Woolly mammoth (Mammuthus primigenius Blum.) and its environment in northern Europe during the last glaciation

Woolly mammoths were large, herbivorous, cold-adapted mammals of the Late Pleistocene. The diet and habitat requirements of the species set certain constraints on the palaeoenvironments it could occupy. The relationship between the mammoth’s shifting range and changing environments can be explored using independent data on ice sheet configuration, temperature, and vegetation, provided the locality and age of the fossil remains can be validated. Here we present a comprehensive record of occurrence of the woolly mammoth in the circum-Baltic region of northern Europe during the last glaciation, based on a compilation of radiocarbon-dated remains. The record shows that the mammoth was widespread in northern and north-eastern Europe during Marine Isotope Stage 3 (MIS 3), at 50,000–30,000 calibrated years ago (50–30 ka). The presence of the species up to 65°N latitude supports the restriction of the Scandinavian Ice Sheet (SIS) during MIS 3. The widest distribution range round 30 ka was followed by a decline that led to the disappearance of mammoths from the area during the maximum extent of the SIS, from 22 to 18 ka. The woolly mammoth re-colonized the Baltic region and southern Scandinavia after the onset of the late-glacial deglaciation at 17 ka. The late-glacial record suggests a markedly fluctuating population changing its range in tune with the rapid environmental changes. The last appearance of mammoth in our study region was in Estonia during the Younger Dryas (Greenland Stadial 1; GS1) at about 12 ka. The two major periods of occurrence during MIS 3 and the late-glacial stadial suggest that mammoth had a wide tolerance of open to semi-open tundra and steppe-tundra habitats with intermediately cold climate, whereas the 22–18 ka disappearance suggests a major southward and/or eastward retreat in response to extremely cold, glacial conditions near the SIS margin. The final regional extinction correlates with the re-forestation during the rapid warming at the Younger Dryas–Holocene boundary.     

AMS 14C chronology of the world's southernmost woolly mammoth (Mammuthus primigenius Blum.)

The world's definite southernmost woolly mammoth record is a molar from Ji’nan (around 36°N), Shandong Province, China. AMS ¹⁴C dating of the specimen, gave a conventional ¹⁴C age of 33,150±250 BP. The period of 40–30 ka BP corresponds to the later phase of the Marine Oxygen Isotope Stage 3 (MIS 3a), recognized as the global interstadial of the last glacial period. However, it is known that the winter monsoon strengthened in Asia during the period 35–33 ka BP, and the age of the woolly mammoth specimen from Ji’nan corresponds to that age. The specimen suggests that this area became cold and dry in 33 ka BP, and grassland or open forest, suitable habitat for woolly mammoths, developed during this short time span. This age is similar to the age of the southernmost woolly mammoth record in Europe, therefore supporting a hypothesis by Porter and An [1995. Correlation between climate events in the North Atlantic and China during the last glaciation. Nature 375, 305–308] that an important component of Chinese palaeoclimates may be linked to changes in North Atlantic oceanic conditions.     

Last straw versus Blitzkrieg overkill: Climate-driven changes in the Arctic Siberian mammoth population and the Late Pleistocene extinction problem

A set of radiocarbon dates on woolly mammoth were obtained from several regions of Arctic Siberia: the New Siberian Islands (n = 68), north of the Yana-Indigirka Lowland (n = 43), and the Taimyr Peninsula (n = 18). Based on these and earlier published dates (n = 201) from the East Arctic, a comparative analysis of the time-related density distribution of ¹⁴C dates was conducted. It was shown that the frequencies of ¹⁴C dates under certain conditions reflect temporal fluctuations in mammoth numbers. At the end of the Pleistocene the number of mammoths in the East Arctic changed in a cyclic manner in keeping with a general “Milankovitch-like” trend. The fluctuations in numbers at the end of the Pleistocene occurred synchronously with paleoenvironmental changes controlled by global climatic change. There were three minima of relative mammoth numbers during the last 50 000 years: 22 000, 14 500–19 000, and 9500 radiocarbon years ago, or around 26 000, 16–20 000, and 10 500 calendar years respectively. The last mammoths lived on the New Siberian Islands, which were connected to the continent at that time, 9470 ± 40 radiocarbon years ago (10 700 ± 70 calendar years BP). This new youngest date approximates the extinction time of mammoths in the last continental refugium of the Holarctic. The adverse combination of environmental parameters was apparently a major factor in the critical reduction in mammoth numbers. The dispersal of humans into the Arctic areas of Siberia no later than 28 000 radiocarbon years ago did not overtly influence animal numbers. Humans were not responsible for the destruction of a sustainable mammoth population. The expanding human population could have become fatal to mammoths during strong the minima of their numbers, one of which occurred at the very beginning of the Holocene.     

Evidence of a two-fold glacial advance during the last glacial maximum in the Tagliamento end moraine system (eastern Alps)

The glacial history of the Tagliamento morainic amphitheater (southeastern Alpine foreland, Italy) during the last glacial maximum (LGM) has been reconstructed by means of a geological survey and drillings, radiocarbon dating and pollen analysis in the amphitheater and in the sandur. Two phases of glacial culmination, separated by a distinct recession, are responsible for glacial landforms and related sediments in the outer part of the amphitheater. The age of the younger advance fits the chronology of the culmination of the last glaciation in the Alps, well established between 24 and 21 cal ka BP (20 to 17.5 ¹⁴C ka BP), whereas the first pulse between 26.5 and 23 cal ka BP (22 to 21 ¹⁴C ka BP), previously undated, was usually related to older (pre-LGM) glaciations by previous authors. Here, the first pulse is the most extensive LGM culmination, but is often buried by the subsequent pulse. The onset and final recession of the late Würm Alpine glaciation in the Tagliamento amphitheater are synchronous with the established global glacial maximum between 30 and 19 cal ka BP. The two-fold LGM glacial oscillation is interpreted as a millennial-scale modulation within the late Würm glaciation, caused by oscillations in inputs of southerly atmospheric airflows related to Dansgaard-Oeschger cycles. Phases of enhanced southerly circulation promoted increased rainfall and ice accumulation in the southern Alps.     

Chronology of the impact of Quaternary climate change on continental environments in the Arabian Peninsula

The remains of former lakes show that in the past the Arabian Peninsula experienced much wetter conditions than today. The last of these humid periods dates to about 10 to 5.5 ka ago. The chronological framework for an earlier humid phase, radiocarbon dated to some 35–20 ka, is inconsistent with evidence from other records from the region. Possibly, these ages are significantly underestimating the true depositional age due to methodological problems. The earliest phase of dune accumulation known so far is dated to the penultimate glaciation maximum of the mid latitudes (ca. 150 ka). Subsequently, dune accumulation occurred around 110 ka, 65 ka and 20 ka ago. All these phases concur with rapid drops in global sea level that caused a drying out of the Persian Gulf basin and of the shelf of the Oman coast. In contrast to some previous interpretations, it is concluded here that aeolian deposition has been limited by sediment supply and not by preservation potential.     

Marine-derived C calibration and activity record for the past 50,000 years updated from the Cariaco Basin

An expanded Cariaco Basin ¹⁴C chronology is tied to ²³⁰Th-dated Hulu Cave speleothem records in order to provide detailed marine-based ¹⁴C calibration for the past 50,000 years. The revised, high-resolution Cariaco ¹⁴C calibration record agrees well with data from ²³⁰Th-dated fossil corals back to 33 ka, with continued agreement despite increased scatter back to 50 ka, suggesting that the record provides accurate calibration back to the limits of radiocarbon dating. The calibration data document highly elevated Δ¹⁴C during the Glacial period. Carbon cycle box model simulations show that the majority of observed Δ¹⁴C change can be explained by increased ¹⁴C production. However, from 45 to 15 ka, Δ¹⁴C remains anomalously high, indicating that the distribution of radiocarbon between surface and deep ocean reservoirs was different than it is today. Additional observations of the magnitude, spatial extent and timing of deep ocean Δ¹⁴C shifts are critical for a complete understanding of observed Glacial Δ¹⁴C variability.     

Patterns of faunal extinction and paleoclimatic change from mid-Holocene mammoth and polar bear remains, Pribilof Islands, Alaska

Qagnaxˆ Cave, a lava tube cave on St. Paul Island in the Pribilofs, has recently produced a mid-Holocene vertebrate faunal assemblage including woolly mammoth, polar bear, caribou, and Arctic fox. Several dates on the mammoth remains converge on 5700 ¹⁴C yr BP. These dates, ~ 2300 yr younger than mammoth dates previously published from the Pribilof Islands, make these the youngest remains of proboscideans, and of non-extinct Quaternary megafauna, recovered from North America. Persistence of mammoths on the Pribilofs is most parsimoniously explained by the isolation of the Pribilofs and the lack of human presence in pre-Russian contact times, but an additional factor may have been the local existence of high-quality forage in the form of grasses enriched by nutrients derived from local Holocene tephras. This interpretation is reinforced by stable carbon and nitrogen isotope values obtained from the mammoth remains. The endpoint of mammoth survival in the Pribilofs is unknown, but maybe coterminous with the arrival of polar bears whose remains in the cave date to the Neoglacial cold period of ~ 4500 to 3500 ¹⁴C yr BP. The polar bear record corroborates a widespread cooling of the Bering Sea region at that time.     

Age constraints on Pleistocene megafauna at Tight Entrance Cave in southwestern Australia

A well-stratified succession of fossiliferous sediments occurs in Tight Entrance Cave, southwestern Australia. These infill deposits contain the remains of megafauna and have accumulated intermittently since the Middle Pleistocene: >137, 137–119 and 50–29 ka, according to the results of ¹⁴C, U–Th, ESR and OSL dating techniques. Megafaunal species richness was highest in the latest part of the penultimate glacial maximum and during the subsequent last interglacial (137–119 ka), but remains are less abundant following an apparent 70 ka depositional hiatus in the sequence. Most megafaunal specimens from the upper (<44 ka) units are fragmentary, and reworking from older strata cannot yet be ruled out. However, one specimen of Simosthenurus occidentalis (a large extinct kangaroo), a pair of articulated dentaries showing no signs of secondary transportation, was found within a sedimentary layer deposited between 48 and 50 ka. This represents one of the youngest demonstrably in situ occurrences of an Australian megafaunal taxon.     

Late Pleistocene glaciation of the Mt Giluwe volcano,Papua New Guinea

The Mt Giluwe shield volcano was the largest area glaciated in Papua New Guinea during the Pleistocene. Despite minimal cooling of the sea surface during the last glacial maximum, glaciers reached elevations as low as 3200 m. To investigate changes in the extent of ice through time we have re-mapped evidence for glaciation on the southwest flank of Mt Giluwe. We find that an ice cap has formed on the flanks of the mountain on at least three, and probably four, separate occasions. To constrain the ages of these glaciations we present 39 new cosmogenic ³⁶Cl exposure ages complemented by new radiocarbon dates. Direct dating of the moraines identifies that the maximum extent of glaciation on the mountain was not during the last glacial maximum as previously thought. In conjunction with existing potassium/argon and radiocarbon dating, we recognise four distinct glacial periods between 293–306 ka (Gogon Glaciation), 136–158 ka (Mengane Glaciation), centred at 62 ka (Komia Glaciation) and from >20.3–11.5 ka (Tongo Glaciation). The temperature difference relative to the present during the Tongo Glaciation is likely to be of the order of at least 5 °C which is a minimum difference for the previous glaciations. During the Tongo Glaciation, ice was briefly at its maximum for less than 1000 years, but stayed near maximum levels for nearly 4000 years, until about 15.4 ka. Over the next 4000 years there was more rapid retreat with ice free conditions by the early Holocene.     

Cosmogenic nuclide chronology of pre-last glacial maximum moraines at Lago Buenos Aires, 46°S, Argentina

At Lago Buenos Aires, Argentina, ¹⁰Be, ²⁶Al, and ⁴⁰Ar/³⁹Ar ages range from 190,000 to 109,000 yr for two moraines deposited prior to the last glaciation, 23,000–16,000 yr ago. Two approaches, maximum boulder ages assuming no erosion, and the average age of all boulders and an erosion rate of 1.4 mm/10³ yr, both yield a common estimate age of 150,000–140,000 yr for the two moraines. The erosion rate estimate derives from ¹⁰Be and ²⁶Al concentrations in old erratics, deposited on moraines that are >760,000 yr old on the basis of interbedded ⁴⁰Ar/³⁹Ar dated lavas. The new cosmogenic ages indicate that a major glaciation during marine oxygen isotope stage 6 occurred in the mid-latitude Andes. The next five youngest moraines correspond to stage 2. There is no preserved record of a glacial advance during stage 4. The distribution of dated boulders and their ages suggest that at least one major glaciation occurred between 760,000 and >200,000 yr ago. The mid-latitude Patagonian glacial record, which is well preserved because of low erosion rates, indicates that during the last two glacial cycles major glaciations in the southern Andes have been in phase with growth and decay of Northern Hemisphere ice sheets, especially at the 100,000 yr periodicity. Thus, glacial maxima are global in nature and are ultimately paced by small changes in Northern Hemisphere insolation.     

Late Pleistocene glaciations at Lake Donggi Cona,eastern Kunlun Shan (NE Tibet): early maxima and a diminishing trend of glaciation during the last glacial cycle

The Burhan Budai Shan in NE Tibet represents a key location for examining the variable influence of the mid‐latitude westerly and monsoonal circulations on late Quaternary glaciations in this sector of the Tibetan Plateau. Our study investigates the glacial history of mountains near Lake Donggi Cona (35°17′N, 98°33′E) using field mapping in combination with ¹⁰Be surface exposure dating and numerical reconstructions of former glacial equilibrium line altitudes (palaeo‐ELA). A set of 23 new exposure ages, collected from moraines in four glacial valleys, ranges from 45 to 190 ka, indicating ice expansion during the early and middle part of the last glacial cycle, and during the penultimate and possibly an earlier Mid‐Pleistocene glaciation. Ice advances reaching 12–15 km in length occurred at around 190–180 ka (≥MIS 6), between 140–100 ka (late MIS 6/MIS 5), and 90–65 ka (late MIS 5/early MIS 4), with a maximum ELA depression of 400–500 m below the estimated modern snowline. Exposure ages from the valley headwaters further indicate a small glaciation between c. 60–50 ka (late MIS 4/early MIS 3), which was essentially restricted to the cirque areas. Significantly, we find no evidence for any subsequent glaciation in the area during MIS 2 or the Holocene period. These results indicate a diminishing trend of glaciation in the region since at least MIS 4, and corroborate the case of a ‘missing LGM’ in the more interior parts of the northeastern Tibetan Plateau. The emerging pattern suggests that the most favourable conditions for glaciation during the Late Pleistocene correspond to periods of relatively moderate cooling combined with an intermediate or rising East Asian monsoon strength.     

Late Pleistocene glaciers in Darhad Basin, northern Mongolia

Late Pleistocene glaciers around Darhad Basin advanced to near their maximum positions at least three times, twice during the Zyrianka glaciation (at ∼ 17-19 ka and ∼ 35-53 ka), and at least once earlier. The Zyrianka glaciers were smaller than their predecessors, but the equilibrium-line altitude (ELA) difference was < 75 m. End moraines of the Zyrianka glaciers were ∼ 1600 m asl; ELAs were 2100-2400 m asl. ¹⁴C and luminescence dating of lake sediments confirm the existence of paleolake highstands in Darhad Basin before ∼ 35 ka. Geologic evidence and ¹⁰Be cosmic-ray exposure dating of drift suggests that at ∼ 17-19 ka the basin was filled at least briefly by a glacier-dammed lake ∼ 140 m deep. However, lake sediments from that time have not yet been recognized in the region. A shallower paleolake briefly occupied the basin at ∼ 11 ka, but between ∼ 11 and 17 ka and after ∼ 10 ka the basin was probably largely dry. The timing of maximum glacier advances in Darhad appears to be approximately synchronous across northern Mongolia, but different from Siberia and western Central Asia, supporting the inference that paleoclimate in Central Asia differed among regions.     

Glacial and environmental changes over the last 60 000 years in the Polar Ural Mountains,Arctic Russia,inferred from a high‐resolution lake record and other observations from adjacent areas

Our knowledge about the glaciation history in the Russian Arctic has to a large extent been based on geomorphological mapping supplemented by studies of short stratigraphical sequences found in exposed sections. Here we present new geochronological data from the Polar Ural Mountains along with a high‐resolution sediment record from Bolshoye Shchuchye, the largest and deepest lake in the mountain range. Seismic profiles show that the lake contains a 160‐m‐thick sequence of unconsolidated lacustrine sediments. A well‐dated 24‐m‐long core from the southern end of the lake spans the last 24 cal. ka. From downward extrapolation of sedimentation rates we estimate that sedimentation started about 50–60 ka ago, most likely just after a large glacier had eroded older sediments from the basin. Terrestrial cosmogenic nuclide (TCN) exposure dating (¹⁰Be) of boulders and Optically Stimulated Luminescence (OSL) dating of sediments indicate that this part of the Ural Mountains was last covered by a coherent ice‐field complex during Marine Isotope Stage (MIS) 4. A regrowth of the glaciers took place during a late stage of MIS 3, but the central valleys remained ice free until the present. The presence of small‐ and medium‐sized glaciers during MIS 2 is reflected by a sequence of glacial varves and a high sedimentation rate in the lake basin and likewise from ¹⁰Be dating of glacial boulders. The maximum extent of the mountain glaciers during MIS 2 was attained prior to 24 cal. ka BP. Some small present‐day glaciers, which are now disappearing completely due to climate warming, were only slightly larger during the Last Glacial Maximum (LGM) as compared to AD 1953. A marked decrease in sedimentation rate around 18–17 cal. ka BP indicates that the glaciers then became smaller and probably disappeared altogether around 15–14 cal. ka BP.     

Landslide deposits as stratigraphical markers for a sequence‐based glacial stratigraphy: a case study of a Younger Dryas system in the Eastern Alps

Detailed ¹⁰Be and ¹⁴C dating and supporting pollen analysis of Alpine Lateglacial glacial and landslide deposits in the Hohen Tauern Mountains (Austria) constrain a sequence‐based stratigraphy comprising a major landslide (13.0±1.1 ka) overlain by till and termino‐lateral moraines of an advancing (12.6±1.0 ka) and retreating (11.3±0.8 ka) glacier in turn overlain by a minor landslide (10.8±1.1 ka). These results define glacier activity during the Younger Dryas age Egesen stadial bracketed by landslide activities during the Bølling‐Allerød interstadial and the Preboreal. In contrast to recent studies on Holocene glaciation in the Alps, no traces of any Holocene glacier advance bigger than during the Little Ice Age are documented. Furthermore, this study demonstrates the advantages of using an allostratigraphical approach based on unconformity‐bounded sedimentary units as a tool for glacial stratigraphy in formerly glaciated mountain regions, rather than a stratigraphy based on either isolated morphological features or lithostratigraphical characteristics.     

High-time resolution AMS C data sets for Lake Baikal and Lake Hovsgol sediment cores: Changes in radiocarbon age and sedimentation rates during the transition from the last glacial to the Holocene

High-time resolution ¹⁴C dating of Lake Baikal sediment cores indicates negative and positive anomalies of calculated linear sedimentation rate (LSR; 1.1 and 35.6 cm/ka, respectively) during the period of climate transition from the last glacial to Holocene. The timing of the Lake Baikal apparent LSR anomalies is consistent with that of the changes in the atmospheric radiocarbon concentration (Δ¹⁴C) during Younger Dryas rapid cooling event. ¹⁴C dating of lipids in the Lake Baikal surface sediments revealed that the sources of sedimentary lipids were different in each basin. In the Northern Basin of Lake Baikal, the ¹⁴C age of total lipids from the surface sediment (4.0 ¹⁴C ka) was found to be older than that of TOC (1.6 ¹⁴C ka). By contrast, the ¹⁴C age of total lipids in the Southern Basin was younger than that of the TOC by ca. 0.7–3.0 ka.In the Lake Hovsgol sediment cores, ages of the main lithologic boundaries during the last glacial–interglacial transition were estimated based on new ¹⁴C data sets. TOC concentration in the cores started to rapidly increase at 13.8 ± 0.3 ¹⁴C ka at the base of the basinwide finely laminated layer deposited during Bølling/Allerød. The base of the layer diatomaceous mud corresponds to the end of Younger Dryas event (10.6 ± 0.1 ¹⁴C ka).     

Revised radiocarbon ages on woolly rhinoceros (Coelodonta antiquitatis) from western central Scotland: significance for timing the extinction of woolly rhinoceros in Britain and the onset of the LGM in central Scotland

Woolly rhinoceros bones, from a number of sites in Britain, have been AMS radiocarbon dated following ultrafiltration pre-treatment. These determinations give a coherent set of ages between >50 and c. 35 cal ka BP. The youngest (35,864–34,765 cal BP) come from the area around Bishopbriggs in western central Scotland and are derived from glaciofluvial sand and gravel overlain by till, both deposited during the Last Glacial Maximum (LGM) glaciation. A previous radiocarbon date from the site suggested that woolly rhinoceros lived c. 27 ¹⁴C ka BP and the region was ice-free at the time. This date has had significant influence on the timing of extinction of woolly rhinoceros and the onset of glaciation over Britain during the LGM. The new dates revise this earlier determination and confirm that woolly rhinoceros became extinct in Britain after c. 35 cal ka BP, that central Scotland was ice-free at this time, and glaciation extended across this region sometime after 35 cal ka BP.     

Palaeoglaciology of Bayan Har Shan,NE Tibetan Plateau: exposure ages reveal a missing LGM expansion

The Bayan Har Shan, a prominent upland area in the northeastern sector of the Tibetan Plateau, hosts an extensive glacial geological record. To reconstruct its palaeoglaciology we have determined ¹⁰Be exposure ages based on 67 samples from boulders, surface pebbles, and sediment sections in conjunction with studies of the glacial geology (remote sensing and field studies) and numerical glacier modelling. Exposure ages from moraines and glacial sediments in Bayan Har Shan range from 3 ka to 129 ka, with a large disparity in exposure ages for individual sites and within the recognised four morphostratigraphical groups. The exposure age disparity cannot be explained by differences in inheritance without using unrealistic assumptions but it can be explained by differences in post-depositional shielding which produces exposure ages younger than the deglaciation age. We present a palaeoglaciological time-slice reconstruction in which the most restricted glaciation, with glaciers less than 10 km long, occurred before 40–65 ka. More extensive glaciations occurred before 60–100 ka and 95–165 ka. Maximum glaciation is poorly constrained but probably even older. The Bayan Har Shan exposure age dataset indicates that glaciers on the northeastern Tibetan Plateau have remained surprisingly restricted for at least 40 ka, including the global last glacial maximum (LGM). This case of a missing LGM is further supported by high-resolution glacier modelling experiments.     

Timing and extent of Holocene glaciations in the monsoon dominated Dunagiri valley (Bangni glacier), Central Himalaya,India

Field stratigraphy and optical and radiocarbon dating of lateral moraines in the monsoon dominated Dunagiri valley of the Central Himalaya provide evidence for three major glaciations during the last 12 ka. The oldest and most extensive glaciation, the Bangni Glacial Stage-I (BGS-I), is dated between 12 and 9 ka, followed by the BGS-II glaciation (7.5 and 4.5 ka) and the BGS-III glaciation (∼1 ka). In addition, discrete moraine mounds proximal to the present day glacier snout are attributed to the Little Ice Age (LIA). BGS-I started around the Younger Dryas (YD) cooling event and persisted till the early Holocene when the Indian Summer Monsoon (ISM) strengthened. The less extensive BGS-II glaciation, which occurred during the early to mid-Holocene, is ascribed to lower temperature and decreased precipitation. Further reduction in ice volume during BGS-III is attributed to a late Holocene warm and moist climate. Although the glaciers respond to a combination of temperature and precipitation changes, in the Dunagiri valley decreased temperature seems to be the major driver of glaciations during the Holocene.     

Climatic variability during the last 90 ka of the southern and northern Levantine Basin as evident from marine records and speleothems

The influence of the northern Atlantic and tropical monsoonal systems, as recorded by the River Nile, on the climate variability of the southeastern Mediterranean was studied in two cores taken by the R/V Marion Dufresne: one core taken SE of Cyprus representing the northern Levantine Basin (core 9501, 980 m water depth) and the other located ˜380 km further south, represents the southeastern Levantine Basin in an area influenced by the River Nile plume (core 9509, 884 m water depth). The study was performed at relatively high resolution using several proxies: δ¹⁸O of Globigerinoides ruber, sediment characteristics and index colour parameters in core sections representing the last 86 ka. A low-resolution alkenone sea surface temperature record was also measured. The time frame in both cores was mostly constrained by ‘wiggle’ matching with the nearby well-dated δ¹⁸O and δ¹³C record of the Soreq Cave, which is mainly influenced by the eastern Mediterranean water vapor. The sedimentary record of the southern core is strongly influenced by the River Nile contribution throughout the last 86 ka, as evidenced by the higher sedimentation rates compared with the northern core (20 cm/ka vs. 5 cm/ka), continuously darker sediment colour, and higher TOC values (0.6–0.9 vs. 0.25 wt% not including sapropels). During sapropels S1 and S3, present in both cores, the influence of the River Nile became more widespread, reaching as far as Cyprus. Yet, the influence of the River Nile remained stronger in the south, as evident by the higher TOC values in the southern core throughout the entire 90 ka period and the longer duration of S1 in the southern core. An anomalous low δ¹⁸O interval that is not recorded in western Mediterranean occurred between 58 and 49 ka in the Levantine Basin and is more developed in the northern core. This period correlates with D-O interstadial 14 and maximum northern hemisphere insolation during the lastglacial cycle, suggesting that the warming mainly impacted the northern Levant.The Eastern Mediterranean Sea and land area was considerably warmer than the western Mediterranean throughout the LGM – Holocene transition, and the δ¹⁸O_{G. ruber} drop of 4.5‰ is significantly greater than the 3‰ shift found for the western Mediterranean δ¹⁸O_{G. bulloides}, both differences reflecting an increased continental effect from the western to eastern Mediterranean. Comparison between the marine and the land δ¹⁸O records suggests that the origin of rain over the land is composed of mixed signal from the southern and northern Levantine Basin. The study of Δδ¹⁸O_sea–land variations demonstrates that various factors have influenced the sea–land relationship during the last 90 ka. The ‘amount effect’ has an important influence on rainfall δ¹⁸O during interglacial periods (particularly sapropel periods), whereas during glacial periods, increased land distances and elevation differences arising from decrease in sea level may have brought about decrease in δ¹⁸O of rainfall due to Rayleigh distillation processes. These influences were superimposed on those of sea surface water δ¹⁸O changes brought about by continental ice melting, and the strong effects felt in the southern Levantine Basin of the high River Nile input during periods of enhanced monsoonal activity.