response to an increase in available moisture. This pattern of vegetation change is supported by the results from another core drilled at a lower elevation at Lake Mirabad in the Saidmarreh valley. The change from a late Pleistocene vegetation formed under arid conditions to a savanna and then later an oak forest as temperature and available moisture increased matches both the Ghab core and those from northern Greece ...