Quaternary paleoclimate of the north-eastern boundary of the Saharan Desert: reconstruction from speleothems of Negev Desert, Israel Vaks Anton This work was submitted for the degree "Doctor of Philosophy" to the Senate of the Hebrew University, Jerusalem. The study was carried out under the supervision of: Dr. Miryam Bar-Matthews, Geological Survey of Israel Prof. Alan Matthews, Institute of Earth Sciences, Hebrew University of Jerusalem Prof. Amos Frumkin, Department of Geography, Hebrew University of Jerusalem Report GSI /14/08 Jerusalem, June 2008
Acknowledgements: I would like to thank everyone who assisted me to bring this work to its successful completion. I had an opportunity to learn from excellent advisors: Dr. Miryam Bar-Matthews, Prof. Alan Matthews and Prof. Amos Frumkin, and I thank them very much for guidance, advising, fruitful discussions, encouragement, being ready to help every time, even during personal vacations, or in 90-th moment before the deadlines. I would like to very much thank Dr. Avner Ayalon for guidance and advice, and for his generous help with the stable isotope analyses. I would like to thank: Dr. Bettina Shilman and Efrat Eliani for help with performing the stable isotope analyses; Natalya Tepliakov, Dr. Ludwik Halicz, Dr. Irena Segal, Dr. Olga Yoffe, Dr. Sarah Ehrlich, Dina Stiber and Galit Sharabi for guidance and help with chemical and mass spectrometric analyses; to Dr. Ahuva Almogi-Labin for fruitful discussions and help with understanding of the marine records, and students Liraz Laor, Ruthi Kiro, Sharona Shlomi, Neta Shalev, Asaf Wunsch, Gilad Garber and Yael Neumieir for the assistance in the laboratory and in the field. I would like to thank to the previous Director of the Geological Survey of Israel Dr. Amos Bein and the present Director, Dr. Benjamin-Ze'ev Begin for help and encouragement, to Onn Crouvi, Dr. Yoav Avni, Dr. Rivka Amit, Dr. Ezra Zilberman, Dr. Naomi Porat and Dr. Rani Calvo for help with understanding soil and sedimentary records of the Negev climate, to Michael Dvoracheck for guidance and help with SEM analyses, to Dr. Amir Sandler, Yoetz Deutsch, Uri Simchai and Shlomo Leyzenbah for help with XRD analyses, and to Michael Kitin for help with sieving the soil samples. I would like to thank Dr. Henry Foner for help with understanding of the dust fall patterns in the region. Many thanks to Prof. Jon Woodhead from University of Melbourne, Australia, and to Prof. Robert Cliff from University of Leeds, United Kingdom, for U-Pb dating of speleothems which are older than U-Th dating limit. I would especially like to thank Shlomo Ashkenazi for his great help in the field and laboratory and especially for being good friend, to Yehuda Peled, Eli Ram, Yaakov Mizrachi, Ariel Gai, Robert Knafo, Yaakov Rafael for the help in field work,
to Moshe Peri for preparation of the thin sections, to Camille Lousqui and Ariel Belmas for help with purchasing an equipment, to Channa Netzer-Cohen, Bat-Sheva Cohen, and Nili Almog for help with graphics and publications, to vice manager of Geological Survey of Israel Shmuel Zitronblat, to Ira Peer, Anna Feigin, Masha Kahalani, Valery Doudin, David Sidi, Oksana Zaripov and to other people of Geological Survey of Israel that was for me a good home during many years. Special thanks to Dr. Mordechai Stein and Prof. Yehouda Enzel for fruitful discussions. Although we had disagreements (and maybe because of them), these discussions contributed a lot to improve the quality of this work. I would like to thank Prof. Uri Dayan and Prof. Baruch Ziv for help with understanding of meteorological systems in the region. Many thanks to Dr. Lior Grossman, Prof. Erella Hovers, Dr. Rivka Rabinovich, Prof. Na'ama Goren-Inbar from the Institute of Archeology of the Hebrew University for help with understanding the archaeological issues, to Dafna Kadosh, Nicolas Waldman, Adi Torfshtein, and Itai Haviv for fruitful discussions, to Yael Yakobi, Surin Lisker, Roi Porat, Azriel Raikin and other members of Israeli Cave Research center for the assistance in the field work. Many thanks to my parents Evgenya Vaks-Reznik and Evgeny Vaks from Be'erSheva, Nadav Bloch, Smadar Bloch-Wiener and Zvi Wiener from Arad, Rakefet Dar from Neot-Smadar and Dr. Georgy Shenbrot from Ramon Science center for the assistance with the rain sampling. Many thanks to the secretaries of the Institute of Earth Sciences, Hebrew University, Maggi Parkin, Carmela Lev, Batia Moshe, Mali Shisha-Halevi. I would like to thank to anonymous reviewers whose remarks helped to increase the quality of this work. Thanks a lot to my friends Dan Asael, Boriana Kalderon, Liraz Laor and Dina Vachtman for their support, help and friendly good word of encouragement. For my dear parents Evgenya Vaks-Reznik and Evgeny Vaks, and my brother David Vaks many thanks for their love and support. I would like to thank and apologize to all people that I maybe inadvertently omitted from these acknowledgements.
IAbstract Carbonate cave deposits (speleothems) from caves of the Israeli deserts were used to reconstruct the paleoclimate conditi ons of the northern boundary of SaharanArabian Desert. The deserts studied in this wo rk are of two types: the Negev Desert in southern Israel, which is part of the Saharan-Arabian desert belt; and the "rain shadow" desert located east of Central Mountain Ridge (CMR ) in the rift valleys of the Jordan River and Dead-Sea (Judean De sert). The presence of speleothems in numerous caves in these present-day arid regions indicates that humid climatic conditions (i.e., periods with positive eff ective precipitation/infiltration index) occurred in the past. In this study, the timing of th e speleothem growth, as an indicator of periods of increased effective precipitation, was obtained by pr ecise U-Th dating; the origin of rainfall, its amounts and the types of vegetation were examined from studies of the speleothem 18O and 13C values and the D values of their fluid inclusions, and the sources of dust were studied through the Sr isotopic composition of the speleothems. The minimum precipitation amounts necessary to deposit speleothems are estimated to be: 200-275 mm/year during glacial periods and 300-350 mm/year during interglacial periods. Speleothem deposition in the Jordan Valley mainly occurred during the three last glacial periods , with minor deposition occurring during Termination II (~135 ka) and MIS-7 interglaci al (225-205 ka), and no speleothem deposition taking place during the Holocene, the peak of last glacial maximum (~19 ka), and for most of the previous inte rglacial MIS-5 (130-75 ka). The 18O and 13C profiles of speleothems deposit ed between 67 ka and 25 ka in the Jordan Valley match the general isotopic trends of previously studied speleothems from central and northern Israel; suggesting a similar Easter n Mediterranean (EM) Sea source for the precipitation and similar climatic conditions . Decrease in temperature and evaporation, and the consequent increase in effective precipitation, were probably the major factors controlling the decay of the "r ain shadow" effect during glacial periods.
II In caves of the northern Negev located on the present-day 150-160 mm isohyet, most of the speleothem deposition occurred during glacial periods (as in the Jordan Valley), and only restricted, episodic deposition occurred during the interglacial periods and glacial maxima. No speleoth em deposition occurred at 117-96 ka and 9284 ka (MIS-5(3-2)) and during the Holocene, indicating that the climate in these periods was similar to present or even more arid. The 18O values of the speleothems mostly show the same trends as speleoth ems from central and northern Israel; thus pointing to the EM Sea as the precipitation source. This source is also indicated by the D18O relations derived from fluid inclusions. Speleothem 13C values in the northern Negev show that the vegetation was usually C3+C4 steppe, with rare invasions of Mediterranean C3 species during the most humid events, or that the area experienced episodes of desertification. Generally the conditions were drier than in the Jordan Valley. The maximum southwar d shift of the boundary between the Mediterranean climate and the semi-deser t zone during humid periods was 20-25 km. Most speleothem deposition in the Judean Desert also occurred during the three last glacial periods, with minor deposition during the two last inte rglacials. Climate conditions there were generally drier than in the northern Negev and Jordan Valley. Major humid periods in the central and southern Negev Desert (Negev Humid Periods â€“ NHP) during the last 350 ka o ccurred at: 350-310 ka (NHP-4), 310-290 ka (NHP-3), 220-190 ka (NHP-2), and 142-109 ka (NHP-1). NHP-4 was most intense in the southern Negev and probably occurred as a result of rainfall of tropical origin, whereas NHP-3, 2 and 1 were dominant in the central Negev, and only NHP-1 reached the southern Negev. Only NHP-3 was associated with a glacial period (MIS8). The rainfall episodes during NHP were short-lived, as evident from very thin speleothem laminae and the very minor de velopment of semi-desert C4/steppe C3+C4 indicated by 13C values. Precipitation during NHP 3, 2 and 1 originated in the EM Sea (Atlantic-Mediterranean Cyclones), as evident from the D18O relations in the fluid inclusions and the systematic decrea se in speleothem numbers from north to south. The 18O values of all Negev speleothems are, however, systematically lower
III than in contemporaneous speleothems from the Jordan Valley, central and northern Israel. This in part is attributed to th e increased rainout of the heavy isotopes by Rayleigh fractionation processes. The dominant speleothem deposition during interglacial periods in the central and southern Negev contrasts with the northern Negev, Jordan Valley and Judean Deserts, where most deposition occurred during glac ial periods. Glacial speleothem deposition in the latter areas is associated with hi gh infiltration coeffici ents during the cold glacial periods. During humid interglacial episodes the intensity of the Cyprus cyclones was higher than during the glacial s, bringing precipitation high as 300-350 mm as far south as the pres ent-day ~50 mm isohyet. Ho wever, higher temperatures and lower effective precipitation only allowed limited water infiltration to the caves and the deposition of thin speleothem laminae. The major decreases of speleothem 18O in Soreq Cave at 132 ka and 127-126 ka and in Peqiâ€™in Cave at 200 ka occurred simultaneously with the highest peaks of speleothem deposition in the Negev Desert during the NHP-1 and NHP-2. These events were probably associated with mo st intensive humid episodes above the EM Sea. NHP events were contemporaneous with periods where the monsoon index 51 cal/cm2 day and sapropel formation occurred in the Mediterranean Sea. Simultaneous intensification of the mons oon and Atlantic-Mediterranean cyclones probably relates to weakening of the high pre ssure cell above th e sub-tropical Atla ntic Ocean. Absence of speleothem deposition in the Negev dur ing the Holocene and MIS-11 (430-400 ka) interglacials could be explained by lower Northern Hemisphere insolation and lower African Monsoon index than during the NHP s. Thus, the high pressure cell above subtropical Atlantic Ocean was probably stronger, leading to more northern trajectories of Atlantic-Mediterranean cy clones and preventing from Mediterranean precipitation to reach the Negev.
IV Speleothem 87Sr/86Sr ratios in the â€œrain shadowâ€ and the Negev deserts mostly vary between 0.7082 and 0.7085, and show only slight change s between glacial and interglacial periods. Compared to the wider variation observed in central Israel, this range probably reflects lower host rock di ssolution rates, or higher dust supply from the proximal Sinai-Nile Delta source characterized by 87Sr/86Sr ratios of ~ 0.7084. Most speleothems of the central and southe rn Negev Desert are older than the limit of uranium series dating ( ca 550 ka). Preliminary U-Pb studies give ages of 3.3-2.7 Ma, with a few speleothems deposited at ~1.3 Ma. These ages correspond to periods when lakes were known to exist in the region. A reduction in the rate of speleothem deposition, together with an increase in speleothem 87Sr/86Sr ratios after ~2.7 Ma, probably reflects increased de sertification of the region and of the entire SaharanArabian Desert, which increased the dust supply from Sinai and Saharan-Arabian sources. Multiple growth / hiatus cycles in a single speleothem over >3 Ma show that in arid environment the fractures conducting wa ter from the surface to the caves remain open over long times. The Sinai-Negev land bridge was the ma jor, and possibly the only land-bridge, connecting the Africa with Asia since the Miocene, and the Saharan-Arabian Desert was a significant obstacle for hominid and anim al migration from Africa to other parts of the world. The occurrence of humi d phases in the Sinai-Negev deserts contemporaneous with humid phases in th e southern and central Saharan-Arabian Desert probably opened climatic â€œwindowsâ€ for migration of hominids and animals out of the African continent. Migration of the early modern humans from Africa to the Levant was probably associat ed with NHP-1 at 142-109 ka.
Table of contents 1 1. Introduction 3 2. Geography, climate and geology of the Israeli deserts 3 2.1 Geographical definition and climatic zoning 5 2.2. Rainfall gradients in the northern Negev Desert and Jordan/Dead Sea Valleys 6 2.3 Geology 8 2.4 Speleothems as paleoclimatic records 11 2.5. Speleothem climate record from the Mediterranean climate zone of Israel 14 3. Research goals 14 3.1 The major aims of the study 15 3.2 Structure of the thesis 17 4. Methodology 17 4.1 The studied caves and sampling 18 4.2 U-Th and U-Pb dating 20 4.3 Stable isotope analyses of speleothems and present day rainwater 22 4.4 D/H analyses of speleothem fluid inclusions 23 4.5 Sr isotope analyses 25 5. Vaks, A., Bar-Matthews, Ayalon., A., Schilman, B., M., Gilmour, M., Hawkesworth, C. J., Frumkin, A., Kaufman, A., and Matthews, A. (2003), Paleoclimate reconstruction based on the timing of speleothem growth, oxygen and carbon isotope composition from a cave located in the â€œrain shadowâ€ in Israel., Quaternary Research , 59, 2, pp 182-193. 37 6. Vaks , A., Bar-Matthews, M., Ayalon, A., Matthews, A., Frumkin, A., Dayan, U., Halicz, L., Almogi-Labin, A. and Schilman, B. (2006), Paleoclimate and location of the border between Mediterranean climate region and the Saharo-Arabian Desert as revealed by speleothems from the northern Negev Desert, Israel., Earth and Planetary Science Letters , 249, 3-4, 384-399. 53 7. Vaks, A., Bar-Matthews, M., Ayalon, A., Matthews, A., Halicz, L. and Frumkin, A., (2007), Desert speleothems reveal climatic window for African exodus of early modern humans , Geology , 35, 9, 831-834. 57 8. Paleoclimate of northern Saharan-Arabian Desert from speleothem record in central and southern Negev and Judea Desert, Israel 57 8.1 Introduction 57 8.2 Results
57 8.2.1 Field and petrographic characterization of the speleothems and the cave sediments 57 184.108.40.206 Vadose speleothems and preliminary U-Pb dating results 60 220.127.116.11 Phreatic speleothems 61 18.104.22.168 Cave sediments and paleosols washed into the caves 62 8.2.2 Uranium concentrations of central and southern Negev speleothems 63 8.2.3 U-Th ages and depositional periods 63 22.214.171.124 Kanaim Cave, Judea Desert and the modified record of Ma'ale-Efrayim Cave, Jordan Valley 63 126.96.36.199 Caves from the central and southern Negev Desert 65 8.2.4 Oxygen and carbon isotopic compositions of the central and southern Negev speleothems 66 188.8.131.52 Oxygen isotopic compositions 68 184.108.40.206 Carbon isotopic compositions 69 8.2.5 Present-day rainfall in the Negev Desert: amounts and hydrogen and oxygen isotopic compositions 70 8.2.6 D values of speleothem fluid inclusions from northern, central and southern Negev 71 8.2.7 Sr concentrations and 87Sr/86Sr ratios in speleothems, cave host rocks and soils for the Jordan Valley and the Negev Desert 73 8.3 Discussion 73 8.3.1 Humid episodes in central /southern Negev, and Judea deserts 73 220.127.116.11 Speleothem deposition as recorder of effective precipitation 76 18.104.22.168 Humid periods in central and southern Negev Desert 79 22.214.171.124 Humid periods in Judea Desert 79 8.3.2 Origin of precipitation in the Negev Desert under present-day conditions and during the past 550 ka 79 126.96.36.199 Present-day rainfall 81 188.8.131.52 Origin of precipitation in the past 85 8.3.3 Vegetation in the central and southern Negev Desert determined from the 13C of the speleothems and its relation to regional climatic change 88 8.3.4 Correlation between the Negev Humid Periods and the regional and global climate 90 8.3.5 "North â€“ south paradox" of the Negev paleoclimate 93 8.3.6 Pliocene-Pleistocene humid periods in the central and southern Negev Desert as evident from speleothems older than 550 ka
95 8.3.7 Phreatic speleothems in Makhtesh-ha-Qatan as markers of paleogroundwater levels 95 8.3.8 â€“ Humid periods in the northern Saharan-Arabian Desert as climatic â€œwindowsâ€ for Out of Africa dispersals of hominids and animals 97 8.3.9 Sr isotope ratios from Israeli desert speleothems as tracers of the origin of dust and host rock weathering 101 9. Summary and conclusions 107 Bibliography Appendix 1: Figures 119 Figure 1: Geographical location of the study area , annual precipitation and sample sites 121 Figure 2A: Geological stratigraphic section of the research area 122 Figure 2B: Geologic map of the research area. 123 Figure 3: Typical cross sections of the Ashalim Cave speleothems as observed in the field, hand specimens and under SEM 124 Figure 4: Cross section of the flowstone KTO(1)-1 from Ktora Cracks 125 Figure 5: The mineralogy and petrography of the calcite and fine laminae representing hiatuses within the Young member of the stalactite ASH-11 126 Figure 6: Flowstone MMR-7(2) from Maâ€™ale-ha-Meyshar Cave 127 Figure 7: SEM analysis of the minerals 128 Figure 8: Petrography of stalactite KN-8 from the Kanaim Cave 129 Figure 9: Phreatic speleothems in Maâ€™ale-ha-Meyshar Cave 130 Figure 10: The relative frequencies of the speleothem ages in â€œrain shadowâ€ desert 131 Figure 11: Dating of central and southern Negev speleothems 132 Figure 12: Plots of the relative age frequencies 133 Figure 13: Sample ASH-33 from the Ashalim Cave 134 Figure 14: 18O values of central and southern Negev speleothems during the last ~3 Ma 135 Figure 15: Oxygen isotope profiles of central and southern Negev speleothems during the last 350 ka 136 Figure 16: Oxygen isotope profiles in central Negev during the two major periods of speleothem deposition 137 Figure 17: 13C values of central and southern Negev speleothems during the last ~3 Ma 138 Figure 18: Carbon isotope profiles of central and southern Negev speleothems during the last 350 ka. 139 Figure 19: Carbon isotope profiles in central Negev during the two major periods of speleothem deposition. 140 Figure 20: Northern Negev rainfall isotopic composition 141
Figure 21: Central and southern Negev rainfall isotopic composition 142 Figure 22: D values of Negev Desert speleothem fluid inclusions 143 Figure 23: Sr concentrations in the Negev and Jordan Valley speleothems 144 Figure 24: 87Sr/86Sr isotopic ratios of Maâ€™ale-Efrayim Cave speleothems 145 Figure 25: 87Sr/86Sr ratios of the Negev Desert vadose speleothems 146 Figure 26: 87Sr/86Sr ratios in Negev Desert vadose speleothems compared with 87Sr/86Sr ratios of dolomitic host rock, of bulk soil above the caves and of its silicate fraction 147 Figure 27: Hydrogen and oxygen isotopic composition of paleo-cave waters based of speleothem fluid inclusion analysis 148 Figure 28: d-excess values from speleothem fluid inclusions in central and southern Negev Desert during the last 220 ka 149 Figure 29: Thinning of the speleothem sequences from the north to the south 150 Figure 29 (A-G): Width of the NHP-1 and NHP-2 sequences in the Negev speleothems 150 Figure 29 (H, I) Thinning of the speleothem laminae from the north to the south in the Negev Desert during the NHP-1 and during the NHP-2 151 Figure 29J: Shortening of the speleothem deposition periods from north to south. 152 Figure 30: Comparison of the 18O profiles and age frequencies of central Negev speleothems with 18O values of Tzavoa, Soreq and Peqiâ€™in speleothems 153 Figure 31: Comparison of the 13C profiles and age frequencies of central Negev speleothems with 13C values of Tzavoa, Soreq and Peqiâ€™in speleothems. 154 Figure 32: Correlation between the peaks of relative frequencies of the speleothem ages during the last 550 ka and the solar radiation energy 155 Figure 33: Relative frequencies of the speleothem ages compared with African monsoon index and timing of Mediterranean sapropels 156 Figure 34: Correlation between the relative frequencies of the speleothem ages during the NHP-1 and NHP-2 and the: 18O profiles of Soreq and Peqiâ€™in Caves, monsoon index, parameters of solar radiation, and timing of Mediterranean sapropels 157 Figure 35: Correlation between the relative frequencies of the speleothem ages and sapropel events during the period between 550 ka and 250 ka with African monsoon index values and solar radiation 158 Figure 36: Location of the caves where Sr isotopic composition of speleothems was studied 159 Figure 37: 87Sr/86Sr ratios of Maâ€™ale-Efrayim Cave speleothems, Jordan Valley and Negev Desert speleothems compared with those of the speleothems of Soreq and Jerusalem caves 160 Figure 38: Comparison between Sr concentrations in central Israel caves (Soreq and Jerusalem) and caves of northern Negev (Tzavoa and Maâ€™aleDragot), Jordan Valley (Maâ€™ale-Efrayim), and central and southern Negev 161 Figure 39: 87Sr/86Sr vs 1/Sr plot in speleothems, host rocks and bulk soils 162
Figure 40: Annual dust fall (gr/m2) map of Ganor and Foner (1996, 2001) and the location of the studied caves and the caves of central Israel 163 Appendix 2: The studied caves 164 Appendix 3: Analytical procedures of U-Th and U-Pb dating methods 169 A.3.1. U-Th dating: chemical separation between U and Th 169 A.3.2. U-Th dating: measurement of U and Th isotopic ratios by MC-ICP-MS 170 A.3.3 U-Pb dating method 171 A3.3.1 Chemical purification and extraction of U and Pb 171 A3.3.2 Measurement using MC-ICP-MS 172 A3.3.3 Isochrone formalization 173 A3.3.4 Correction for U series disequilibrium 173 References of Appendix 3 174 Appendix 4: Tables of results 177 Table 1: U-Pb and U-Th ages of speleothems 178 Table 2: D and 18O values of present-day rainfall in the Negev Desert 183 Table 3: D values and calculated 18O values of the Negev Desert speleothem fluid inclusions 186 Table 4: Table 4: Sr concentrations and 87Sr/86Sr ratios: speleothems, host rocks, bulk soils and their silicate fractions in the Jordan Valley and the Negev Desert. 188
1. Introduction The Saharan-Arabian Desert (Fig 1A)1 is the largest and the most arid desert in the world, but its history is punctuated by many humid periods and the formation of this desert is connected with the glaciation of the northern hemisphere at about 2.7 Ma were (deMenocal, 1995; 2004). Numerous studies have shown that the paleoclimate of southern and central part of Saharan-Arabian Desert is strongly related to global climate changes (Street and Grove, 1979; Miller et al ., 1991; Szabo et al., 1995; Crombie et al., 1997; Rohling et al., 2002; Fleitmann et al., 2003; Almogi-Labin et al., 2004; Osmond and Dabous, 2004; Drake and Bristow, 2006). Middle-Late Quaternary monsoon activity increased during periods of highest northern hemisphere insolation, mainly associated with precession (19-23 ka) and obliquity (42 ka) cycles of Earth's orbit, bringing about wet conditions and the formation of mega-lak es and rivers in the southern and middle parts of present day hyper-arid Saharan-Arabian Desert. Derr icourt (2005) suggested that humid phases in the Saharan Arabian Desert may have controlled the dispersal of hominids from the tropical Africa to the othe rs parts of the world, via the Sinai-Negev land bridge. The northern part of the Saharan-Arabian Dese rt mainly receives its precipitation from mid-latitude cyclones, but the data on its pale oclimate and the relationship with the global climate change is still fragmented and controversial. Evidence from Tunisia shows that rainfall increased during Holocene African Hu mid period, resulting in the formation of large lakes (Causse et al., 2003). Eviden ce for wetter conditions during the early Holocene in the eastern parts of Saha ro-Arabian Desert was given by the 13C values of the organic matter of the land snails in th e northern Negev Desert, southern Israel (Magaritz and Goodfriend, 1987; Goodfriend, 1990; 1991; 1999). Evidence for clustering of flood events during glacial and late Holocene is given by the presence of tufa and slack water deposits in caves near the wadi channels (Kuperman, 2005; Greenbaum et al., 1 All figures in this work are given in Appendix 1
2006). However, no evidence of Holocene humi dity has been found in the deserts of northern Egypt or the southern Jordan (Petit -Maire et al., 2002; Brook et al., 2002). Older humid periods at 8-15 ka, 25-60 ka, ~75-145 ka, ~150 ka, ~170-250 ka and 280 ka, have been interpreted from studies on: soil carbonate nodules (Goodf riend and Magaritz, 1988), spring tufa (Schwarcz et al., 1979; Livnat and Kronfeld, 1985; 1990; Enmar, 1999), soils on alluvial terraces (Plakht, 1995; 2000), lacustrine sediments from southern Israel (Avni, 1997; Avni et al., 2001-a; 2001-b; Ginat, 1997; Ginat et al., 2003) and southern Jordan (Petit-Maire et al., 2002) and speleothems from Djara Cave in Egypt (Brook et al., 2002). However, the dating precis ion of these humid periods is far from being accurate. In the present work, carbonate cave deposits (speleothems) from caves of Israeli deserts were used to recons truct the paleoclimate conditions of the northern boundary of Saharan-Arabian Desert. The Israeli deserts can be classifi ed in two types: the Negev Desert in southern Israel, which is part of the Saharan-Arabian desert belt; and the "rain shadow" desert of eastern Is rael, located to the east of Central Mountain Ridge (CMR) along the Dead-Sea Transform in the rift valleys of the Jord an River and Dead-Sea, (Fig. 1B). These regions are ideal for paleoclimate research because of: a) the very sharp precipitation gradients between the Med iterranean climate zone (>350 mm of rainfall/year, with cool rainy winters and hot and dry summers), the mildly arid steppe and semi-desert zone (350-150 mm), and the arid to hyper arid desert (<150 mm) (Fig. 1C); and b) the presence of numerous caves containing speleothems in the three climate zones. Speleothems grow in caves when water reaches the unsaturated zone and vegetation is present on the surface to supply the CO2 necessary for limestone dissolution (Hendy, 1971; Schwarcz, 1986), but they do not grow in water and soil-CO2 depleted conditions of arid and hyper-a rid deserts (Holmgren et al., 1995; Fleitmann et al., 2003). At present, the caves in the semi-desert and desert regions of Isra el are dry with no speleothem formation. However, the presence of speleothem s in these caves indicates
that wet conditions prevailed in the past. The focus of this study is to reconstruct (1) the timing of the speleothem growth, i.e., r econstruction of the periods of increased precipitation in present-day desert by precise dating usin g U-Th method, and (2) the origin of precipitation, eolian dust sources, and vegetation above the caves, using the oxygen, carbon and Sr isotopic composition of the speleothems, and the hydrogen isotopic composition of their fluid inclusions. 2. Geography, climate and geology of the Israeli deserts 2.1 Geographical definition and climatic zoning The Negev Desert (Fig. 1B, C) occupies the southern half of Israel between latitudes of 3130'N 2930'N and longitudes 34 30'E 3530'E. Its eastern geographical boundary is marked by the Dead Sea Transfor m (Dead Sea Arava Valley), the western geographical boundary by the Isr aeli-Egyptian border, and the southern limit is the Gulf of Elat (Gulf of Aqaba). The northern bounda ry is defined by the Mediterranean Sea coast of Gaza Strip and by a line ~20 km to the north of Be'er-Sheva Arad Valley. The Be'er-Sheva Arad Valley dissects southern Israel from west to the east along latitude ~3115'N, and separates between the CMR to th e north, and the highla nds of the central Negev to the south. The CMR creates strong "rain shadow" effect in Dead Sea and Jordan Valleys, because of their depressed position relatively to westerly winds. Thus, a 15-30 km strip of desert and semi-desert continue s to the north along the Dead Sea Transform till ~32 20'N: the arid to hyper-arid Judea Desert near the Dead Sea and the mildly-arid semi-desert zone in the Jo rdan Valley (Fig. 1C). Conventionally, the Negev Desert is divided to three geographical regions (Fig. 1C): the northern Negev to the north of Wadi Zin, the central Negev from Wadi Zin to Wadi Paran, and the southern Negev from the Wadi Pa ran to the Gulf of Elat. In this study the division of the Negev desert is different and based mainly on climate and vegetation zoning as follows:
1) The mildly arid northern Negev Desert, bordering the south-east corner of the Eastern Mediterranean (EM) Sea, forms a ~40 km wide belt that includes the coastal plain near Gaza, the southern edge of the CMR, and a few smaller ridges to the east near Arad. The northern border of the area is defined by th e 350 mm isohyet (the southern boundary of Mediterranean climate zone). The Be'er-Sheva Arad Valley dissects the area from west to east and the southern border of the northern Negev is defined by the ~150 mm isohyet along the southern margins of this valley. Vegetation in this region changes southward from C3 Mediterranean steppe forest to a mix of C3 and C4 semi-desert Irano-Turanian vegetation (Vogel, 1986; Cerling, 1993; Feinbr un-Dothan, 1998; Goodfriend, 1990; 1999). 2) The arid central Negev Desert or Negev Highlands, located south of Be'er-Sheva Arad Valley consists of several small NE-SW tre nding ridges with elevations of 500-1033m asl. Rainfall varies from 150 mm in the north to 50 mm in the south. The vegetation changes southward from semi-desert Irano-Turanian vegetation to Saharan-Arabian desert type, both comprising mixed C3 and C4 vegetation (Vogel, 1986; Cerling, 1993; Feinbrun-Dothan, 1998; Goodfriend, 1990; 1999). 3) The hyper-arid southern Negev Desert , located south of the Negev Highlands, presently receives ~30-50 mm average annual rainfall and is characterized by SaharanArabian desert flora (C3+C4) (Vogel, 1986). Whereas the nor thern parts of the Negev mainly receive winter rainfall associated with mid-latitude Atlantic-Mediterranean (Cyprus) cyclones, the southern Negev receives its rainfall mostly at the beginning (October-November) and end of the winter ra iny season (March-May) in sporadic short storms, usually accompanied by local floods. Some of this rainfall is associated with synoptic systems that originate in the trop ical Atlantic Ocean, pass over Africa and approach the region from the south-southw est (Kahana et al., 2002). These synoptic conditions infrequently occur also in the northern parts of the Negev.
The climate zoning in the rain shadow dese rts includes: the mildly arid steppe and semi-desert of northern and central Jordan Valley receiving 350-150 mm annual rainfall, and the arid deserts in southern Jordan Valley and the Judea Desert near the Dead Sea receiving 150-50 mm annual rainfall. These regi ons mainly receive winter rainfall from Cyprus cyclones, although rare thunderstorms bringing tropical moisture occur at the beginning or at the end of the rainy season, similar to the southern Negev (Kahana et al., 2002). South of the Dead Sea in the Arava Vall ey, the climate is hyperarid with less than 50 mm rainfall (Fig. 1B, C). The relative abund ance of C4 vegetation is higher in these regions compared with the Negev Desert, owing to the higher temperatures (Vogel, 1986). 2.2. Rainfall gradients in the northern Negev Desert and Jordan/Dead Sea Valleys At present the northern and central Negev De sert and the Jordan Valley region receive most of their rainfall during November to March from mid-latitude Cyprus cyclones moving eastwards above the EM Sea. The typical Cyprus cyclones (the major contributor of rainfall in Israel) correspond to the majo rity of long fetch maritime air masses crossing the Mediterranean (Dayan, 1986). Summers, between May and September, are hot and dry and result from the sinking air of s ubtropical highs, whic h develop over the Mediterranean Sea as strong high-pressure ridges pushed eastwar ds from the Azores subtropical high. The Mediterranean Sea is the snow/rainfal l moisture source and supplies the latent heat from its warm waters to the air masse s entering the southern Levant through the passage of the Cyprus Lows. When passing ov er the land, the supply of moisture and latent heat becomes dramatically reduced to the south, and more gradually to the east (Shay-El and Alpert, 1991; Enzel et al., 2007). Due to the orographic increase in precipitation above CMR and adiabatic heati ng of descending air masses on its eastern flanks, the rainfall decreases by 400-500 mm over ~25 km from the CMR crest to the
Jordan and Dead Sea Valleys, bringing about the formation of the "rain shadow" desert. Consequently, isohyets run from north to south in northern and central Israel parallel to the coastline and CMR, but then abruptly change their orientation and run to the west parallel to the northern coastline of the Sinai Peninsula. In the Negev Desert the precipitation decrease from north to the south becomes very sharp; even though there isnâ€™t any significant topographic barrier in th e 50 km belt inland of th e coastline (Fig. 1B, C). This sharp precipitation grad ient occurs in the zone wh ere the Cyprus cyclones cross the Levant from the west to the east, b ecause the northern Sinai coastline forms the southern limit at which rain clouds can form (Zangvil and Druian, 1990). Consequently, the rainfall amount north of Be'er-Sheva Arad Valley is much higher than to its south (Fig. 1C). 2.3 Geology During the Mesozoic and Cenozoic epochs Is rael was a marginal region between the land and sea. Thus, most of exposed sedi ments in the Negev are of CretaceousPleistocene age, and are composed of limest ones, dolomites, chalks and marls of marine origin, with fewer exposures of sedimentary rocks of terrestrial origin, such as sandstones, conglomerates, lacu strine marls, and loess (Fig . 2A, B). Triassic â€“ Early Cretaceous sandstones, dolomites, limestone and gypsum are exposed in erosional craters and deep canyons, and the deepest Pre-Cambrian crystalline basement rocks and Cambrian sandstones and dolomites are exposed in Elat Mountains in southern Negev. Karstic caves are mainly associated with dolomites and limestones of Cenomanian and Turonian Judea Group, with few caves in Eocene limestone rocks (Appendix 2). These marine sediments are usually exposed at the ridges. In the valle ys they are locally overlain by sandstones, conglomerates, and l acustrine marls of Miocene Hazeva Group, by lacustrine marls and conglomerates of Pliocene Arava Formation and Early Pleistocene Zhiha Formation, and by loess deposits and alluvium of Middle-Late
Pleistocene. Last major period of loess de position in the northern and central Negev occurred between 90 ka and 14 ka (Crouvi et al., 2007). Partial erosion of the loess occurred during the Holocen e (Avni et al., 2006). The present landscape of the Negev resulted from the two major geological processes: 1) The formation of the Syrian Arc, whic h started during the Senonian period about 70 Ma ago, and probably was associated with the plate movements during the beginning of the Alpine orogenic processes in the nort h. The NW-SE convergence stress re-activated Jurassic normal faults in a reverse directi on, causing the formati on of anticline chains extending from northern Sinai through the Nege v and central Israel to Syria. The folding of the anticlines was syn-se dimentary, with few pulses co ntinuing until the Neogene (Garfunkel, 1982; Garfunkel, 1988-a; Solomon and Garfunkel, 1989; Calvo, 2002). As a result of these processes, thick sequence of marine chalks, marls, cherts and phosphates were deposited in the synclines, and a thin se quence at the tops of the anticlines (Soudry et al., 1985). Senonian-Paleocene soft se diments overlying the hard dolomites and limestones of Judea Group, known as Mount Scopus Group, contain a high concentration of uranium (~120 ppm) (Soudry et al., 2002). The anticlines of the Syrian Arc form the present-day SW-NE ridges of the central Ne gev, small ridges to the east of Be'er-ShevaArad Valley and the CMR. 2) The formation and evolution of the D ead Sea Transform during the Neogene and Quaternary. Following a major marine regression at the end of the Eocene, the southern Levant was relatively flat until the middle Miocene, when the region started to experience left lateral strike slip movement. During the late Miocene â€“ Early Pliocene, a first major tectonic phase shaped the present outlines of the long and narrow depression (rift) along the Dead Sea transform and controlled the uplifting and tilting of its shoulders, which included the Negev Desert (Picard, 1943; Garfunkel and Horowitz, 1966; Garfunkel, 1970; Garfunkel, 1978; Garfunkel, 1988-b; Steinitz and Yosef, 1991; Zilberman, 1991 Avni et al., 2001-a). A second major tectonic phase during the Early Pleistocene caused
the central Arava Valley to subside by 500700 m relative to central Negev, and the western part of the southern Negev to be upl ifted by ~200 m relative to central Sinai. The central and southern Negev were tilted 1-2 eastwards creating a new drainage system flowing towards Dead-Sea Valley (Avni, 1991; Zilberman, 1991; Avni, 1997; Avni et al., 2001-a). As a result, erosion removed most of th e Tertiary and Senonian soft sediments, exposing the hard carbonate rocks of the Ju dea Group on the axes of the Syrian Arc anticlines. Conglomerates and lacustrine marls of P liocene Arava Formation were deposited in the southern Negev and the Arava Valley be tween the two major tectonic phases of the Dead Sea transform (Avni et al., 2001-a). Earl y Pleistocene conglomer ates and lacustrine sediments of Zhiha Formation were deposite d at the beginning of the second tectonic phase (Ginat, 1997; Ginat et al., 2003). Both th ese formations are associated with lakes. The caves were most probably formed betw een Late Eocene and Miocene when the Judea Group was below groundwater table. So me caves (mainly those on the anticline crests) were formed in open aquifer conditions , whereas others formed further from the anticline crests in confined conditions below the aquitards of Mount Scopus marls (Frumkin and Fischhendler, 2005). The caves were uplifted above groundwater table during the two late major tectonic phases, thus enabling the formation of vadose speleothems. 2.4 Speleothems as paleoclimatic records Vadose carbonate speleothems (stalactites, stalagmites and flowstone) form when meteoric waters react with soil CO2, dissolve the carbonate rocks in the unsaturated zone, and deposit secondary CaCO3 deposits in the caves as a result of CO2 degassing. The process of speleothem formation depends on the presence of water in unsaturated zone, which in turn depends of the effective prec ipitation (i.e. precipitation â€“ (evaporation + runoff)); and the presence of vegetation above the cave to supply the soil CO2 necessary
for the dissolution of carbona te host rock (Hendy, 1971; Buchmann and Dreybrodt, 1985; Schwarcz, 1986; Ford and Williams, 1989). Ava ilability of water is a particularly important factor controlling the speleothem growth rates in regions where climate fluctuates between arid and semi arid (Hol mgren et al., 1995; Vaks 2001; Fleitmann et al., 2003). Chemical models also show that f actors such as Ca concentrations in the seepage water, cave temperature, CO2 pressure and water flow patterns influence speleothem growth rates (Baker and Smart, 1995; Dreybrodt et al., 1996). Speleothems are excellent recorders of climate. They are protected from weathering and alteration; and can be dated precisely by the uranium-series di sequilibrium methods, with a precision ap proaching -2% (2 ), using Thermal Ioniza tion Mass Spectrometry (TIMS) and Multi-Collector Inductively C oupled Plasma Mass Spectrometry (MC-ICPMS) techniques. The high ionization efficien cy of the MC-ICP-MS relative to TIMS significantly improves the ioniza tion for thorium, which make this method particularly efficient for low-uranium young speleothems containing relatively little radiogenic 230Th (Robinson et al., 2002; Fleitmann et al., 2003) . Speleothems usually contain low amounts of detrital material, which make their ages more reliable than carbonate sediments with higher detrital contents (i.e., tufa and trav ertine, lacustrine carbonates and soil carbonate nodules). When deposited in isotopic equilibrium with the cave waters, the 18O values of speleothems and the D values of their fluid inclusi ons depend on cave temperature and the 18O and D values of the rainfall. The cave temperature usually approximates the mean annual surface temperature because it buffered by large mass of rock (Hendy and Wilson, 1968; Schwarcz, 1986). Rainfall 18O and D depend on the vapor source (the sea surface water 18O and D), rainfall amounts, the latitude, altitude and distance from the source (Dansgaard, 1964; Frumkin et al., 1999; Bar-Matthews et al., 1997, 2003-a). Thus, speleothem 18O values and the fluid inclusion D values are excellent proxies of the paleo-temperatures, source of water va por, and rainfall amounts (Hendy and Wilson,
1968; Schwarcz, 1986; Bar-Matthews et al ., 1996, 1997, 1999; Frumki n et al., 1999; Matthews et al., 2000; Dennis et al., 2001; Burns et al., 2003; McDermott, 2004; McGarry et al., 2004). The independent chronology of speleothem data offers opportunities to critically assess leads and lags in the climate system (timing and duration of major O isotopedefined climatic events â€“ e.g., Marine Isotope St ages (MIS) and smaller scale climatic changes), which in turn can provide important insights into global forcing and feedback mechanisms (McDermott, 2004). In this respec t, speleothems can of ten offer advantages over many other paleoclimate records. Speleothem 13C values depend on soil CO2 and the carbonate host rock. Under closed system conditions the contribution of the two sources is approximately equal (Hendy, 1971; Bar-Matthews et al., 1996; McDermott, 2004). The 13C values of the marine carbonate host rock is considered to be constant (-1 +1), while the 13C of soil CO2 depends on the vegetation type above the cave. Organic carbon originating from C3 vegetation type (Calvin photosynthesis cycle) has 13C values ranging from -26 to 20, whereas C4 vegetation type (Hatch-Slack cycle) has values from -16 to -10. These differences are preserved as distinctive ranges of 13C in speleothem carbonates: from -14 to -6 for C3 plants and fr om -6 to +2 for C4 plants. However, speleothem 13C values can become higher due to deposition in partially open system conditions, resulting in a kine tic fractionation during rapid CO2 degassing, and incomplete equilibration of water with soil CO2 during periods of heavy rain because of short residence time of the wa ter in the soil zone (Schwarc z, 1986; Bar-Matthews et al., 1996, 1997, 2000; Fleitmann et al., 2003; McDermott, 2004). Trace elements such as Sr and U are incorporat ed into the calcite la ttice. Their isotopic ratios (87Sr/86Sr and 234U/238U) in speleothem calcite ca n provide information on the upper vadose zone water chemistry, water-soil-rock interactions, weathering and origin of
eolian dust (Banner, 1995; Goede et al., 1998; Ayalon et al., 1999; Frumkin and Stein, 2004; Li et al., 2005). Growth rate, growth periods and depositional hiatuses (growth breaks) are important indicators of the presence and amounts of water in the unsaturated zone, and are thus an indication of desertification processes (Holmgren et al., 199 5; Vaks 2001; Fleitmann et al., 2003), or permafrost formation and melti ng in cold climates (Harmon et al., 1977; Schwarcz, 1986; Lauritzen, 1995). Laminae th ickness, crystal morphology, mineralogy and petrography can be indicator s of dripping rates, degree of super saturation of the precipitating water, fluid chemistry and the nature of the fluid transport (Bar-Matthews et al., 1991, 1997; Ayalon et al., 1999; Frisia et al., 2000; 2002). Large columnar crystals are usually indicative of slow and constan tly dripping water where the speleothem is continuously wet and growing at near equilibrium saturation conditi ons with respect to the CaCO3. A fabric consisting of small crystals w ith a high density of crystal defects is indicative of deposition under similar low supe r saturation conditions, but with variable water discharge and presence of growth inhi bitors. Dendritic fabr ics with the highest density of crystal defects are indicative of speleothem deposition under high super saturation disequilibrium conditi ons and prolonged periods of low fluid flow regime that resulted in prolonged out gassing (Frisia et al., 2000). 2.5. Speleothem climate record from the Med iterranean climate zone of Israel Accurate high resolution U-Th dating of speleothems from the Soreq, Jerusalem and Peqi'in caves, in central and northern Israel (Fig. 1B, C), show that speleothem deposition was continuous over several glacial-interglacial cycles, thus providing a powerful tool for understanding the EM paleoclimate (Ayal on et al., 1999; 2004; Bar-Matthews et al., 1997, 1999, 2000, 2003-a; Frumkin et al., 1999, 2000). The 18O isotopic profile of the EM speleothems shows a good correspondence with the stacked oxygen isotopes (Martinson et al., 1987) and with the marine record of the EM Sea, suggesting that their
isotopic composition reflects global and regiona l climatic changes, and that climatic events on sea and land were linked (Frumkin et al., 1999; Bar-Matthews et al., 2000, 2003-a). Periods of low 18O values usually coincide with low 18O values of the planktonic foraminifera G. ruber (i.e., sea surface water), with th e times of increased atmospheric CO2, warming, sea level rise and the formation of Mediterranean sapr opels. Sapropels are black, organic rich layers found in sediment s throughout the Mediterranean Sea. They are interspersed within the normal, whitish, oxidized hemi pelagic sediments and reflect oxygen deficient conditions in the deep-sea due to stratification of the water column resulting from a high input of the fresh water. Periods of sapropel formation in the EM Sea are associated with the minimum speleothem 18O both during interglacials (Early Holocene, MIS-5.1, 5.3, 5.5, 7.1, 7.3, 7.5), but also during glacial inte rstadial MIS-6.4 (Bar-Matthews et al., 2003-a; Ayalon et al., 2004), suggesting increased discharge of freshwater to the Medite rranean Sea via the Nile (Olausson, 1961; Ryan, 1972; Rossignol-Strick and Paterne, 1999), via other rivers from present-day Sahara Desert region (Rohling et al., 2002) and via increase in rainfall over the entire Mediterranean basin (Kallel et al., 2000) The highest speleothem 18O values coincide with higher 18O values of the G. ruber (i.e. colder EM Sea surface temperatures), corresponding to extreme global cooling episodes, such as Hein rich events and the Last Glacial Maximum (LGM) (Ayalon et al., 2002; Bar-Matthew s et al., 1999, 2000, 2003-a). The strong links between sea and land 18O records shows that the 18O of the rainfall above the caves was mainly controlled by the EM Sea prec ipitation source (Frumkin et al., 1999; BarMatthews et al., 2003-a; Kolodny et al., 2005). S uperimposed on this, rainfall amount is important control on the 18O of precipitation and therefore on the 18O of the speleothems (Bar-Matthews et al., 1997; 2003a). The influence of the â€œamount effectâ€ (as originally defined by Dansgaard (1964)) on speleothem 18O in EM Sea region is still under debate. Frumkin et al. (1999, 2000) and Kolodny et al. (2005) argue for a
negligible contribution whereas Bar-Matthews et al., (2000; 2003-a) argue that periods with very little 18O change in EM Sea surface wate r, but larger contemporaneous 18O changes in speleothems, reflect the amount effect. The EM Sea as a precipitation source is also evident by D18O relations based on D values of fluid inclusions (FI) trapped in EM speleothems. Present-day precipitation originating in the EM Sea is characterized by d-excess values ( D-818O) of ~+22, defining the local Mediterranean Meteoric Water Line (MMWL) ( D=8 18O+22) (Gat and Carmi, 1987; Ayalon et al., 1998), which differs from the global Meteoric Water Line (MWL) relationship ( D=8 18O+10) (Craig, 1961). The D18O relationships of paleo cave waters determined from fluid inclus ions show that interglacial waters follow the MMWL, but most glacial waters are sl ightly offset from the MMWL, becoming closer to the MWL during th e last glacial maximum (LGM ) (Matthews et al., 2000; McGarry et al., 2004). Temperature changes on land and sea are very similar as evident from alkenone sea surface temperatures fr om EM sediments (Emeis et al., 2000; BarMatthews et al., 2003-b; Kolosovsky, 2003) an d calculated land temperatures inferred from the D of trapped fluid inclusio ns (McGarry et al., 2004). Speleothem 13C values in central and northern Israel are generally in the range from 9 to -13 and are indicative of the dominance of C3 type Mediterranean vegetation during interglacial periods. During global cooling episodes like the glacial maxima, 13C values are higher (~ -7 -9 ), indicative of higher contri bution of C4 type vegetation and drier conditions (Bar-M atthews et al., 1997, 1999, 2000, 2003-a). During interglacial intervals with low 18O minima at 127-121 ka and 9-7 ka 13C values are anomalously high. Bar-Matthews et al. (1997, 2000, 2003-a) sugge sted by them that these high values represent deluge episodes that stripped the soil cover, and/or the rainwater residence time in the soil zone was too short to equilibrate with the soil CO2. Frumkin et al., (2000) on the other hand had suggested that these periods of high 13C are indicative of dry conditions.
The 87Sr/86Sr ratio of speleothems is a proxy for paleo-dust input and weathering conditions of the cave host rock. Ayalon et al., (1999) and Bar-Matthews et al., (2000-b) found that concentrations of Sr and U and 87Sr/86Sr and 234U/238U ratios in Soreq Cave speleothems are lower (less radiogenic) duri ng interglacials, and interpreted this to indicate wetter conditions cau sing enhanced leaching of soil and rock. They also found a higher values of these parameters during the last glacial period, attributing it to an increase in the contribution of salts derive d from exogenic sources, (sea spray and aeolian dust), with less leaching of the host rock, ow ing to drier conditions. Frumkin and Stein (2004) found the similar patterns in Jerusalem Cave speleothems, and suggested that Sr and U in speleothems were mainly derived fr om the soils above the caves. The increase in 87Sr/86Sr and 234U/238U ratios of the Jerusalem cave speleothems were interpreted by Frumkin and Stein (2004) to represent enhan ced supply of Saharan dust during the glacial periods, due to the increased ar idity of the Sahara Desert. Th e decrease of these values during interglacials was attributed to shut -down of Saharan dust source as a result of wetter conditions in the source regions and an increase in host rock contribution of Sr and U. 3. Research goals 3.1 The major aims of the study The objective of this study is to determine using the speleothems the paleoclimate of the Israeli deserts: the Negev Desert, southern Israel, and th e rain shadow deserts in the Jordan and Dead Sea Valleys, located betw een present day isohyets of 300 mm and 30 mm, in the following terms: 1) Reconstruction of the period icity of wet and dry interval s in: (a) the mildly arid and arid zones of the Jordan and Dead Sea Valleys and northern Negev and (b) the arid and hyper arid zones of the central and southern Negev Desert. To use this data to determine: (c) how the northern boundary of Saharan-
Arabian Desert has shifted in time and space and (d) how these shifts may have influenced the migration of hominids and an imals between Africa and Asia via Sinai-Negev land bridge. These goals are mainly achieved by U-Th dating of speleothems. 2) Reconstruction of rain so urces, rainfall amounts and paleo-vegetation type above the caves. These goals were achieved by the determination of high resolution 18O and 13C records of calcite laminae and the D values of trapped fluid inclusions and comparison of these records with other marine and terrestrial paleoclimate proxies in the region, especially with th e well-dated high-resolution records of speleothems in the Mediterra nean climate area of Israel (Soreq, Jerusalem and Peqiâ€™in caves). 3) Determination of the origins of speleoth em Sr and dust provenance in the Negev Desert and Jordan Valley. This was done by analysis of Sr concentrations and 87Sr/86Sr ratios in speleothems, host rocks and dust-born soils; the relationship between them and the speleothems in the Mediterranean climate region (Soreq and Jerusalem Caves). 3.2 Structure of the thesis The thesis is composed of three published arti cles and one chapter. The research topics in each part are detailed below: The first article is: â€œVaks et al. (2003), Paleoclimate reconstruction based on the timing of speleothem growth, oxygen and carbon isotope composition from a cave located in the â€œrain shadowâ€, Israel., Quaternary Research , 59, 2, pp 182-193â€. The aim of this study was a paleoclimate reconstruction in the mildly-ari d semi-desert of the Jordan Valley, located in the â€œrain shadowâ€ of CMR, using the speleothem record for Maâ€™ale-Efrayim Cave (#1, Fig. 1C, Appendix 2).
The second article is: â€œVaks et al. (2006), Paleoclimate and location of the border between Mediterranean climate region and the Saharo-Arabian Desert as revealed by speleothems from the northern Negev Desert, Israel., Earth and Planetary Science Letters , 249, 3-4, pp. 384-399â€. The aim of this study was the paleoclimate reconstruction in the mildly-arid semi-desert of the northern Negev using the speleothem record from Maâ€™ale-Dragot a nd Tzavoa Caves (#3 and #4 re spectively, Fig. 1C, Appendix 2). The focus was mostly on the Tzavoa Cave, being the southernmost cave in this region containing thick speleothems (few tens of cm) younger than 550 ka (U-Th method limit). The third article is: â€œVaks et al. (2007), Desert speleothems reveal climatic window for African exodus of early modern humans, Geology, 35, 9, pp. 831-834â€. This study reconstructs humid periods in arid and hyper arid central and southern Negev Desert during the last 180 ka using the spel eothem record for five caves: Hol-Zakh, Ashalim, Even-Sid-Ramon, Maâ€™ale-ha-Meyshar, and Ktora Cracks (#6, #8, #9, #12, and #14 respectively, Fig. 1C, Appendix 2). The study examines the possible link between the desert humid periods and migration of hominids and animals out of Africa through the Saharan-Arabian Desert and Sinai-Negev land bridge. The last chapter of the thesis is: â€œPal eoclimate of northern Saharan-Arabian Desert from speleothem record in central and southern Nege v and Judea Desert, Israelâ€. This chapter deals with a detailed paleoclimate reconstruction in the central and southern Negev and Judea deserts, for longer time intervals (Pliocene â€“ present) than were discussed in articles 1-3. The speleothem record of the following caves was used: Kanaim, Izzim, Hol-Zakh, Makhtesh-ha-Qat an, Ashalim, Even-Sid-Ramon, Wadi-Lotz, Maâ€™ale-ha-Meyshar, Shizafon and Ktora Cr acks (#2, #5 #9, #11 #14 respectively, Fig. 1C, Appendix 2). The record of Ma'ale-Efrayim Cave, Jordan Valley was expanded for this study. This chapter also studies the origins of the speleothem Sr and the dust provenance using Sr isotopic compositions from the major studied caves.
4. Methodology 4.1 The studied caves and sampling Fourteen Caves were studied: twelve caves from the Negev Desert, one cave in the Judea Desert near Dead Sea Valley and one cave from the Jordan Valley. Twelve caves are located within Cretaceous (Cenomanian and Turonian) host dolomite and limestone and one within Eocene limestone. The cave na mes and their numbers in Fig. 1C are: Maâ€™ale-Efrayim Cave (#1), Kanaim Cave ( #2) Maâ€™ale-Dragot Caves (#3), Tzavoa Cave (#4), Izzim Cave (#5), HolZakh Cave (#6), Makhtesh-ha-Q atan Cave (#7), Ashalim Cave (#8), Even-Sid-Ramon mini-caves (# 9), Mitzpe-Ramon Cave (#10), Wadi-Lotz Cave (#11), Maâ€™ale-ha-Meyshar Cave (#12), Shizafon mini-caves (#13), Ktora Cracks (#14). The coordinates and geological and geographical background of each cave are given in Appendix 2. Maâ€™ale-E frayim Cave and Kanaim Cave are located in the â€œrain shadowâ€ desert: Maâ€™ale-Efrayim cave in mildly -arid semi-desert in Jordan Valley located between 250 mm and 300 mm isohy ets, and Kanaim Cave in ar id Judea Desert located at ~100 mm isohyet. The twelve Negev caves are located along north-south climatic transect from ~ 300 mm isohyet in the north to ~30 mm isohyet in the south (Fig. 1C and in Appendix 2). 123 in-situ or broken vadose speleothems (flows tones, stalactites and stalagmites) and phreatic speleothems were collected from the fourteen caves. The speleothems vary in size from few mm thick flowstone crusts to 60 cm length 20 cm width stalagmites, stalactites, and flowstones. The samples were sectioned using a diamond saw to expose their internal structure and to eliminate diagenetically alte red samples (Bar-Matthews et al., 1997). The mineralogy an d petrography was determined using petrographic microscope, Jeol 840 scanning electron micros cope equipped with Oxford ISIS EDS system, and Philips PW 3020 X-ray diffractomet er. Preliminary petrographical studies and U-Th dating using alpha-spectrometry indicated that the speleothems older than the U-Th method limit were composed of large columnar crystals a few cm in length,
forming massive layers ra nging from a few cm to 40 cm thick, whereas younger speleothems were thinly la minated. Based on their petrog raphy sixty five speleothems including thinly laminated sections from the thirteen caves (a ll caves except MitzpeRamon Cave) were chosen for high precision U-Th dating: eight speleothems from Maâ€™ale-Efrayim Cave and other small caves in its surroundings (Vaks et al., 2003); five speleothems from Kanaim Cave, Judea Dese rt; six speleothems from Maâ€™ale-Dragot Caves and fifteen speleothems from Tzavoa Cave, northern Negev (Vaks et al., 2006), and thirty one speleothems from other ten caves of central and southern Negev Desert (Vaks et al. (2007) and , Table 1)2. 4.2 U-Th and U-Pb dating 37 speleothem samples from Ma'ale-Efrayim Cave were dated using thermal ionization mass spectrometers (TIMS) at the Open University, Milton Keynes, UK, and six other samples from this cave were dated by alpha-spectrometry in Geological Survey of Israel (Vaks et al., 2003) using methods described in Bar-Matthews et al. (1997), Kaufman et al. (1998), McDermott et al. ( 1999) and Vaks (2001). Most of the U-Th dating (295 speleothem samples from thirteen caves â€“ (Vaks et al ., 2006) and Table 1) was performed using the MC-ICP-MS in the Geological Survey of Israel. For dating purposes up to 1 g material was drilled using 0.8 4 mm diameter drill bits along the growth axis for stalagmites (Fig. 2A), and across the growth axis or crosssection for stalactites and flow stone (Fig. 2B). A few laminae were sampled two times or more by drilling in few different locations along the same lamina (these samples with duplicate or triplicate dating are designated by the roman numerals I and II in Table 1). Depending on the uranium concentration, 10-800 mg calcite powder was dissolved in 7N HNO3 (Table 1). The sample was loaded ont o mini-columns contained 2ml Bio-Rad AG 1X8 200-400 mesh resin. U was eluted by 1N HB r and Th with 6N HCl. The U and Th 2 All tables in this work are given in Appendix 4
solutions were evaporated to dryness and dissolved in 2ml and 5ml of 0.1N HNO3 respectively (McDermott, 2001; Kuperman, 2005). High precision U-Th dating was performed using a Nu Instruments MC-ICP-MS, equipped with 12 Faraday cups and 3 ion counters. The sample was introduced to the MC-ICP-MS through an Aridus micro-concentric desolvating nebuliser sample introduction system. The instrumental mass bias was corrected using an exponential equation by measuring the 235U/238U atomic ratio and correcting with the natural 235U/238U ratio (0.0072). The calibra tion of ion-counters relatively to Faraday cups was performed using several cycles of measurement with different collector configurations in each particular analysis. The age determin ation was possible due to the accurate determination of 234U and 230Th concentrations by isotope dilution analysis using the 236U-229Th spike. U-Th ages were corrected for detrital 230Th (Kaufman et al., 1998), assuming a 232Th/238U isotope atomic ratio of 3.8 (the m ean crustal value) in the detrital components. However, less than 10% of the samples had a 230Th/232Th activity ratio less than 30 and needed this correction (Table 1). The reproducibility of 234U/238U ratio was 0.11% (2 ). For a detailed description of the chem ical preparation of the samples and the dating by MC-ICP-MS see the Appendix 3 and Kuperman (2005). Speleothem deposition in the arid zone of Israel was intermittent and depositional hiatuses with distinct petrographic characteri stics separated the different calcite laminae (Chapter 8.4.2, Vaks et al., (2003, 2006, 2007)). E ach calcite lamina represents a growth period. In order to determine the duration of the speleothem growth periods, the top and base of growth laminae were dated. Where possible, several age determinations were performed across the lamina cross section. Th is enabled calculation of the growth rates during the periods of speleoth em deposition, with the assump tion that the age represents the center of the drilled area, and that there is a constant growth rate between two dated points. When growth laminae were very thin it was possible to obtain only one U-Th age. In such situations, or if the ages at the t op and at the botto m of the lamina were within
each otherâ€™s 2 uncertainty range, it was assumed th at the entire grow th period was not longer than 1000 years. Five samples of speleothems older than the U-Th dating limit, ASH-15-C, ASH-15-D and ASH-15-E from Ashalim Cave and KTO( 1)-J and KTO(1)-1-K from Ktora Cracks, were dated using U-Pb method in University of Melbourne, Australia, according to the method described in Woodhead et al., (2006), and summarized Appendix 3, section A3.3. The sample ASH-15-D+E from Ashalim Cave (including both the ASH-15-D and ASH15-E in one piece), was also dated using UPb method in University of Leeds, UK, according to the method of (Walker et al., 2007). 4.3 Stable isotope analyses of speleothems and present day rainwater For 18O and 13C analyses, samples of 1-2 mg mate rial were drilled using a 0.8-1 mm diameter drill, either along or across the growth axis. Hendy tests (55 measurements of 18O and 13C) were performed on several stalagmites and flowstone samples from Ma'ale-Efrayim and Tzavoa Caves, Ashalim Cave, Hol-Zakh Cave and Ma'ale-haMeyshar Cave to ensure that speleothem calcite deposition occurred in isotopic equilibrium with the dripping water (Hendy, 19 71). These tests were most important in Tzavoa and Ashalim Caves because these cave s have natural openings. At all locations, the speleothems were indicated to have depos ited in equilibrium. More than a thousand measurements of 18O and 13C were performed in high resolution on 5 stalagmites, 7 flowstone samples and 9 stalactites from Ma'ale-Efrayim, Tzavoa, Hol-Zakh, Ashalim, Even-Sid-Ramon, Ma'ale-ha-Meyshar and Shizafon caves using VG SIRA-II Mass Spectrometer with ISOCARB system for carbona te analysis (Bar-Matthews et al., 1997). 18O and 13C values of calcite are reported in permil () relative to the PDB standard. 18O and 13C profiles were measured in speleoth ems having dateable laminae thicker than 4 mm, which enabled sampling of at least 3-4 samples across the lamina. The isotope profiles together with the U-Th ages enabled wiggle-matching between the 18O
profiles of different speleothems that grew in the same time period (from the same cave or the adjacent caves) within 2 dating uncertainty of each sample. Data on the present day rainfall was obtai ned by collecting rainwater samples during two rainy seasons, winters of 1998-99 and 1999-2000 (September to May) in village of Ma'ale-Efrayim, Jordan Valley; and winters of 2004-5 and 2005-6, from the five stations located on a north-south tran sect of the Negev Desert (Fig. 1C and Table 2): 1) The city of Beâ€™er-Sheva, northern Negev, 380 m asl, and ~40 km from the EM Sea coast; 2) The town of Arad, northern Negev, 600 m as l, and ~80 km from the EM Sea coast; 3) Makhtesh-ha-Qatan, on the western fla nk of the Dead-Sea Valley and in rain shadow of the hills to the west, 1 km from the cave 3, ~ 0 m asl (on sea level), and 98 km from the EM coast; 4) The town of Mitzpe-Ramon in the highla nds of the central Negev, 2 km from cave #5; 820 m asl, ~100 km from the EM Sea coast); 5) The village of Neot-Smadar, at the s outhern Negev, located between the caves 9 and 10; 400 m asl, ~165 km from the EM coast). In order to estimate 18O and D of meteoric water in sout hern Negev during the years preceding the program of rainfall sampling, th e water of Ein-Netafim spring near Elat was sampled in June 2004. This spring water represents the mean 18O and D values of major rainfall events during the preceding years. Each rain event was sampled in Ma'ale-Efrayim, Beâ€™er-Sheva and in Neot-Smadar during the first year, whereas at the other sites the rain was allowed to accumulate in plastic bucket with a 1 cm thick oil layer a dded to prevent evaporation. Rainwater that accumulated below the oil was sampled at 20-45 days intervals and transferred to sealed plastic bottles. 18O and D measurements of the cave an d rain water were performed using methods described in Bar-Matthews et al. (1996) and Ayalon et al. (1998), and are reported in the SMOW scale.
4.4 D/H analyses of speleothem fluid inclusions The thermal decrepitation technique was used for extraction of the fluid inclusions (FI) water from the calcite. The ex tracted water was reduced to H2. The measurement of D followed the procedures of Yonge (1982), modified by Matthews et al. (2000) and McGarry et al. (2004). This technique has been found most applicable to the speleothems containing small amounts of water (McGa rry et al., 2004). Nine speleothems, 2 stalagmites, 2 stalactites and 5 flowstone samples from Tzavoa, Hol-Zakh, Ashalim and Maâ€™ale-Ha-Meyshar caves were studied fo r D/H analyses of FI. Speleothem FI D of Ma'ale-Efrayim Cave were stud ied by McGarry et al. (2004). D analyses were made on 32 samples, each representing a growth lamina from a particular speleothem. The laminae were chosen using the two following criteria: 1) they represent the main depositional periods in northern, central and southern Nege v Desert based on U-Th dating; and 2) they are composed of pure calcite w ith typical columnar fabric with no evidence of alteration and contain very low detrital material such as clay, to avoid extr action of water of noncalcitic origin. The thermal decrepitation t echnique involves an isotopic fractionation accompanying the extraction of calcite bound water. This fr actionation needs to be calibrated. For Soreq Cave speleothems Matthews et al. (2000) and McGarry et al. (2004) calibrated for the fractionation: ( D FI water = D extracted water D cave water) = -30. In the present study it was found that vacuum line blank has increased relatively to the previous studies. The contribution of the vacuum line blank was es timated by the analysis of samples with different water amounts giving the following relation, which was used to correct the results in this study: DFI water = Dextracted water (6.6 VS 40) where Vs amount of the extracted water in ml measured by the mass spectrometer.
4.5 Sr isotope analyses Strontium concentrations and isotopic ratio s were studied on 115 speleothem samples from the following nine caves: Maâ€™ale-Efr ayim, Maâ€™ale-Dragot, Tzavoa, Hol-Zakh, Makhtesh-ha-Qatan, Ashalim, Even-Sid-Ramon, Maâ€™ale-ha-Meyshar, Shizafon and Ktora Cracks. In order to determine the Sr conten ts and isotopic composition of soil and rock, which are the main contributors of Sr to the speleothems (Aya lon et al., 1999; BarMatthews et al., 1999; Frumkin and Stein, 2004), 9 samples of Cenomanian, Turonian, Senonian and Eocene rocks adjacent to the caves (Fig. 2) were analyzed as well as 11 samples of the soils above the caves. 3 samp les of paleosols washed to the Maâ€™ale-haMeyshar and Even-Sid-Ramon caves (See sec tion 8.2.7) were also analyzed. Strontium contents of samples were measured us ing Perkin Elmer Optima 3300 ICP-AES. For speleothem and host rock samples, 100 mg of the powdered calcite was dissolved in 3N HNO3. For silicate soils, 0.25 g samples of the sieved < 60 m fraction dried in 110 C, were reacted with 1.25 g LiBO3 in 900C in platinum-gold crucibles. The glass formed after cooling was dissolved in 4% HNO3 and 500 ml of distilled water was added. Sc was added as an internal standard prior to chemical analysis of samples. In order to separate silicate fraction from the carbonates in the soil samples, 10-40 mg of sample containing 1-2 g Sr was dissolved in 1.25M HNO3 for 24 hours in order to dissolve the carbonate fraction, then centrifuge d to deposit the silicates. The aliquots were removed and remaining material was washed in double distilled H2O to remove the remains of aliquots, and evaporated to dryness. For extraction of Sr from the silicates 5 ml HF+1ml HNO3 was added to 0.1-0.5 g of the sample (according to the Sr content) and evaporated to dryness. The resi dues were dissolved in 3.5M HNO3, and the solution was then loaded on a column containing the resin Eichrom-Sr-spec (50-100 mesh), and the matrix was rinsed away with 3 portions of 1 ml and one portion of 0.5 ml of 3.5M HNO3. Strontium was eluted from the column by 3X1ml 0.05M HNO3.
87Sr/86Sr ratios were measured using the MC-ICP-MS following the procedure of (Ehrlich et al., 2001). Five Faraday collector s were used for measurements of masses 83 and 85-88 for isotopes of Kr (83), Rb (85) and Sr (86, 87 and 88). 83Kr was measured in order to correct for 86Kr interference (86Kr=1.52 83Kr), due to the Kr contamination of the argon gas. 85Rb was measured in order to correct for 87Rb interference. 87Rb is calculated from the ratio 87Rb=0.3860 85Rb, after mass discrimina tion correction with the exponential law. The correction factor is calculated from the measured 87Sr/86Sr ratio, employing the exponential law mass bias correction, using the 87Sr/86Sr natural ratio of 0.1194.
Paleoclimatereconstructionbasedonthetimingofspeleothemgrowth andoxygenandcarbonisotopecompositioninacave locatedintherainshadowinIsraelAntonVaks,aMiryamBar-Matthews,a,*AvnerAyalon,aBettinaSchilman,aMabsGilmour,bChrisJ.Hawkesworth,cAmosFrumkin,dAaronKaufman,eandAlanMatthewsfaGeologicalSurveyofIsrael,30MalcheiIsraelStreet,Jerusalem95501,IsraelbDepartmentofEarthSciences,TheOpenUniversity,WaltonHall,MiltonKeynes,MK76AA,UKcDepartmentofEarthSciences,TheUniversityofBristol,Bristol,UKdDepartmentofPhysicalGeography,TheHebrewUniversityofJerusalem,Jerusalem91905,IsraeleDepartmentofEnvironmentalSciences,TheWeizmannInstituteofScience,Rehovot76100,IsraelfDepartmentofEarthSciences,TheHebrewUniversityofJerusalem,Jerusalem91905,Israel Received29July2002 Abstract High-resolution230Th/234Uagesand18Oand13CcompositionsofspeleothemsinMaâ€™aleEfrayimCavelocatedtotheeastofthe centralmountainridgeofIsraelenableustoexaminethenatureoftherainshadowaridityduringglacialandinterglacialintervals. Speleothemgrowthoccurredduringmarineglacialisotopicperiods,withnogrowthduringthetwolastmarineisotopeinterglacialintervals andduringthepeakoftheLastGlacialMaximum.Thiscontrastswithspeleothemgrowthincaveslocatedonthewesternankofthecentral mountainridge,intheEasternMediterraneansemiaridclimaticzone,whichcontinuedthroughoutthelast240,000yr.Thus,duringglacial periodswaterreachedbothsidesofthecentralmountainridge.Acomparisonofthepresent-dayrainandcavewaterisotopiccompositions andamountsattheMaâ€™aleEfrayimCavesitewiththoseonthewesternankshowsthatevaporationandhighertemperaturesontheeastern ankaremajorinuencesonisotopiccompositionandthelackofrainfall.The18Oand13Cprolesofthespeleothemsdepositedbetween 67,000and25,000yrB.P.matchthegeneraltrendsoftheisotopicprolesofSoreqCavespeleothems,suggestingasimilarsource(eastern MediterraneanSea)andsimilarclimaticconditions.Thus,duringglacialperiodsthedesertboundaryeffectivelymigratedfurthersouthor eastfromitspresent-daylocationontheeasternank,whereasinterglacialperiodsappeartohavebeensimilartothepresent,withthedesert boundaryatthesameposition.Thedecreaseinoveralltemperatureandaconsequentreductionintheevaporationtoprecipitationratioson theeasternankareviewedasthemajorfactorscontrollingthedecayoftherainshadoweffectduringglacialperiods. 2003ElsevierScience(USA).Allrightsreserved.Keywords: Speleothems;18O;13C;Paleoclimate;EasternMediterranean;Rainshadow;Effectiveprecipitation Introduction IsraelislocatedintheeasternMediterraneanregionin thetransitionzonebetweenahumidclimateinthenorthand anextremelyaridclimateinthesouthandsoutheast.The aridareasoftheLevantcanbedividedintotwotypes.One typeincludestheNegevDesertinsouthernIsraelandthe desertsinSinaiandsouthernJordanwhicharepartsofthe Saharo-Arabiandesertbeltformedbythesubtropicalhighpressurezonelocatedmainlytothesouthoflatitudes30â€“ 31N.Theothertypeisthelocalrainshadowdesertwhich islocatedontheeasternanksofthemountainridges,along theDeadSeatransform.Thetemperaturesontheeastern sideofthecentralridgeofIsrael(Fig.1)areusuallywarmer byseveraldegreesthanthoseonthewesternside,andthe raincloudscomingfromthewest(i.e.,fromtheeastern MediterraneanSea)areadiabaticallycooledbythemoun*Correspondingauthor.Fax: 00-972-2-5380688. E-mailaddress: firstname.lastname@example.org(M.Bar-Matthews). R Available online at www.sciencedirect.com QuaternaryResearch59(2003)182 www.elsevier.com/locate/yqres 0033-5894/03/$â€“seefrontmatter2003ElsevierScience(USA).Allrightsreserved. doi:10.1016/S0033-5894(03)00013-9
Fig.1.Locationmap,showingthestudiedarea.ThearrowshowsthedirectionofthecentralmountainridgeofIsraelwithMaâ€™aleEfrayimCave(1)tothe eastandSoreqCave(2)tothewest.Thegraylevelsrepresenttheaverageannualrainfall.Thetopographicheightsaregiveninmeters. 183 A.Vaksetal./QuaternaryResearch59(2003)182
tainridge,whichrisesto800 â€“1000m(Fig.1),producing orographicprecipitation.Ontherainshadowsideofthe ridgetheairundergoescompressionalwarminganddrying. Moststormtracksreachingtheareaoriginateinthe AtlanticOceanandpassovertheeasternMediterranean Sea.Theinteractionofcoldandrelativelydrycontinental airmasseswiththewarmMediterraneanSeaduringthe winterresultsinrainfallwithauniqueisotopicsignaturein which18Oâ€“DrelationshipsdenethelocalMediterraneanMeteoricWaterLine(MMWL)withD 8*18O 20 â€“30(e.g.,GatandCarmi,1970,1987;Gat,1996).The18OandDvaluesofrainfallbecomemoredepletedasthe rainmovesinlandawayfromtheeasternMediterraneanSea andaselevationincreases(e.g.,Dansgaard,1964;Gat, 1996).Asaresult,underpresent-dayconditionsthe18O valuesofrainfallingneartheshorelineareabout 4to 5,whereasonthecentralmountainridgeofIsrael(Fig. 1)the18Ovaluesdecreasetoabout 5to7 asa functionofelevation(Bar-MatthewsandAyalon,2001; Bar-Matthewsetal.,2003). Thepresentresearchdealswiththetimingandisotopic compositionofspeleothemdepositionfromMa â€™aleEfrayim Cave.Thecaveislocatedintherainshadowontheeastern sideofthecentralmountainridge.Thetimingofspeleothem growthprovidesanidealindicationoftheclimaticconditions,becausetheformationandgrowthratesofspeleothemsdependontheavailabilityofwater.Incolder climates,speleothemsdonotgrowduringglacialperiods becauseofthefreezingconditions(e.g.,Gascoyneetal., 1982;Gordonetal.,1989;Lauritzen,1995;Schwarcz, 1986).Thus,inwarmerareas,suchasthesiteofthisstudy intheLevantregionoftheeasternMediterranean,thetimingofspeleothemgrowthcanserveasanindicatorforwhen theamountofrainfallislessthantheamountofwaterlost throughevaporation. Thepurposeofthepresentstudywastoenhanceour understandingoftheclimaticeffectoftherainshadowin threemajoraspects:(1)todeterminethepresent-dayeffect oftherainshadowandevaporationontherainfallamount anditsisotopiccomposition;thiswascarriedoutthrough measurementsoftherainfallamountandtheoxygenand hydrogenisotopiccompositionofpresent-dayrainandcave waterinMaâ€™aleEfrayimandthroughcomparisonwiththe rainandcavewatercompositionfromSoreqCave,located intherainierregiononthewesternsideoftheridge;(2)to evaluatetheeffectoftherainshadowonspeleothemgrowth duringinterglacialandglacialintervals;thiswascarriedout throughdetailed230Th-UTIMSdating;and(3)toevaluate theeffectoftherainshadowonthepaleoclimaticconditions onbothsidesofthecentralmountainridgeduringthepast 80,000years;Thiswasachievedthroughmeasurementsof thecarbonandoxygenisotopecompositionoftheMa â€™ale Efrayimspeleothemsandthroughcomparisonofthisrecord withthewell-studiedrecordoftheSoreqCave(e.g.,Ayalon etal.,2002;Bar-Matthewsetal.,2000,2003). Cavelocationandresearchmethods Studyarea Maâ€™aleEfrayimCaveislocatedintherainshadowonthe easternsideofthecentralridgeofIsrael,250mabovesea level,60kminlandfromtheMediterraneanSea,andnear theJordanRiftValley(Fig.1).Thepresent-dayaverage annualrainfallnearthecaveis250 â€“300mm,themean annualtemperatureis21â€“22C,andtheannualpotential evaporationis1800mm.Theareaiskarsti edCretaceous dolomitewithscarceIrano-Turanicvegetationlocated mainlyinsoilpockets.ThecavewasformedalongatectonicfractureinthedolomiticCenomanianAmminadav Formationandispresently2 â€“20mbelowthesurface.The cavewidthvariesfrom3minthelowerleveltoabout8m intheupperlevelandthelengthisabout50m.Untilits discoveryduringroadconstructionabout12yearsago,the cavewasclosedwithnonaturalopening.Conditionstoday arearidandtheinteriorofthecaveisalmostdry,withonly alittlewaterdrippingmainlyduringthewintermonths.No recentspeleothemdepositionhasoccurred.However,the largenumberofspeleothemsinsidethecaveindicatesthat depositionwasintensiveinthepast. Analyticalmethods Forthepurposeof230Th-234Udating,15speleothems fromMaâ€™aleEfrayimCaveweresampled,includingboth stalagmitesandstalactites,andagesofonly9speleothems werewithintherangeofthedatingmethod.Thespeleothemsweresectionedperpendicularandparalleltotheir growthaxestoexposetheirgrowthlayers.Seriesoflaminae about0.5â€“2cmthickwereseparatedfromeachotherbya diamondsaw,resultinginsubsamplesof2 â€“20geach.An effortwasmadetopreventthecontaminationofthesubsamplesbynearbylaminae.AgedeterminationswereperformedbyalphaspectroscopyattheGeologicalSurveyof IsraelandbyTIMSattheOpenUniversity(UK),following theproceduresdescribedbyBar-Matthewsetal.(1997), Kaufmanetal.(1998),andMcDermottetal.(1999).The datingresultsaresummarizedinTable1. Twospeleothemswerechosenfordetailedcarbonand oxygenisotopicstudies.Fourgrowthlayersofthestalagmitewereisotopicallyanalyzed:foreachlayer,fouror ve pointswereanalyzed,anditwasevidentthattheywere depositedinisotopicequilibriumaccordingtoHendy â€™s (1971)criteria.Thecarbonandoxygenisotopicanalyses wereperformedonsamplesfromatransectalongthe growthaxis,withsamplingevery1.0mm.Isotopicanalyses ofcalcitespeleothemswereperformedonaSIRA-IImass spectrometerattheGeologicalSurveyofIsraelandare giventheusualnotation.All18Oand13Cvalueswere calibratedagainsttheinternationalstandardNBS-19andare reportedinper-mil(),relativetothePeeDeeBelemnite standard.Instrumentprecisionwasbetterthan 0.1,and184 A.Vaksetal./QuaternaryResearch59(2003)182â€“193
theexternalreproducibilitybasedonduplicatemeasurementsofreferencestandardwas 0.05. Acontinuousisotopicrecordwasobtainedunderthe assumptionsthatthemeasuredagerepresentsthecenterof thedatedlaminaandthatthegrowthratefromthecenterof thedatedlaminatothecenteroftheadjacentdatedlamina isconstant.Consequentlytheageofanisotopiceventis moreprecisefortimeintervalswithalargernumberof datedlaminae. AllraineventswerecollectedinMa â€™aleEfrayimvillage, 800mfromthecave,duringtworainyseasons,fromOctober1998toMay2000.Thecollectionofeachrainevent enablestheprecisedeterminationoftheamountanddistributionoftherain.Therainwateramountwasmeasuredby allowingwatertoaccumulateinalargefunnelanddripinto anarrow-headedbottleaccordingtothemethoddescribed inAyalonetal.(1998).CavewaterwassampledfromApril 1999toMay2000at10differentsiteswithinthecave. Oxygenisotopiccompositionsofrainandcavewaterwere determinedafterequilibrationwithCO2byshakinga2-ml watersamplefor4 â€“ 6hat25 C(EpsteinandMayeda, 1953).Hydrogenisotopiccompositionsweredetermined afterthewatersamplesreactedwithZnreagentat500 Cfor 50min(Colemanetal.,1982;Tanweeretal.,1988).Analyticalreproducibilityofduplicatesmeasuredondifferent daysisbetterthan0.1 for18Oand1.0 forD.18O/16O andD/HmeasurementsweremadeontheVGSIRA-IImass spectrometerattheGeologicalSurveyofIsraelandareTable1 U-Thanalyticalresults:Maâ€™aleEfrayimcave Sample Lamina238Uppb234U/238U230Th/234U230Th/232ThAge(103yr) StalactiteME-2 ME-2-A1 484 1.1215 0.05200.2840 0.0120245 36.0 1.9 ME-2-A2 409 1.0701 0.01780.3526 0.0072 22 47.0 1.3 ME-2-A3 317 1.0763 0.00960.4857 0.0058320 71.6 1.3 ME-2-B 501 1.0627 0.00880.7517 0.006712876 148.0 3.3 ME-2-C1 453 1.0723 0.01100.8170 0.01503344 177.9 9.1 StalactiteME-3 ME-3-A1 519 1.1141 0.00550.2596 0.0019319 32.5 0.3 ME-3-A2 505 1.0771 0.00410.3202 0.0050292 41.8 0.8 ME-3-B 444 1.0846 0.00420.3457 0.0025283 45.9 0.4 ME-3-C 366 1.1042 0.00740.3653 0.0027132 49.1 0.5 ME-3-D1 479 1.0865 0.00400.3980 0.0016139 54.8 0.3 ME-3-D2 415 1.0797 0.00520.4471 0.0026274 63.9 0.5 ME-3-E 426 1.0843 0.00350.7198 0.0034 27 134.7 1.4 ME-3-F 380 1.0968 0.00490.7796 0.0042260 158.0 2.2 StalactiteME-5 ME-5-A1 289 1.0768 0.01060.4563 0.0091 31 65.7 1.9 ME-5-A2 276 1.0918 0.01020.5182 0.0109 41 78.4 2.2 ME-5-B 300 1.0599 0.01000.7741 0.0167 19 157.8 8.6 StalactiteME-7 singlelamina408 1.1117 0.00100.3016 0.00211862 37.3 0.3 StalactiteME-8 singlelamina793 1.1036 0.00640.2915 0.00204513 38.8 0.3 StalagmiteME-12ME-12-A 690 1.0966 0.01930.1438 0.0026175 16.9 0.3 ME-12-B 366 1.1166 0.01600.2186 0.0033467 26.7 0.5 ME-12-C 249 1.0682 0.00440.2476 0.0024147 30.8 0.3 ME-12-D 235 1.0696 0.00320.2536 0.0016245 31.6 0.4 ME-12-E 293 1.1439 0.00620.2314 0.0020439 28.5 0.5 ME-12-F 186 1.0866 0.00350.2528 0.0025395 31.6 0.4 ME-12-G1 202 1.0827 0.00410.2538 0.0011209 31.9 0.2 ME-12-I1 226 1.0965 0.05030.2914 0.0154 98 37.3 2.4 ME-12-I2 270 1.0877 0.00510.2813 0.0033213 35.8 0.5 ME-12-I3 242 1.1013 0.00780.2967 0.0031151 38.1 0.5 ME-12-I4 146 1.0935 0.00440.3018 0.0022160 38.7 0.2 ME-12-J1 169 1.1052 0.00760.3592 0.0043 42 48.0 0.7 ME-12-J2 175 1.1028 0.00630.3424 0.0095 45 45.3 1.6 ME-12-J4 146 1.0816 0.00620.3931 0.0028 89 53.9 0.5 ME-12-J5 234 1.0720 0.00620.4214 0.0050223 59.1 1.0 ME-12-K3 182 1.0969 0.00490.4355 0.0058114 61.6 1.1 ME-12-K4 198 1.1321 0.00700.4138 0.0030 71 57.4 0.6 ME-12-K5 163 1.0998 0.00520.4429 0.0047147 63.0 0.9 ME-12-K7 232 1.0967 0.00460.4363 0.0033343 61.7 0.7 Alphaages StalactiteME-9 ME-9-A 2546 1.1942 0.01080.2954 0.0218 25 37.7 3.3 ME-9-B 1804 1.1401 0.01100.7693 0.01211008 152.0 5.5 StalagmiteME-12ME-12-G2 210 1.1249 0.05630.2440 0.0275 35 33.4 4.0 ME-12-H1 217 1.1071 0.04650.2954 0.0249 17 37.9 3.9 ME-12-H2 207 1.0833 0.03870.3130 0.0203 15 40.6 3.0 StalagmiteMME-1MME-1-A 481 1.4444 0.04200.7527 0.0241327 136.4 8.9 185 A.Vaksetal./QuaternaryResearch59(2003)182â€“193
giveninper-milnotationrelativetoSMOW(Craig,1961), calibratedusingtheViennaSMOWstandard. Theaveragecumulativeannualrainfallandtheaverage cumulativeisotopiccompositionwerecalculatedafter weighingtoallowfortheproportionoftheannualprecipitationcontributedbyeachraineventAyalonetal.,1998; (Bar-Matthewsetal.,1996).Theisotopiccompositionand theamountofrainfallintheMa â€™aleEfrayimregionand cavewaterwerecomparedtotherainandcavewaterof SoreqCavecollectedduringthesametimeperiods. Results Growthperiods TheresultsofthedatingareshowninTable1andare plottedinFig.2foreachspeleothem,inorderoftheyoungesttotheoldestlamina.Thedistributionshowsthatalmost alltheagesareinaccordwiththestratigraphicorder,except forseverallaminaefromonestalagmitewheresomeslight ageoverturnsweremeasured(Fig.2,sampleME-12).These overturnsmayhavebeencausedduringthesawingofthe samplebycontaminationfromthenearbyouteryounger laminae.Theagedistributionshowsthatthereareperiods duringwhichspeleothemsgrewinMa â€™aleEfrayimCave andperiodsduringwhichnogrowthoccurred,i.e.,periods whennoagesweredeterminedinanyofthespeleothems. Theagedistributionsforalleightspeleothems(Fig.2) showthatspeleothemgrowthoccurredduringthefollowing periods:glacialmarineoxygenisotopestage6,from 185,000to130,000yrB.P.;glacialstages4and3,from 80,000to25,000yrB.P.;andtheshortintervalneartheend ofthestage2LastGlacialMaximumfromabout19,000to 16,000yrB.P.Somespeleothemsareolderthan185,000yr B.P.butthemagnitudeoftheuncertaintiesprecludestheir condentinterpretation. Hiatuses(periodsofnospeleothemdeposition)areindicatedforthefollowingtimeintervals:thelastinterglacial, i.e.,marineisotopestage5from130,000to80,000yrB.P.; theLastGlacialMaximumatabout25,000to19,000yr B.P.;andduringtheendofstage2andthroughoutthe Holocenefromabout16,000yrB.P.untilthepresentday. Thehiatusesinspeleothemgrowthinferredfromtheages arealsodenedinthepetrographyofthespeleothemsby theoccurrenceofverysharpchangesincolorandthe presenceofthin(1-mm)whitemicriticlaminae,reddetrital-richlaminae,andpaleo-coraltites. TheceilingofMaâ€™aleEfrayimCaveischaracterizedby numeroussmallconicalstalactites(ME-7,ME-8,ME-9), whosetipsrangeinagebetween39,000and37,000yrB.P. Nosmallconicalstalactitesyoungerthanthesewerefound, suggestingthatthiswasaperiodwhenthecavewashighly active.Incontrast,agesofsimilarsmallconicalstalactites fromcavesonthewestern anksofthecentralridgeareof theHolocenetimeinterval,showingthatthosecaveswere activeforatleastpartoftheperiodduringwhichMa â€™aleEfrayimCavewasdry. Isotopiccompositionofpresent-dayrainandcavewaters18Oâ€“Drelationshipsofindividualrainfalleventsare plottedinFig.3andshowthattherainwaterinMa â€™aleEfrayimCavefallsbetweentheMMWLandtheglobal MeteoricWaterLine(MWL).Forcomparison,the18OandDvaluesoftherainfallandcavewaterattheSoreqsite followtheMMWLtrend(Ayalonetal.,1998;Bar-Matthewsetal.,2003)andshowsigni cantlylower18Oand theDvaluesthanthoseofMa â€™aleEfrayimrainandcave Fig.2.AgeswitherrorbarsofeightspeleothemsfromMa â€™aleEfrayim Caveinstratigraphicorder(fromtheyoungesttotheoldestlamina).Solid circlesmarktheTIMSages;opencirclesshowthealphaspectrometric ages.Theshadedareasmarkhiatuses. Fig.3.18Oâ€“DrelationshipsofallraineventsintheMaâ€™aleEfrayimsite andtheirweightedaverageforthehydrologicalyears1998 â€“1999and 1999 â€“2000.AlsoshownisMaâ€™aleEfrayimCavewatersandtheweighted average.TheweightedaverageSoreqrainandcavewatersfor1998 â€“2000 arealsoshown(afterBar-Matthewsetal.,2003).TheMediterranean MeteoricWaterLine(MMWL)andtheglobalMeteoricWaterLine (MWL)areshownforreference. 186 A.Vaksetal./QuaternaryResearch59(2003)182â€“193
water(Fig.3).Theaverageisotopiccompositionoftherain andcavewaterineachsiteispresentedinTable2. Duringthedryyearof1998 â€“1999theannualrainfall amountwasmuchbelowaverage,andhighaverage18O andDvaluescharacterizetherainwater.During1999 â€“ 2000,whentheamountofannualrainfallwasclosertothe meanannualaverage,18OandDvaluesoftherainfall werelower.However,theisotopiccompositionofMa â€™aleEfrayimCavewaterduringbothyearswassimilar(Table2, Fig.3)despitethelargedifferencesoftherainwatercompositioninthe2years.Thissuggeststhatthecavewater18OandDvaluesaretheaveragesofseveralyearsof rainfall,becauseofthewaterstorageintheepikarst. Therainfall18Ovaluesofindividualeventscandivided intotwodistinctgroups(Fig.4):thelightrainevents( 5 mm)rangingfrom 14to 5 andtherelativelymore massiverainevents(5mm)with18Ovaluesranging from0to 10.Thisissimilartothetrendobservedfor rainsinthewestern ankoftheridge(Ayalonetal.,1998; Gat,1996). IsotopiccompositionofMa â€™aleEfrayimspeleothems StalagmiteME-12,about42cmlong,wasdatedwith highresolution,showingcontinuousgrowthfrom67,000to 24,000yrB.P.withnopetrographicevidenceforhiatuses duringthisperiod.Thereforethissamplewasconsideredto bethebestforhigh-resolutionoxygenandcarbonisotopic studies.Asecondgrowthperiodfrom19,000to16,000yr B.Pisobservedinthissample.Betweenthesetwogrowth periodsthereisapetrographicallydistinctwhitelaminaand asharpchangeinthecolorofthelaminaefromwhiteto brown.StalactiteME-5,smallerinsize,wasdepositedbetween80,000and72,000yrB.P. TheoxygenisotopiccompositiondiagramshowninFig. 5aisacompositeprolebasedonstalagmiteME-12and stalactiteME-5.Theoverallpatternofthe18Oproleisof ageneralincreasein18Ovaluesfromabout 5.5 inthe oldestlaminatoapproximately 2 intheyoungest.Superimposedonthegeneraltrendarefrequentisotopicoscil-Table2 Weightedaverage18OandDvaluesofallrainfallandcavewaterduring1998 â€“2000intheMaâ€™aleEfrayimandSoreqcaves Years Rainfall (mm) Rain18O ( SMOW) RainD ( SMOW) Cave18O ( SMOW) CaveD ( SMOW) Maâ€™aleEfrayim1998â€“1999100 2.45 5.1 3.45 10.3 1999â€“2000250 5.41 25.1 3.6 15.8 Soreq 1998â€“1999198 4.43 15.0 4.1 12.1 1999â€“2000450 6.00 25.6 4.1 15.4 Fig.4.18OofraineventsvsrainfallamountattheMaâ€™aleEfrayimsite. Therainfall18Ovaluesofindividualeventscandividedintotwodistinct groups:thelightrainevents(5mm)rangingfrom 14to 5 andthe relativelymoremassiveraineventsrangingfrom0to 10. Fig.5.Superpositionofthe18O(a)and13C(b)recordsofMaâ€™ale EfrayimCave(thisstudy;black)andSoreqCave(gray)speleothems(after Bar-Matthewsetal.,1999),showingthestrikingsimilaritybetweenthetwo records.Thesquaresatthetopof(a)indicatethemeasuredTIMSages (opensquaresâ€”SoreqCave,solidsquaresâ€”Maâ€™aleEfrayimCave). 187 A.Vaksetal./QuaternaryResearch59(2003)182â€“193
lationsofapproximatelyonetoseveralhundredyearsduration(exceptfortheperiodsbetween46,000and40,000yr B.P.)characterizedbyamplitudeswingsaslargeas3 . Low-18Oeventsoccurat80,000yrB.P.( 5.3),63,000 yrB.P.(5.0),53,000and52,000yrB.P.(5.4), 36,000yrB.P.(4.8),and32,000yrB.P.(4.8). High-18Oeventsoccurat77,000yrB.P.( 3.6),72,000 yrB.P.(3.6),66,000yrB.P.(3.8),58,000yrB.P. (3.6),54,000yrB.P.(2.9),50,000yrB.P. (3.4),46,000yrB.P.(3.7),38,000yrB.P. (3.0),35,000yrB.P.(2.9),32,000yrB.P. (3.1),28,000yrB.P.(2.8),24,000yrB.P. (2.7),and19,000yrB.P.(2.1).Theseevents closelymatchtheisotopiceventsdeterminedforSoreq Cavespeleothems(Fig.5a). The13CproleofMaâ€™aleEfrayimspeleothems(Fig. 5b)showsatrendsimilartothatofthe18Oprole,witha slighttrendtowardhigher13Cvaluesovertime.Generally,13Cvaluesrangefromabout 11to 9.However, manyofthehigher13Cvaluesofabout 8 occurat 54,000and35,000yrB.P.,similartothetrendforSoreq Cave(Fig.5b).Superimposedonthegeneraltrendare frequentoscillationslastingonetoafewhundredyears.A 3 uctuationoccursataround54,000yrB.P.,fromthe lowestrecordedvaluesof 11to 8.0,thehighestvaluesobservedthroughouttheperiodofdeposition. Discussion Present-dayconditions Theisotopicvaluesofpresent-dayrainandcavewaters inMaâ€™aleEfrayimaregenerallyhigherthanthoseofSoreq Cave(Table2,Fig.3).Thistrendismoreevidentinthe drieryear1998 â€“1999thanintherelativelyrainyyear 1999 â€“2000(Fig.3).Duringthewintersof1998 â€“1999and 1999 â€“2000mostofthestormtracksbringingraintoboth cavesitesoriginatedfromtheeasternMediterraneanSea.If aRayleigh-typerainouteffectwasoperatingsolely(asbetweenthecoastandtheridge:Bar-MatthewsandAyalon, 2001),lessrainbutlowerisotopiccompositionwouldbe expectedontheeasternside.Theloweramountofrainfall andthehigher18OandDvaluesoftherainontheeastern sidecomparedwiththehigheramountofrainfallandthe lower18OandDvaluesoftherainonthewesternside thussuggestthatamajorpartoftherainfallevaporated beforereachingtheMa â€™aleEfrayimarea,duetotherain shadoweffectoftheridge.Thisviewissupportedbythe18Oâ€“Drelationshipsoftherainfall,whichfallbetween theMMWLandtheMWLandlie(Fig.3)onalowerslope deningevaporationtrends.Thesetrendsindicateevaporationbelowtheclouds(e.g.,Gat,1982,1996).Asimilar trendwasobservedbyRindsbergeretal.(1990)forrainfall atseveralstationsalongatransectfromthewesterntothe eastern ankofthecentralmountainridgeinIsrael.18OandDvaluesofMaâ€™aleEfrayimCavewaterhave similarcompositionandre ectatrendsimilartothatofthe localrainwater(Fig.3).Thisindicatesthatthesourceof Maâ€™aleEfrayimCavewatersisfromlocalprecipitation.The differencebetweenthe18Ovaluesoftherainfallatthe Maâ€™aleEfrayimCavesiteandthoseoftheSoreqrainfall duringthesameevents(b18OME-SR)isusuallypositive. The b18OME-SRdifferenceincreaseswithdecreasing amountofrainfall,bothduringasingleeventandduringthe wholerainyseason(12and15raineventsfor1998 â€“1999 and1999 â€“2000,respectively;Fig.6).Ayalonetal.(1998) showedthatrainfalleventsoflessthan20mmattheSoreq Cavesiteoccurmainlywhenairtemperaturesexceed10 C, resultinginevaporationbeneaththecloud.Thus,wesuggest thatthe2â€“3ChighersurfacetemperaturesattheMaâ€™ale EfrayimsiteduringrainfalleventsresultedinmoreevaporationthanthatattheSoreqsiteandincreased b18OME-SR. Onlyduringrelativelymassiverainfallevents(i.e.,above10 mmattheMaâ€™aleEfrayimsiteandabove20mmatthe SoreqCavesite;Fig.6),whenairtemperaturesinSoreqare between5 and10Candevaporationdecreases,doesthe b18OME-SRvaluefalltobetween2and 4.Under present-dayconditionsmostoftherainfallattheMa â€™ale Efrayimsiteundergoesevaporation.Inaddition,waterloss throughevaporationintheunsaturatedzone,evapotranspiration,andrunoffresultsinonlysmallamountsofdrip waterinthecaveandconsequentlytheabsenceofpresentdayspeleothemdeposition. Paleoclimaterecord Growthperiods SpeleothemdepositioninMa â€™aleEfrayimCaveoccurred duringmostglacialintervals,marineoxygenisotopicstages 2,3,4,and6(Fig.2),withnodepositionoccurringduring Fig.6.DifferencebetweenMaâ€™aleEfrayim(ME)rainfall18Ovaluesand Soreq(SR)rainfallvalues(b18OME-SR)vsrainfallamount(mm)in Maâ€™aleEfrayim,showingthatthe b18OME-SRincreaseswithdecreasing amountofrainfall.Thecirclesrepresentindividualrainevents,andthe squaresrepresenttheaverageofthe12raineventsin1998 â€“1999(solid square)andthe15raineventsduring1999 â€“2000(opensquare). 188 A.Vaksetal./QuaternaryResearch59(2003)182â€“193
interglacialstage5andtheHolocene.Inaddition,nodepositionoccurredduringtheheightoftheLastGlacialMaximum,fromabout25,000to19,000yrB.P.(Fig.2).Thus, waterwasnotavailableintheunsaturatedzoneduringthese periods,unlikethesituationonthewesternsideofthesame ridge,wherespeleothemdepositionwascontinuousduring thelast240,000yr(Bar-Matthewsetal.,2003). Previousstudiesofspeleothemsshowedthatincentral andnorthernIsrael,onthewesternsideofthecentralmountainridge,speleothemdepositionwascontinuousthrough bothglacialandinterglacialintervals.Theisotopiccompositionofspeleothemsrecordedbothglobalandregional paleoclimatechanges,becausetherainfalloriginatedin boththeAtlanticOceanandtheeasternMediterraneanSea (Ayalonetal.,2002;Bar-Matthewsetal.,1997,1999,2003; Frumkinetal.,1999). Maximumrainfallonthiswestern ankofthecentral mountainridgewassuggestedtooccurinperiodswhen sapropelswereformedintheeasternMediterraneanregion, mainlybetween128,000and120,000yrB.P.andduringthe earlyHolocenebetween8500and7000yrB.P.Kalleletal. (1997)showedthatthesalinityoftheentireMediterranean basinwashomogeneousduringsapropelevents.They claimedthatthisreectedmainlyenhancedrainfallthroughouttheMediterraneanbasin,becauseotherprocessesoccurringduringsapropelevents,suchasmeltwaterdischarge fromtheEuroasiancontinentalicesheetorenhancedNile discharge,wouldhaveresultedinstrongsalinitygradients. DuringthemiddleHolocenetheboundaryoftheNegev DesertinIsraelshifted20kmsouthofitspresentposition (GoodfriendandMagaritz,1988;Goodfriend,1991,1999). Goodfriend(1999,Fig.3)supportedthisclaimonthebasis oftheisotopiccompositionoflandsnails. Theassociationofextremelyhighspeleothem13Cwith minimum18Ovaluesbetween128,000and120,000yr B.P.(timeequivalenttosapropelS5)leadFrumkinetal. (2000)tosuggestthatthistimeintervalwashotanddry. Theyinterpretedtheveryhigh13Cvaluesofastalagmite fromacaveinJerusalemasre ectingcarbonaterockoutcropsthatwerestrippedofvegetationandsoilcoverdueto hotanddryconditions,possiblywiththeaidofforest res. Analternativeexplanationforthestronglyde nedcoupling oflow18Owithhigh13Cthatwasalsofoundforthistime intervalandduringthebeginningoftheHoloceneinspeleothemsfromSoreqCave(Bar-Matthewsetal.,2000, 2003)isthatitisassociatedwithenhancedrainfallduring sapropelintervals.Thehigh13Cvaluesarecausedby strippingofthesoilcoverduetodelugeevents,which resultedinwaterreachingthecaveafterlittleinteraction withsoilCO2.Evidenceforthisprocessisprovidedbythe higherpoollevelswithinSoreqCave(Ayalonetal.,2002; Bar-Matthewsetal.,2000,2003),thelargeamountofdetritalmaterialincorporatedinSoreqCavespeleothemsduringthesetimeintervals(Ayalonetal.,1999),andthesharp dropintheconcentrationsofSr,Ba,andUandintheratios of(234U/238U)0and87Sr/86Sr,whichreectenhanced weatheringofthehostrock(Ayalonetal.,1999;BarMatthewsetal.,2000;Kauffmanetal.,1998).Apresentdayanalogisthehigh13Cvaluethatoccursduringmassive rainstorms,wherebythe13Cvaluesofdissolvedinorganic carboninwaterinltratingthecavethroughthinrockcover havemuchhigher13Cvaluesofca. 5,comparedto valuesof 14to 10 observedduringnormalrainfall events(Bar-Matthewsetal.,1996). Theoccurrenceof â€œwetâ€ easternMediterraneanconditionsinwhichrainfalloccursonthewestern anksofthe centralmountainridgeofIsrael,butisabsentontheeastern anks,isconsistentwiththerainshadoweffectoftheridge, asseenatthepresentday,leadingtothequestions:under whatconditionsdoesthiscomeaboutandwhyweremostof theglacialperiods â€œwet?â€ Thepresentdayisclearlyausefulanalogsincethereis nowasharpcontrastbetweentheconditionsonthewestern andthoseontheeastern anksofthecentralmountain ridge.Undersuchconditionsverylittlewaterreachesthe unsaturatedzoneintheMa â€™aleEfrayimregion,causingthe cavetoremaindry.Thisnegativewaterbalanceisduetothe highevaporationtoprecipitationratio,wherebyeffective precipitationisextremelylow.Weinferthatinterglacial conditionsweresimilarandthathighevaporationtoprecipitationratiosledtoaridconditions.Duringtheglacialperiods,whentemperatureswerecolder,evaporationwas lower,andtheeffectiveprecipitationincreased,allowing watertoresideintheunsaturatedzone.Consequently,speleothemdepositionoccurredinMaâ€™aleEfrayimCave.A similarscenariowassuggestedtoexplainspeleothemdepositioninsoutheastAustraliaduringglacialperiods(Ayliffe etal.,1998).TheexistenceofLakeLisan(amuchlarger anddeeperlatePleistoceneprecursoroftheDeadSea) duringthelastglacialinterval(e.g.,Bartovetal.2002; Beginetal.,1974;NeevandEmery,1967)isastrong indicatorthatrelativelyhumidconditionsprevailedonboth sidesofthecentralridgeduringglacialintervals. Theshorthiatusduringthelastglaciation,betweenabout 24,000and19,000yrB.P.,representstheariditythatdevelopedintheareaprobablyasaresultofasharpdropin therainfallthen(Bar-Matthewsetal.,1997),whichledalso toasuddendropinthelevelofLakeLisan(Bartovetal., 2002)andisevidentintheRedSea(Almogi-Labinetal., 1986). Isotopicrecord TheisotopicprolesofMaâ€™aleEfrayimCaveandSoreq Cavespeleothemsforthistimeintervalarecomparedin Fig.5.Thereisaverygoodcorrespondenceinthegeneral trendofbothrecords.Basedonthegoodmatchingbetween thewell-datedpartsoftherecords,we â€œwiggle-matchedâ€ theMaâ€™aleEfrayimCaverecordwiththeSoreqCaverecord forthosetimeintervalsforwhichonlyafewagedeterminationsweremade.Theseperiodsare80,000 â€“72,000, 67,000 â€“ 63,000,30,000 â€“25,000,and19,000 â€“16,000yr B.P.Thegapbetween72,000and67,000yrB.Pisprobably189 A.Vaksetal./QuaternaryResearch59(2003)182â€“193
duetolackofsampling.Suchsimilarityofspeleothem recordsfromtwocaveslocated 60kmawayfromeach otherondifferentsidesofthecentralmountainridgeandin differentpresent-dayclimaticconditionscon rmsourinferencethatthesourcefortherainfallfromwhichthespeleothemsacquiredtheirisotopicsignalwasthesameand,as shownbyBar-Matthewsetal.(2003)fortheSoreqand Peqâ€™iincaves,thissourcewastheeasternMediterranean Sea.Moreover,thesimilaritybetweentheiroxygenisotopic recordtrendssuggeststhat,unlikethepresentday,the climateinbothsitesduringthistimeintervalwasgenerally similar.Thematchingalsosupportsthissuggestionobservedbetweenthe13Cisotopicprolesofthespeleothems frombothcaves(Fig.5b),whichindicatesthatbothwere coveredbysimilarvegetation.However,from46,000to 36,000yrB.P.andfrom30,000to16,000yrB.P.,the13C valuesofMaâ€™aleEfrayimCavespeleothemswerehigherby 0.4 thanthoseofSoreqCavespeleothems,possibly becauseduringtheseperiodsthevegetationinMa â€™ale EfrayimcontainedaslightlyhigherproportionofC4type vegetationthanthatinSoreqand/orbecauseofdifferences inthesoilproductivityatthetwosites(Frumkinetal., 2000). The18OvaluesofMaâ€™aleEfrayimCavespeleothems areloweronaverageby 0.5 thanthoseofSoreqCave speleothems(Fig.5a;integratedaveragesare 4.06 0.54 and 3.53 0.45,respectively).Tworeasonscanexplainthis0.5 difference;lower18Oofrainfallatthe Maâ€™aleEfrayimsiteand/orslightlyhighertemperaturesin whichspeleothemsweredepositedattheMa â€™aleEfrayim site.Underpresent-dayconditions,therainattheMa â€™ale Efrayimsiteis18O-enrichedrelativetothatattheSoreqsite, duetotheevaporationprocessesbroughtaboutbythesharp increaseintemperaturesontheeasternsideofthemountain ridgeandthecorrespondingrainfall/temperatureeffectof 0.4/C(Rozanskietal.,1993).Thus,itispossiblethat duringglacialintervals,whentemperatureswerecolder, theseeffectswerenegligibleandtherainwater18Ovalues becamelowerastherainpassedoverthecentralmountain ridgeandbecameisotopicallydepletedthroughRayleigh fractionationprocesses(rainouteffect;Gat,1996).This alternativeissuggestedbythepresent-dayanalogue,showingthatwhenrelativelymassiverainfalleventsoccur,and airtemperatureisbelow10C,the b18OME-SRislowand evennegative(Fig.6).Thesmaller b18OME-SRofthe rainieryear(Fig.6)alsosupportsthispossibility.Analternativepossibilityisthatthe18Oofrainwatersdidnot changeovertheridgeandthatthelower18Ovaluesinthe Maâ€™aleEfrayimspeleothemsindicatetemperatures2 â€“3C higherattheMaâ€™aleEfrayimsitethanattheSoreqsite.In bothcases,lowevaporationisafundamentalfactorcontrollingtheavailabilityofwaterintheepikarstsystem. Duringtheshorttimeintervalsbetween19,000and 16,000yrB.P.,atthetransitionfromglacialtointerglacial conditions,thegeneralpatternchangesand18Ovaluesof Maâ€™aleEfrayimCavespeleothemsareonaverage 0.5 higherthanthoseofSoreqCave.Thispositive 0.5 differenceisthatwhichwouldbecalculatedifdeposition werehypotheticallyoccurringintheMa â€™aleEfrayimCave underpresent-dayconditions.The18Ovaluescalculated forspeleothemsforminginequilibriumfromtheMa â€™ale EfrayimCavewatersare 5.0 (calculatedbysubstituting theaveragecavewater18Ovalueof 3.5 andanaveragecavetemperatureof21CintotheOâ€™Neiletal.(1969) calciteâ€“waterfractionationequation).Thisvalueishigher by 0.4 thantheaverage18Ovalueof 5.4 ofmodernspeleothemsfromSoreqCave,whichprecipitateat averagecavetemperatureof18 Cfromcavewaterwithan average18Ovalueof 4.7 (Bar-MatthewsandAyalon, 2003).Thehigherhypothetical18Ovalueofpresent-day speleothemsinMaâ€™aleEfrayimCavereectswatersthatare isotopicallyenrichedthroughevaporation;highertemperatureswouldcausearelative18Oloweringinthespeleothems.Similarly,their13Cvaluealsowouldhavebeen higherthanthatofSoreqCavespeleothems,becauseofthe dominanceofC4vegetationintheMa â€™aleEfrayimarea. Thus,thetransitionfromcolderglacialtowarmerinterglacialintervalsintheMaâ€™aleEfrayimareainvolvedbotha temperatureincreaseandthedevelopmentofaridityonthe easternsideofthecentralmountainridge.Asimilartrend andenrichmentof0.5 in18Oisalsoobservedforanotherspeleothem(ME-5)depositedbetween80,000and 72,000yrB.P.However,sampleME-5wascollectedfrom adifferentsitewithinthecave,anditmayre ectlocal variationinthetemperatureandthedripisotopiccomposition. ThecomparisonoftheisotopicrecordsofMa â€™ale EfrayimCaveandSoreqCaveshowsthatduringcoolglacialconditions,whentemperatureswere 6 â€“10Clowerin theentireeasternMediterraneanregion(Emeisetal.,2000), theevaporationwasmuchlesssigni cant,theamountof rainfallatbothsiteswasmoresimilar,andthedifferencesin theisotopiccompositionsofrain,cavewater,andspeleothemsreecteitherthe2â€“3Cgradientbetweenthesites orthelower18Ovaluesoftherain.Thesimilar13Cvalues (Fig.5b)showthat,unlikepresentday,thevegetationat bothsiteswasalmostthesame,suggestingalsothatthe climateduringcoolglacialintervalswassimilaronboth sidesoftheridgeandthe uctuationsin13Cvalueswere causedbythesamereasons(changesintheC3/C4proportions)aswassuggestedforSoreqCavespeleothems(BarMatthewsetal.,1997).Increasingtemperaturesduringthe transitionfromcoolglacialintervalstowarminterglacial conditionswereassociatedwithadecreaseintherainfall amountandanincreaseinitsisotopiccomposition.Asa resultthespeleothem18OvaluesbecamehigheratMaâ€™ale EfrayimCavethanatSoreqCave(Fig.5a).Astemperatures increased,theeffectiveprecipitationdecreased,causing verylittlewatertoresideintheunsaturatedzone,and eventuallyspeleothemdepositionceasedinMa â€™aleEfrayim Cave. Thenatureoftheoxygenandcarbonisotopic uctua-190 A.Vaksetal./QuaternaryResearch59(2003)182â€“193
tions(the â€œspikesâ€ and â€œlowsâ€)inMa â€™aleEfrayimCavecan beexplainedinthelightoftheSoreqCavespeleothem oscillationsduringthelastglaciation(Bar-Matthewsetal., 1999).Theyusuallycorrelatewithglobalandregionalclimaticevents,suchasHeinricheventsandtheLastGlacial Maximum.Becauseofthesimilarityintimeandamplitudes oftheisotopiceventsinMa â€™aleEfrayimCaveandSoreq Cavespeleothemswesuggestthattheyrecordthesame events. Growthrates Todeterminethegrowthrate,wechosesampleME-12, alargeinsitustalagmitethatgrewduringmostofthelast glacialandwasdatedwithveryhighresolution(Table1). GrowthratevariationsfromtheMa â€™aleEfrayimandSoreq cavescloselymatchthe uctuationsin18Oand13Cvalues(Fig.7),inwhichperiodswithrelativelyhighgrowth ratearecharacterizedbyrelativelylow18Oand13C values.Thisismostpronouncedat 60,000,50,000and 34,000yrB.P.Thetrendtowardslowergrowthratesusually followsperiodsshowingincreasein18Oand13Cvalues. Wesuggesttherefore,thathigherisotopicvalueswithlower growthratesreectperiodswithdecreasingrainfalland consequentlydecreasesintheeffectiveprecipitationandthe amountofwaterreachingtheunsaturatedzone.Thesimilarityingrowthrateatthetwosites(Fig.7c)furthersupportstheargumentthattherainshadoweffectdidnotexist duringglacialconditions. Conclusions PeriodsofspeleothemdepositioninMaâ€™aleEfrayim Cave,locatedintherainshadowontheeastern ank ofthecentralmountainridge,occurredduringmarine isotopicglacialconditionsduringperiodsofincreased effectiveprecipitation.Speleothemdepositiondidnot occurininterglacialperiods.Rainfallduringearly stage2at25,000 â€“19,000yrB.P.,theheightofthe LastGlacialMaximum,appearstohavebeensuf cientlylowthatspeleothemdepositionceasedthen. Waterlossthroughevaporation,evapotranspiration, andrunoffresultsinnegligibleamountsofdripwater inMaâ€™aleEfrayimCaveandconsequentlytheabsence ofpresent-dayspeleothemdeposition.Theabsenceof speleothemgrowthduringinterglacialperiodsisinterpretedtoreecthydrologicconditionssimilarto thoseofthepresentday,withthedesertboundaryin thesameposition. Fig.7.Thecalculatedgrowth-rate(black)ofaMaâ€™aleEfrayimCavespeleothem(stalagmiteME-12)superimposedonthe18O(a)and13C(b)proles (gray).Fastestgrowthratescoincidewithminimum18Oand13Cvalues.In(c)thegrowthrateoftheSoreqCavespeleothems(thinline)issuperimposed onthatoftheMaâ€™aleEfrayimCavespeleothem(thickline). 191 A.Vaksetal./QuaternaryResearch59(2003)182â€“193
RainfallintheMaâ€™aleEfrayimareaoriginatesinthe easternMediterraneanSea.However,the18Oâ€“D relationshipsofbothrainfallandcavewatersfallbetweentheMMWLandtheMWLandde neevaporationtrends.Underpresent-dayconditions,therainat theMaâ€™aleEfrayimCavesiteismoreenrichedin18O thanthatattheSoreqCavesiteonthewestern ank, duetotheevaporationprocessesbroughtaboutbythe sharpincreaseintemperaturesontheeasternsideof themountainridge. Thegeneral18Oand13CtrendsofMaâ€™aleEfrayim Cavespeleothemsdepositedduringthelastglacial intervalissimilartothoseoftheSoreqCavespeleothemlocatedonthewestern ankofthecentral mountainridge, 60kmaway.Thisindicatesthat, unlikethepresentday,theclimateonbothsidesofthe ridgeduringthelastglacialintervalwassimilar,and thesourcefortherainfallfromwhichthespeleothems acquiredtheirisotopicsignalinbothsiteswasthe easternMediterraneanSea.The18OvaluesofMaâ€™ale EfrayimCavespeleothemsdepositedduringthelast glaciationareonaverage 0.5 lowerthanthoseof SoreqCavespeleothems,becausethetemperaturesin Maâ€™aleEfrayimwereslightlyhigherand/orthe18O ofrainfall,associatedwithhigherregionalprecipitation,waslower. Increasingtemperaturesduringthetransitionfrom coolglacialintervalstowarminterglacialconditions resultedintherelativeincreaseofboth18Oand13C valuesofMaâ€™aleEfrayimCavespeleothemscomparedwiththoseofSoreqCavespeleothems,dueto decreaseintherainfallamountandincreaseinthe rainfallisotopiccomposition.Thus,theeffectofthe rainshadowbecomesdominantonlyduringwarm interglacialclimateconditions,similartothepresent day. The uctuationsinMaâ€™aleEfrayimCavespeleothem growthratescloselymatchthe18Oand13C uctuations,withrelativelyhighergrowthratesbeingassociatedwithrelativelylower18Oand13Cvalues, thuspointingtothecorrelationbetweenlowerisotopic valuesandincreasingamountsofrainfallandeffective precipitation.ThesimilarityincontemporaryspeleothemgrowthratesattheMa â€™aleEfrayimandSoreq cavesitesareconsistentwiththeabsenceofasignificantrainshadowduringglacialperiodsandaconsequentmigrationofthedesertboundary. Acknowledgments ThisresearchwassupportedbyTheIsraelScienceFoundation(GrantNo.151/98-13.0andNo.20/01-13.0).We expressourgratitudetotheIsraelNatureandNationalParks ProtectionAuthorityfortheircooperation.ToIrenaSegal forhelpwithU-ThmeasurementsandtoYehudaPeled, ShlomoAshkenazi,EliRam,RamiSolomon,MarkTzepelevich,andMeiravDinorfortheirhelpinthe eldwork. ReferencesAlmogi-Labin,A.,Luz,B.,Duplessy,J.,1986.Quaternarypaleo-oceanography,pteropodpreservationandstableisotoperecordoftheRed Sea.Palaeogeography,Palaeoclimatology,Palaeoecology57,195â€“211. Ayalon,A.,Bar-Matthews,M.,Sass,E.,1998.Rainfall-rechargerelationshipswithinakarsticterrainintheEasternMediterraneansemi-arid region,Israel,18OandDcharacteristics.JournalofHydrology207, 18 â€“31. Ayalon,A.,Bar-Matthews,M.,Kaufman,A.,2002.Climaticconditions duringmarineoxygenisotopestage6intheeasternMediterranean regionfromtheisotopiccompositionofspeleothemsofSoreqCave, Israel.Geology30,303â€“306. Ayliffe,L.K.,Marianelli,P.C.,Moriarty,K.C.,Wells,R.T.,McCulloch, M.T.,Mortimer,G.E.,Hellstrom,J.C.,1998.500kaprecipitation recordfromsoutheasternAustralia:evidenceforinterglacialrelative aridity.Geology26,147â€“150. Bar-Matthews,M.,Ayalon,A.,2001.EasternMediterraneanpaleoclimate duringthelast250,000yearsasderivedfromthepetrography,mineralogy,traceelementandisotopiccompositionofcavedeposits(speleothems),Israel.GeologicalSurveyofIsraelReportGSI/41/01,44pp. Bar-Matthews,M.,Ayalon,A.,2003.Speleothemsaspaleoclimateindicators,acasestudyfromSoreqcavelocatedintheEasternMediterraneanregion,Israel,in:Battarbee,R.W.,Gasse,F.,Stickly,C.E. (Eds.),PastClimateVariabilitythroughEuropeandAfrica,Kluwer AcademicPublisher(inpress). Bar-Matthews,M.,Ayalon,A.,Matthews,A.,Sass,E.,Halicz,L.,1996. Carbonandoxygenisotopestudyoftheactivewater-carbonatesystem inthekarsticMediterraneancave:implicationsforpaleoclimateresearchinsemiaridregions.GeochimicaetCosmochimicaActa60, 337â€“347. Bar-Mattews,M.,Ayalon,A.,Kaufman,A.,1997.LateQuaternarypaleoclimateintheEasternMediterraneanRegionfromstableisotopeanalysisofspeleothemsinSoreqcave,Israel.QuaternaryResearch47, 155â€“168. Bar-Matthews,M.,Ayalon,A.,Kaufman,A.,Wasserburg,G.J.,1999.The EasternMediterraneanpaleoclimateasareectionofregionalevents: Soreqcave,Israel.EarthPlanetaryScienceLetters166,85â€“95. Bar-Matthews,M.,Ayalon,A.,Kaufman,A.,2000.TimingandhydrologicalconditionsofSapropeleventsintheEasternMediterranean,as evidentfromspeleothems,Soreqcave,Israel.ChemicalGeology169, 145â€“156. Bar-Matthews,M.,Ayalon,A.,Gilmour,M.,Matthews,A.,Hawkesworth, C.J.,2003.Sea-landoxygenisotopicrelationshipsfromplanktonic foraminiferaandspeleothemsintheEasternMediterraneanregionand theirimplicationforpaleorainfallduringinterglacialintervals. GeochimicaetCosmochimicaActa(inpress). Bartov,Y.,Stein,M.,Enzel,Y.,Agnon,A.,Reches,Z.,2002.Lakelevels andsequencestratigraphyofLakeLisan,thelatePleistoceneprecursor oftheDeadSea.QuaternaryResearch57,9 â€“21. Begin,Z.B.,Ehrlich,A.,Nathan,Y.,1974.LakeLisan,thePleistocene precursoroftheDeadSea.GeologicalSurveyofIsraelBulletin63,30. Coleman,M.L.,Shepherd,T.J.,Durham,J.J.,Rouse,J.E.,Moore,G.R., 1982.Reductionofwaterwithzincforhydrogenisotopeanalysis. AnalyticalChemistry54,993â€“995. Craig,G.,1961.Isotopicvariationsinmeteoricwaters.Science133, 1702â€“1703. Dansgaard,W.,1964.Stableisotopesinprecipitation.Tellus16,438 â€“ 468. Emeis,K.C.,Struck,U.,Schulz,H.M.,Rosenberg,R.,Bernasconi,S., Erlenkeuser,H.,Sakamoto,T.,Martinez-Ruiz,F.,2000.Temperature andsalinityvariationsofMediterraneanSeasurfacewatersoverthe last16,000yearsfromrecordsofplanktonicstableoxygenisotopesand 192 A.Vaksetal./QuaternaryResearch59(2003)182â€“193
alkenoneunsaturationratios.PalaeogeographyPalaeoclimatology Palaeoecology158,259 â€“280. Epstein,S.,Mayeda,T.K.,1953.Variationsof18Oofwatersfromnatural sources.GeochimicaetCosmochimicaActa4,213â€“224. Frumkin,A.,Ford,D.C.,Schwarcz,H.P.,1999.Continentaloxygenisotopicrecordofthelast170,000yearsinJerusalem.QuaternaryResearch51,317â€“327. Frumkin,A.,Ford,D.C.,Schwarcz,H.P.,2000.PaleoclimateandvegetationofthelastglacialcyclesinJerusalemfromaspeleothemrecord. GlobalBiochemicalCycles14,863â€“ 870. Gascoyne,M.,Schwarcz,H.P.,Ford,D.C.,1982.Uraniumseriesagesof speleothemfromNorthwestEngland:correlationwithQuaternaryclimate.PhilosophicalTransactionsRoyalSocietyofLondonB-301, 143â€“164. Gat,J.R.,1982.Precipitation,groundwaterandsurfacewaters,in:PaleoclimatesandPaleowaters.InternationalAtomicEnergyAgency,Vienna,pp.3â€“12. Gat,J.R.,1996.Oxygenandhydrogenisotopesinthehydrologiccycle. AnnualReviewsEarthPlanetaryScience24,225â€“262. Gat,J.R.,Carmi,I.,1970.EvolutionintheisotopiccompositionofatmosphericwatersintheMediterraneanSeaarea.JournalGeophysical Research75,3039 â€“3048. Gat,J.R.,Carmi,I.,1987.Effectofclimatechangesontheprecipitation patternsandisotopiccompositionofwaterinclimatictransitionzone: caseofEasternMediterraneanseaarea,in:TheinuenceofClimatic ChangeClimaticVariabilityontheHydrologicRegimeandWater Resources,IANS168Symp.Proc.pp.513â€“523. Goodfriend,G.A.,1991.Holocenetrendsin18Oinlandsnailshellsfrom theNegevDesertandtheirimplicationsforchangesinrainfallsource areas.QuaternaryResearch35,417â€“ 426. Goodfriend,G.A.,1999.TerrestrialstableisotoperecordsofLateQuaternarypaleoclimatesintheeasternMediterraneanregion.Quaternary ScienceReviews18,501â€“513. Goodfriend,G.A.,Magaritz,M.,1988.PalaeosolsandlatePleistocene rainfall uctuationsintheNegevDesert.Nature332,144 â€“146. Gordon,D.,Smart,P.I.,Ford,D.C.,Andrew,J.N.,Atkinson,T.C.,Rowe, P.J.,Christopher,N.S.J.,1989.DatingofLatePleistoceneinterglacial andinterstadialperiodsintheUnitedKingdomfromspeleothem growthfrequency.QuaternaryResearch31,14 â€“26. Hendy,C.H.,1971.TheisotopiccompositionofthespeleothemsI.The calculationoftheeffectsofdifferentmodesofformationonthe isotopiccompositionofspeleothemsandtheirapplicabilityaspaleoclimaticindicators.GeochimicaetCosmochimicaActa35,801â€“ 824. Kallel,N.,Paterne,M.,Duplessy,J.-C.,Vergnaud-Grazzini,C.,Pujol,C., Labeyrie,L.,Arnold,M.,Fontugne,M.,Pierre,C.,1997.Enhanced rainfallintheMediterraneanregionduringthelastsapropelevent. OceanologicaActa20,697â€“712. Kaufman,A.,Wasserburg,G.J.,Porcelli,D.,Bar-Matthews,M.,Ayalon, A.,Halicz,L.,1998.U-ThisotopesystematicsfromtheSoreqCave Israelandclimaticcorrelations.EarthPlanetaryScienceLetters156, 141â€“155. Lauritzen,S.E.,1995.High-resolutionpaleotemperatureproxyrecordfor thelastinterglaciationbasedontheNorwegianspeleothems.QuaternaryResearch43,133â€“146. McDermott,F.,Frisia,S.,Huang,Y.,Longinelli,A.,Spiro,B.,Heaton, T.H.E.,Hawkesworth,C.J.,Borsato,A,Keppens,E.,Fairchild,I.J., Borg,K.,Verheyden,S.,Selmo,E.,1999.Holoceneclimatevariability inEurope:evidencefrom18O,texturalandextensionratevariationsin threespeleothems.QuaternaryScienceReviews18,1021â€“1038. Neev,D.,Emery,K.O.,1967.TheDead-Sea;depositionalprocessesand environmentsofevaporites.GeologicalSurveyofIsraelBulletin41, 147pp. Oâ€™Neil,J.R.,Clayton,R.N.,Mayeda,T.K.,1969.Oxygenisotopefractionationofdivalentmetalcarbonates.JournalofChemicalPhysics30, 5547â€“5558. Rindsberger,M.,Jaffe,Sh.,Rahamin,Sh.,Gat,J.R.,1990.Patternsofthe isotopiccompositionofprecipitationintimeandspace:datafromthe Israelistormwatercollectionprogram.Tellus42B,263â€“271. Rozanski,K.,Aragua s-Aragua s,L.,Gonantini,R.(1993).Isotopicpatternsinmodernglobalprecipitation.ClimateChangeinContinental IsotopicRecords,GeophysicalMonograph78,AmericanGeophysical Union,1â€“36. Schwarcz,H.P.,1986.Geochronologyandisotopicgeochemistryofspeleothems,in:Fritz,P.,Fontes,J.C.(Eds.),HandbookofEnviromental IsotopeGeochemistry,Vol.2.Elsevier,Amsterdam,pp.271â€“300. Tanweer,A.,Hut,G.,Burgman,J.O.,1988.Optimalconditionsforthe reductionofwatertohydrogenbyzincformassspectrometricanalysis ofthedeuteriumcontent.ChemicalGeology(IsotopeGeoscienceSection)73,199 â€“203. 193 A.Vaksetal./QuaternaryResearch59(2003)182â€“193
PaleoclimateandlocationoftheborderbetweenMediterranean climateregionandtheSaharo â€“ ArabianDesertasrevealedby speleothemsfromthenorthernNegevDesert,IsraelA.Vaksa,b,,M.Bar-Matthewsb,A.Ayalonb,A.Matthewsa,A.Frumkinc,U.Dayanc, L.Haliczb,A.Almogi-Labinb,B.SchilmanbaInstituteofEarthSciences,HebrewUniversityofJerusalem,Jerusalem91904,IsraelbGeologicalSurveyofIsrael,30MalcheiIsraelSt,Jerusalem95501,IsraelcDepartmentofGeography,HebrewUniversityofJerusalem,Jerusalem91905,Israel Received9January2006;receivedinrevisedform23May2006;accepted10July2006 Availableonline30August2006 Editor:S.KingAbstract SpeleothembearingkarsticcavesofthenorthernNegevDesert,southernIsrael,providesanidealsiteforreconstructingthe paleoclimateandpaleo-locationoftheborderbetweenMediterraneanclimateregionandtheSaharo â€“ ArabianDesert.Majorperiodsof speleothemdeposition(representinghumidperiods)weredeterminedbyhighresolution230Th â€“ Udatingandcorrespondingstudiesof stableisotopecompositionwereusedtoidentifythesourceofrainfallduringhumidperiodsandthevegetationtype.Majorhumidintervals occurredduringglacialsat190â€“ 150ka,76â€“ 25ka,23 â€“ 13kaandinterglacialsat200 â€“ 190ka,137â€“ 123kaand84 â€“ 77ka.Thedominant rainfallsourcewastheEasternMediterraneanSea,withapossiblesmallcontributionfromsoutherntropicalsourcesduringtheinterglacial periods.WhentheinterglacialintervalrainfallwasofEasternMediterraneanorigin,theminimumannualrainfallwas 300â€“ 350mm; approximatelytwicethanofthepresent-day.Lowerminimumamountsofprecipitationcouldhaveoccurredduringglacialperiods,dueto thecoolertemperaturesandreducedevaporation.Althoughduringmostofthehumidperiodsthevegetationremainedsteppewithmixed C3+C4vegetation,MediterraneanC3typesteppe-forestvegetationinvadedsouthwardforshortperiods,andtheclimateinthenorthern NegevbecameclosertoMediterraneantypethanatpresent.Theclimatewassimilartopresent,orevenmorearid,duringintervalswhen speleothemdepositiondidnotoccur:150 â€“ 144ka,141â€“ 140ka,117â€“ 96ka,92 â€“ 85ka,25 â€“ 23ka,and13kaâ€“ present-day. PrecipitationincreaseoccurredinthenorthernNegevduringtheinterglacialmonsoonalintensitymaximaat198ka,127ka, 83kaandglacialmonsoonalmaximaat176ka,151ka,61kaand33ka.However,duringinterglacialmonsoonalmaximaat 105kaand11ka,thenorthernNegevwasaridwhereasduringglacialmonsoonalminimaitwasusuallyhumid.Thisimpliesthat thereisnotalwayssynchroneitybetweenmonsoonalactivityandhumidityintheregion. OxygenisotopicvaluesofthenorthernNegevspeleothemsaresystematicallylowerthancontemporaneousspeleothemsofcentral andnorthernIsrael.ThispartisattributedtotheincreasedrainoutoftheheavyisotopesbyRayleighfractionationprocesses,possiblydue tothefartherdistancefromtheMediterraneancoast. 2006ElsevierB.V.Allrightsreserved.Keywords: Speleothems;Paleoclimate;NegevDesert;NorthernSaharoâ€“ ArabianDesert;230Th â€“ Uages EarthandPlanetaryScienceLetters249(2006)384 â€“ 399 www.elsevier.com/locate/epsl Correspondingauthor.Tel.:+97225314343;fax:+97225380688. E-mailaddress: email@example.com (A.Vaks). 0012-821X/$-seefrontmatter2006ElsevierB.V.Allrightsreserved. doi: 10.1016/j.epsl.2006.07.009
1.Introduction AkeysubjectofpaleoclimatestudiesisthereconstructionofclimatechangesoftheSaharoâ€“ Arabiandesertbelt ( Fig.1A),andhowthisclimatewasaffectedbyglobal events.Atpresent,theSaharo â€“ ArabianDesertisinfluencedbytwomainclimatesystems.Oneassociatedwith polarfrontsoriginatinginthenortheastAtlanticOceanand producingseriesofcyclonesthatpassoverWesternEurope beforereachingtheMediterraneanSea.Theabsenceof topographicbarriersallowstheseAtlanticlow-pressure cellstoprogresseastwardsabovethewarmMediterranean Sea,whichprovidesasecondarymoisturesourcefor winterrainfall ,thatmainlyaffectsthenorthernSaharoâ€“ ArabianDesert(e.g., [2 â€“ 4] ).Thesecondsystemoriginates inthetropicalAtlanticandIndianOceans,andis manifestedasthelowlatitudemonsoonthatmainlyaffects thesouthernpartsofSaharo â€“ ArabianDesert(e.g., ). DuringtheMiddle â€“ LateQuaternary,thesouthern boundaryoftheSaharo â€“ ArabianDesertshiftedseveral hundredkilometersnorthofitspresentlocation,as demonstratedbytheformationoflacustrinesediments [6,7] ,travertines,uraniumoresandspeleothems [8,9] . Theiroccurrenceindicatesthatwetterclimateconditions prevailedduringpartofpreviousinterglacialperiods andduringtheearlyHolocene  .Wetperiodsalso occurredinthesouth-easternSaharaduringseveral glacialinterstadials.Allthesehumidepisodesresulted fromthenorthwardmigrationoftheAfricanandIndian monsoonsystems [8,10,11,12] .Duringthelast200ka theperiodicityofthewetphaseswasusuallydrivenby the23kaprecessioncycle . PaleoclimatedataonthenorthernborderofSaharo â€“ ArabianDesertissignificantlymorelimited,particularlyin theEasternMediterranean(EM)region.DatafromEgypt, TunisiaandMorocco [13 â€“ 15] showthatwetterconditions occurredduringtheearlyHolocene.Lakesedimentsin southernJordanindicatethathumidconditionsprevailed duringpreviousinterglacialmarineisotopicstages(MIS) 5.5and7.1,andprobablyalsoduringglacialinterstadials inMIS-6  .Themajorsourceofpaleoclimaticdataon thenorthernSaharo â€“ ArabiandesertintheEMareacomes Fig.1.Geographical,isohyetandlocationmapsofthestudyarea.(A)MapindicatingtheextentoftheSaharo â€“ Arabiandesert(ingray).Therectangle markstheresearcharea.(B)PrecipitationmapofIsraelandadjacentlands:PalestinianAuthorityinGaza(PA),north-easternEgypt(easternSinai Peninsula),westernJordan,south-westernSyriaandsouthernLebanon.Isohyetsareindicatedbyblacklinesandpresentbordersbygreylines,the villageofNeve â€“ Ativ(NA)isshownbyawhitecircleandPeqi'inCavebyablackcirclelabelled5.(C)Mapshowingtherelief  andprecipitation intheresearchareainmoredetail.Theisohyetsaremarkedbywhitelinesandthecavesbyblackcircleswithnumbersasfollowing:1)Ma'aleâ€“ Dragotcavesystem,2)TzavoaCave,3)SoreqCave,and4)Ma'ale â€“ EfrayimCave.CMRistheCentralMountainRidgerunningfromnorthtosouth, andendingnorthofBe'er â€“ Sheva â€“ Arad(BA)Valley.Citiesareshownbywhitecircles. 385 A.Vaksetal./EarthandPlanetaryScienceLetters249(2006)384 â€“ 399
fromtheNegevDesert,southernIsrael(Fig.1 B,C), where230Th â€“ Uagesoftravertines [17,18] showthat wetterperiodsoccurredat N 500ka, 200ka,140â€“ 120ka, 70â€“ 25kaand 15kaandsimilarly14Cdatesofcarbonate nodulesfromloesssoils,at N 37ka, 28ka,and 13ka  .However,theseproxiesareonlysporadically availableandaresubjecttosecondaryprocessessuchas erosionandweathering. ThelocationofthenorthernboundaryofSaharo â€“ ArabianDesertisoneofthecriticalissuesforpaleoclimate reconstructionoftheEMareaanddeterminingifclimate changesatitsnorthernmarginsweresynchronouswith thoseatthesouthernboundaryandwiththemonsoonal activity.Inthisworkwedeterminethetimingofshiftsin thenorthernboundaryofthedesert,andtheassociated climaticchanges,throughastudyofcavedeposits (speleothems)fromkarsticcavesinthenorthernNegev Desert( Fig.1 C).Today,thisregionencompassesthe transitionbetweentheMediterraneanclimateregioninthe north(withannualrainfallof350 â€“ 1200mm,coolrainy wintersandhotdrysummers)andtheariddesertclimate regiontothesouth,withlessthan200mm.Thisstudy focusesontheregionbetweenpresent-day300mmand 150mmisohyets.Theexistenceofnumerouskarsticcaves richwithspeleothemsmaketheregionidealfor reconstructingthepaleo-positionofthedesertboundary andunderstandingitspaleoclimate.Speleothemsgrowin caveswhenwaterreachestheunsaturatedzoneand vegetationispresentonthesurfacetosupplytheCO2necessaryforlimestonedissolution.Thespecificaimsof thestudyare:1)toreconstructthepaleo-climateofthe northernNegevregionbydatingperiodsofspeleothem growthusing230Th â€“ Udating;2)todeterminetherainfall moisturesourceandthepaleovegetationusing 18Oand 13Cofcalcitespeleothems.Theregionalsignificanceof thenewdatawillbeexploredbyacomparisonwith speleothemrecordsfromtheMediterraneanclimateregion incentralandnorthernIsrael(mainlywithSoreqcave speleothems) [20 â€“ 28] andtheJordanValley rain shadow desertontheeasternflankoftheCentral MountainRidge(CMR)ofIsrael  ( Fig.1 B,C). 2.Climatologicalbackground 2.1.ClimaticzonesofNegevDesert Thepresent-dayNegevDesertcanbedividedinto threemainclimaticzones( Fig.1 C): 1)The northernNegevDesert ,locatedatsouth-eastern corneroftheEMSea,formsa 40kmwidebeltthat includesthecoastalplainnearGaza,thesouthernedge oftheCMR,andfewsmallerridgestotheeastnearthe townofArad.Thenorthernborderoftheareais approximatelydefinedbythe350mmisohyet,which marksthesouthernboundaryofMediterraneanclimate zone.ThesouthernborderisBe'er â€“ Sheva â€“ Arad Valley,dissectingtheareafromwesttoeast( Fig. 1 C).Annualaveragerainfallvariesfrom 300to 350mminthenorthernpartto 150mminthesouth. VegetationchangessouthwardfromC3Mediterranean steppe-foresttoamixofC3andC4semi-desertIrano â€“ Turanianvegetation [30 â€“ 34] . 2)The NegevHighlands ,southofBe'er â€“ Sheva â€“ Arad ValleyconsistofseveralsmallNE â€“ SWtrending ridgeswithelevationsof500 â€“ 1033masl.Therainfall variesfrom150mminthenorthto50mminthesouth. Vegetationchangessouthwardfromsemi-desert Irano â€“ TuranianvegetationtoSaharo â€“ Arabiandesert type,bothcomprisingmixedC3andC4vegetation [30,32 â€“ 34] . 3)The southernNegevDesert ,locatedsouthofthe NegevHighlands,receives 30 â€“ 50mmaverage annualrainfallandischaracterizedbySaharo â€“ Arabiandesertflora.Theregionreceivesitsrainfall mostlyatthebeginning(October â€“ November)and endoftherainyseason(March â€“ May)insporadic shortstorms,usuallyaccompaniedbylocalfloods. Someofthisrainfallisassociatedwithsynoptic systemsthatoriginateinthetropicalAtlanticOcean, passoverAfricaandapproachtheregionfromthe south-southwest.ThesesynopticconditionsinfrequentlyoccurinthenorthernNegev. 2.2.RainfallgradientinthenorthernNegevDesert AtpresentthenorthernNegevDesertreceivesmostof itsrainfallduringthewintermonths(Decemberto February)frommid-latitudecyclones(Cypruscyclones) movingeastwardsabovetheEMSea.Dayan  found thatthetypicalCypruscyclones(themajorcontributorof rainfallinIsrael)correspondtothemajorityoflongfetchof maritimeairmassescrossingtheMediterranean.Summers, betweenMayandSeptember,arehotanddryandtheresult fromthesinkingairofsubtropicalhighs,whichdevelop overtheMediterraneanSeaasstronghigh-pressureridges pusheastwardsfromtheAzoressubtropicalhigh. Rainfallisohyetsrunfromnorthtosouthinnorthern andcentralIsraelparalleltothecoastline,butthenabruptly changetheirorientationandruntothewestparalleltothe no rtherncoastlineoftheS in aiPeninsulaandtheN â€“ S rainfalldecreasebecomesharpesteventhoughthereisn't anysignificanttopographicbarrierinthe50kmbeltinland ofthecoastline( Fig.1 B,C).Thissharpprecipitation386 A.Vaksetal./EarthandPlanetaryScienceLetters249(2006)384 â€“ 399
gradientchangeoccurswheretheCypruscyclonescross theLevantregionfromthewesttotheeast:thenorthern Sinaicoastlineformingthesouthernlimitatwhichrain cloudscanform  .Consequently,therainfallamount northofBe'er â€“ Sheva â€“ AradValleyismuchhigherthanto thesouth( Fig.1 C).SincethepositionofthenorthernSinai coastchangedinthepastduetoQuaternarysealevel fluctuations [37,38] ,itcouldaffecttheamountofrainfall intheNegevDesert. 3.Thestudiedcaves Speleothemsweresampledfromtwocavesystems locatedinlimestoneanddolomitehostrockof Cenomanian â€“ TuronianJudeaGroup(100 â€“ 88Ma): Ma ' ale â€“ DeragotCaves (#1in Fig.1 C)located betweenpresent-day280 â€“ 300mmisohyets,atan elevationof630 â€“ 720masl,10 â€“ 90mbelowthe surface,and65kmfromtheMediterraneanSea.Thisis theclosestNegevcavesystemtotheregionwith Mediterraneanclimate.Untilthecaveswerediscovered duringquarrying,theyhadnonaturalopenings. Tzavoacave (#2in Fig.1 C)locatedwithinpresentday150 â€“ 160mmisohyets,atanelevationof550masl, 20 â€“ 50mbelowthesurface,and80kmfromthe MediterraneanSea.Mostspeleothemsusedinthe presentstudycomefromthiscavebecauseofitscritical locationatthesouthernandmostaridpartofthe researcharea.Thecavehasseveralnaturalopenings. 4.Analyticalmethodsandbackgrounddata Fifteenspeleothems(3stalagmitesand12stalactites) werecollectedfromvariouslocationswithintheTzavoa cave.Twostalactiteswerein-situ,allotherswere broken.Thestalagmitesrangeinsizefrom5to15cm inlengthand 3to15cminwidth.Thestalactitesare upto30cmlongandupto8cmwide.Sixrandomly locatedspeleothemswerecollectedfromtheMa'ale â€“ Deragotcavesystem:twostalagmitesabout0.5mlong and20cmwide,andfourstalactitesvaryinginsizefrom afewcmto 0.5mlongand20cmwide. Thespeleothemsweresectionedusingadiamondsawto exposetheirinternalstructu reandtoeliminatediageneticallyalteredsamples  .Themineralogyandpetrography wasdeterminedusingpetrographicmicroscope,Jeol840 scanningelectronmicroscopeequippedwithOxfordISIS EDSsystem,andPhilipsPW3020X-raydiffractometer. Fordatingpurposes,upto1gmaterialwasdrilled using0.8 â€“ 4mmdiameterdrillbitsalongthegrowth axis(forstalagmites)( Fig.2 A)andacrossthegrowth axis(forstalactites).Sixlaminaeweresampledtwiceby drillingtwodifferentspotsonthesamelamina(these sampleswithduplicatedatingaredesignatedbythe romannumeralsIandIIinSupplementalTable1). Dependingontheuraniumconcentration(Supplemental Table1),80 â€“ 900mgcalcitepowderwasdissolvedin 7NHNO3.Thesamplewasloadedontomini-columns thatcontained2mlBio-RadAG1X8200 â€“ 400mesh resin.Uwaselutedby1NHBrandThwith6NHCl  .AfterwardstheUandThsolutionswere evaporatedtodrynessanddissolvedin2mland5ml of0.1NHNO3respectively.230Th â€“ Udatingwasperformedon125samplesof speleothemsfromTzavoacaveand17samplesfrom Ma'ale â€“ Dragotcaves,usingmulticollectorinductively coupledplasmamassspectrometer(MC-ICP-MS)Nu InstrumentsLtd(UK)equippedwith12Faradaycupsand 3ioncounters.ThesamplewasintroducedtotheMCICP-MSthroughanAridusmicro-concentricdesolvatingnebulisersampleintroducingsystem.Theinstrumentalmassbiaswascorrected(usingexponentialequation) bymeasuringthe235U/238Uratioandcorrectingwiththe natural235U/238Uratio.Thecalibrationofion-counters relativelytoFaradaycupswasperformedusingseveral cyclesofmeasurementwithdifferentcollectorconfigurationsineachparticularanalysis.Theagedetermination waspossibleduetotheaccuratedeterminationof234U and230Thconcentrationsbyisotopedilutionanalysis usingthe236U â€“229Thspike.230Th/Uageswerecorrected fordetrital230Th  ,assuminga232Th/238Uisotope atomicratioof3.8(themeancrustalvalue)inthedetrital components.However,lessthan20%ofthesampleshada230Th/232Thactivityratiolessthan100andneededthis correction(SupplementalTable1).Thereproducibilityof234U/238Uratiowas0.11%(2 ). For 18Oand 13Canalyses,samplesof1 â€“ 2mg materialweredrilledusingan0.8mmdiameterdrill, eitheralongoracrossthegrowthaxis( Fig.2 B).504 measurementsof 18Oand 13Cweremadeinhigh resolutionon3stalagmitesand5stalactitesusingVG SIRA-IIMassSpectrometerwithISOCARBsystemfor carbonateanalysis  .Hendytests  were performedtoensurethatthespeleothemsweredepositedinisotopicequilibrium.Thistestwascriticalfor Tzavoaspeleothemssincethecavehasseveralpresentdaynaturalopenings.Calcite 18Oand 13Cvaluesof calcitearereportedrelativetoPDBstandard. Forlaminaethinnerof b 5mmonlysingledatingwas performed.Thickerlaminaeweredatedattheirtopand base,andinseveralcaseswheretheagedifference betweenthetwowasgreaterthan3kaweincreasedthe datingpoints.Thisenabledtopreciselycalculatethe growthratealongthecrosssection,withtheassumption387 A.Vaksetal./EarthandPlanetaryScienceLetters249(2006)384 â€“ 399
thattheagerepresentsthecenterofthedrilledarea,and thatthereisaconstantgrowthratebetweentwodated points.The 18Oand 13Cprofilesweremeasuredalong thesameline,allowingustodefinepreciselytheageof eachpoint. Inordertoprovideadditionaldataonthepresent-day conditions,threesamplesofdrippingwaterfromTzavoa cavewerecollectedinMarch2002,attheendofthe 2001 â€“ 2rainyseason.Rainsampleswerecollected during2004 â€“ 5winterfromthecityofBe'er â€“ Sheva, 380masland 40kmfromtheEMSeacoast,where eachraineventwassampled,andinthetownofArad, 600masl,and 80kmfromtheEMSeacoast.AtArad therainaccumulatedinaplasticbucketandtoprevent evaporation,a1cmthickoillayerwasadded.Rainwater thataccumulatedbelowtheoilwassampledat40 â€“ 50dayintervalsandtransferredintosealedplastic bottles. 18Oand Dmeasurementsofthecaveand rainwaterwereperformedusingmethodsdescribedin  andarereportedintheSMOWscale. RainfallamountsinBe'er â€“ ShevaandAradduring the2004 â€“ 5winterwere390mmand160mmandtheir 18Oand Dvaluesrangefrom 11.8 to1.3 ,and 68 to6 ,respectively( Fig.3A).AtBe'er â€“ Sheva 18O â€“ Drelationshipsofraineventsabove10mm usuallyfollowtheMediterraneanMeteoricWaterLine (MMWL)  ,consistentwiththeirEMorigin,withthe exceptionoftwoeventson29.10.2004and8â€“ 10.3.2005 thatoriginatedinthetropicalAtlanticOcean(Israel MeteorologicalServicedata).The 18O â€“ Ddatafor theseeventsfittheglobalMeteoricWaterLine(MWL). Raineventsbelow10mmusuallyfollowtheMMWL, withsomefallingonevaporationtrends. 18O â€“ D relationshipsofrainfallinAradfollowtheMMWL. Present-day 18OvaluesofTzavoaCavewaterarefrom 4.8 to 5.6 and Dvaluesarefrom 14 to Fig.2.(A).Typicallaminarcrosssectionalongthegrowthaxisofstalagmite(TZ-15)fromTzavoaCave;individuallaminaearemarkedA â€“ N.Two230Th â€“ UagesareindicatedinlaminaeB1andN.ThreewhitelaminaeseenbetweenthelaminaeB2andC1,C2andD,andDandE,markhiatusesin deposition.Drillinggroovescanbeseenwheresamplesweretakenfor230Th â€“ Udating.(B)Perpendicularcrosssectionoftwojoinedstalactites:TZ22(1)ontheright,andTZ-22(2)ontheleft.Theholesmarkthedrillingpointsfor 18Oand 13Canalyses.(C).Opticalmicroscopeimageofthe laminabetweenCandDinTZ-22(2)incrossedpolarizedlight.Thislaminathatdefinesanunconformityiscomposedofdark,opaque,micrometer size(micritic)calcitecrystalstogetherwithdetritalmaterial(mainlysilica,ironoxideandapatite).Thismicritizationofcalcite,asthevoidsinthe laminaDbelowtheuncomformityarepossiblycausedbycorrosionduringtheceaseofthestalactitegrowthbetween 135kaand 80ka. 388 A.Vaksetal./EarthandPlanetaryScienceLetters249(2006)384 â€“ 399
19.3 ( Fig.3 B),andalsofollowtheMMWL, consistentwithanEMSearainsource  . Nospeleothemdepositionoccurstodaybecausethe amountsofrainfallanddrippingwaterinthecaveare extremelylow,andthereforethepresent-dayrelationshipbetweentheisotopiccompositionsoftherainfall andofthespeleothemscannotbedetermined.Inthe Mediterraneanclimateregion,thespeleothemisotopic composition(oxygenofcalciteandhydrogenoffluid inclusions)reflectstheisotopiccompositionofrainfall [42,24,43] . 5.Results 5.1.Growthperiodsandpetrogaphy StalagmitesandstalactitesfromTzavoacaveareall composedoflowmagnesiumcalcite(upto1%Mg) withlaminatedtexture.The230Th â€“ Uagesshowthat theygrewfrom 200kato13ka(SupplementalTable 1, Fig.4 ).Noneoftheindividualspeleothemsgrew throughouttheentiretimespan.Nospeleothems youngerthan13kawerefound. 5.1.1.Stalagmites Thestalagmitesarecomposedof0.1to4cmthick laminae.Columnarhoneybrowncalcitecrystalsareup to1cminsize,devoidofdetritus.Ineachsample,itis possibletoidentifythinwhitelaminaecuttingthrough thebrownishcrystals( Fig.2 A).Microscopically,these whitelaminaearecomposedofmicrometer( m)size, dark,opaque,slightlycorrodedandporouscalcite crystals,frequentlyrichwithdetritus.Thelaminae usuallymarkgrowthbreaks(hiatuses)( Fig.2 B,C). Thestalagmiteagesarebetween78kaand13ka,(i.e. duringtheglacialintervalsandatthebeginningofthe deglaciation).StalagmiteTZ-15grewbetween78ka and31ka( Figs.2Aand4 )withhiatusbetween46ka and40ka;StalagmiteTZ-6grewbetween46kaand 30ka,withanhiatusbetween43kaand40kaand stalagmiteTZ-14grewfrom 23kato13kawith hiatusesbetween 23kaand 20kaand 16kato 14ka. 5.1.2.Stalactites Thelaminaeofmostofthestalactitesarethinnerthan thoseofthestalagmites( Fig.2 B).Incontrasttothe stalagmites,stalactitesgrewfrom 200kato 13ka, duringbothinterglacialandglacialperiods( Fig.4). Stalactitesthatgrewduringcoolglacialintervalsat MIS-6andduringthelastglacial(MIS-4to2),are characterizedbybrownhoneycoloredlaminaealternatingwiththinwhitelaminae( Fig.2 B). Thelargestalactites:TZ-4,TZ-21andTZ-22(upto 8cmthickandupto30cmlong)arecomposedofhoney colored,columnarcalcitecrystals,upto 1cminsize, preferablyorientatedperpendiculartothegrowthaxis. Theirgrowthperiodsare:TZ-4between196and47ka, TZ-21from189to38kaandTZ-22from200to27ka ( Fig.4 ).Thegrowthduringthelastglacialperiodin Fig.3.(A)The 18Oand DvaluesofrainwaterfromBe'er-ShevaandAradduringthewinter2004â€“ 5.SamplesfromBe'er-Shevawith b 10mm rainfallareshownassmalldots;eventswith N 10mmrainfallareshownasboldblackcircles.Raineventswith N 10mmrainfallbutfromtropical originaremarkedasboldgreycircles.IsotopecompositionsofsamplesfromArad(shownassmallopenrectangles)arefromfivesamplingperiods: 29.10.2004event(thefirstrainoftheseason);1.11to14.12.2004;14.12.2004to2.2.2005;2.2to13.3.2005;and13.3to22.5.2005.TheMWLis markedbythecontinuouslineandMMWLbybrokenline.(B) 18Oand DvaluesofthedrippingwaterinTzavoaCave(crosses)andSoreqCave (opendiamonds)duringMarch2002,relativetoMMWLandMWL. 389 A.Vaksetal./EarthandPlanetaryScienceLetters249(2006)384 â€“ 399
thesestalactiteswaslimitedtotheuppermostlaminae withthicknessoflessthan1cm,withmajorhiatuses between151kaand144ka,141kaand140ka,117ka and 96ka,andfrom 92kato85ka(thesehiatuses arebasedontheagesincludingerrors).Allthese hiatusesaredefinedbysimilarpetrographyasin stalagmites( Fig.2 B,C).Smallerstalactitesgrewonly duringthelastglacialperiod,atsimilartimeintervalsto thestalagmites.Thestalactitesceasedgrowingatabout 13 â€“ 14ka(samplesTZ-1andTZ-9),similartothe cessationfoundforstalagmiteTZ-14( Fig.4 ). Thelargenumberofagedeterminationsfromthe TzavoaCaveenabledthecalculationoftheagerelative frequencydiagram,usingIsoplot3software  .This Fig.5.RunningaverageofspeleothemagerelativefrequencydiagramfortheTzavoaCaveduringthelast200ka(seetextforexplanation).Itshows whichfractionofthesamplesformedincertainagewith95%confidence.Marineisotopicstagesandsubstagesaremarkedatthetopofthegraph. Fig.4.PeriodsofspeleothemdepositioninMa'ale â€“ Dragot(MD)(bottom)andTzavoaCave(TZ)(centerandtop)asindicatedbytheage determinations.Thehorizontalaxismarkstheage(ka).Agedatafordifferentspeleothemsamplesareenclosedinrectangles,togetherwith speleothemsamplenumber.Stalagmitessamplesareindentifiedbythesuffix Stg .Blackdotswitherrorbarsdefineindividualmeasurements. Hiatusesaremarkedbyverticalpatternedrectangles.Marineisotopicstages(1 â€“ 7)aremarkedatthetopofthefigure.Alldataistakenfrom SupplementalTable1. 390 A.Vaksetal./EarthandPlanetaryScienceLetters249(2006)384 â€“ 399
analysis( Fig.5 )showsthatthenumbersofmeasured agesdecreaseastheagesbecomeolder.Thisis consistentwithasamplingbiastowardsyounger speleothems,whicharemoreaccessibleinthecave andprobablycovertheolderspeleothems.Thepeaksof higherfrequenciesalsoindicatemoreintensivespeleothemdepositionandhigherhumidity.Periodsof highestspeleothemagefrequenciesoccurredduringthe lastglacialperiod:at48 â€“ 46ka, 38ka,33 â€“ 31ka,21 â€“ 20ka,and19 â€“ 17ka.Longerintervalswithrelatively highagefrequencyoccurredbetween64and58ka,and between40and26ka.Lesssignificantpeaksofage frequenciesclusterat200 â€“ 196ka,178 â€“ 167ka,162 â€“ 154ka,152â€“ 150ka,137â€“ 132ka,130â€“ 123ka, 118ka,96 â€“ 92ka,85 â€“ 81ka,80 â€“ 78ka,75 â€“ 72ka, 70ka, 54ka,and 14ka. GrowthintervalsofspeleothemsfromMa'aleDragot CavesaresimilartothoseoftheTzavoacave (SupplementalTable1, Fig.4 ).However,speleothems olderthan200kawerealsofound,aswellasonesample withanageof113kathatoccurswithinalonghiatus determinedforTzavoacave,andonelaminaoflate Holoceneage 3.8ka.Thesetwoadditionalagesare consistentwiththelocationofMa'aleDragotcaves northofTzavoa,inmorehumidconditions. 5.2.Oxygenandcarbonisotopiccompositionof speleothems Eightspeleothems,threeconicalstalagmitesandfive conicalstalactitesfromTzavoaCavewereanalysed. Wherepossible,isotopiccompositionsofspeleothems thatgrewinthesametimeintervalweremeasuredto checkiftheyshowsimilarvaluesandtrendsinorderto verifyfurtherthatthespeleothemswereformedin isotopicequilibrium.Similar 18Otrendswereobserved inmostsamples:stalagmiteTZ-6andstalactiteTZ-19 fromthe49 â€“ 43katimeinterval;stalagmiteTZ-15and stalactiteTZ-16between64and58ka,butwithanoffset of 0.5 .Onlyinonecase,partialmatchingisevident Fig.6. 13C(lowerplot)and 18Oprofiles(upperplot)ofthesampledTzavoaCavespeleothemsforthelast200ka.Thespeleothemagesareshown atthetopoftheplotsasblackdiamondswithhorizontal2sigmaerrorbars. 391 A.Vaksetal./EarthandPlanetaryScienceLetters249(2006)384 â€“ 399
betweenstalagmiteTZ-6,stalactiteTZ-19andstalagmiteTZ-15forthetimeintervalbetween39kato30ka. SinceHendy'stestsuggeststhatallsampleswere depositedinisotopicequilibrium,itisunlikelythatthese differencesarearesultofkineticisotopicfractionation. Wethereforesuggestthatcalculationoftheages assumingconstantgrowthrateonbothsidesofthe datingpointisnotalwaysaccurate,causingaslight offsetwhenaverydetailedmatchingisattempted. Theoverallrangeof 18Ovaluesisfrom 11 to 3 ( Fig.6 A).Three 18Ominimanotablyoccurin interglacialandglacialintervalswhensapropelformationoccurredintheEMSea  :200 â€“ 196ka(MIS-7.1) witha 18Ovariationof 6.0 to 10.4 ;177 â€“ 173ka(MIS-6.4)witha 18Ovariationof 7.5 to 11 andbetween131and123ka(MIS-5.5)witha 18Ovariationof 8.5 to 10.7 . DuringthepreviousglacialMIS-6.2andMIS-6.3,at 166 â€“ 150ka,andthetransitionbetweenMIS-6.1and MIS-5.5,at137 â€“ 132ka,the 18Ovaluesrangebetween 7.0 and 4.5 . 18Ovaluesofspeleothems depositedduringthelastglacialperiod(MIS-4 â€“ MIS-2, between 76kaand 13ka)varybetween 6.5 and 3.0 .Theisotopiccompositionsofsmall samplesandsampleswithpoorageresolutionwere notmeasured;thustheisotopicprofilesdonotcoverall periodsofspeleothemdepositioninthecaveandnotall gapsintheisotopicprofilesreflectlackofdeposition ( Figs.4 â€“ 6 ). TheoxygenisotopicprofileperformedforTzavoa CavespeleothemswascomparedwithSoreqCave speleothems  ,showingalinearcorrelationsbetween the 18Oprofiles( R2=0.7to0.8)forthetimeintervals between176kaand173ka,161and150ka,and137 and123ka,butwithroughcorrelationduringthelast glacialperiod.Becauseofthegeneralcorrespondence betweenthetrendsofthetwoprofiles,weconsidered usingwigglematching(onlyifitwaspossiblewithinthe 2 errorbarsoftheages)incaseswhentherecordsdid notexactlymatcheachother. 13Cvaluesareshownin Fig.6 B.Theoverallrange ofvaluesisfrom 0.0 to 9.5 ,averagingat about 6 . 13Cvaluesduringthelastglacialperiod varybetween 9.2 and 1.45 .Asharpdrop Fig.7.Comparisonofthe 18Oand 13CprofilesoftheTzavoacavewiththoseoftheSoreqcave. 18OprofilesofTzavoaspeleothems(upperleft plot)comparedwithSoreqCavespeleothems(lowerleftplot). 13CprofilesofTzavoaspeleothems(upperrightplot)comparedwithSoreqCave speleothems(lowerrightplot).TheagesofTzavoaCavespeleothemsareshownatthetopsoftheupperplotsbyblackdiamondswithhorizontal2 sigmaerrorbars.TheagesofSoreqspeleothemsaremodifiedfromBar-Matthewsetal.  . 392 A.Vaksetal./EarthandPlanetaryScienceLetters249(2006)384 â€“ 399
from0.0 to 6.0 inspeleothemsthatgrewbetween 131and123ka.Anextremechangein 13Cvalues occursbetween200and196kawithadropfrom 1.0 to 9.4 ,whichwasimmediatelyfollowedby ariseto 5.5 . 13Cvaluesalsodroppedto 9.5 duringtheintervalbetween177and173ka. 6.Discussion 6.1.SourcesofrainfallinnorthernNegevduringthe last200ka TzavoaCaveislocatedtodayinadesertenvironment atthesouthernmostlimitofMediterraneancyclone tracks.Onlyasmallfractionoftherainfallisoftropical originfromthesouth [46,47] .Althoughnospeleothem depositionoccurstodayintheTzavoaCave,thick speleothems(withsectionsoftensofcm)were depositedatintervalsduringthelast200ka.Furtherto thesouth,speleothemdepositionduringthelast200ka wasveryminor,andwhenitoccurreditisrepresented byverythinlaminae(1 â€“ 3cmthick).TzavoaCavethus representsthesouthernlimitofthickspeleothem depositioninthepast200ka.Thissuggeststhatmost oftherainfallthatreachedTzavoaCaveinthepast originatedfromthenorth(i.e.fromEMSeamoisture) ( Fig.3 ).Thissuggestionissupportedbythesimilar 18Otrends(butlargeramplitudevariations)ofthe TzavoaspeleothemscomparedwithSoreqCavespeleothems( Fig.7 )formostdepositionalintervals. 6.2.Theminimumamountofannualprecipitation requiredforspeleothemdeposition Depositionofspeleothemsdependsontheavailability ofwaterintheunsaturatedzone.Inaridenvironments, periodsofdepositionareindicativeofwateravailabilityin thiszone [29,48] .Weusetheterm humid(orwet) period fortimeintervalswithahighratioofprecipitation ( P )toevaporation(E ),allowingmorewatertoenterthe unsaturatedzoneforspeleothemdeposition.Correspondingly,theterm arid(ordry)period referstoalow P / E ratio,whichreducestheamountofwaterenteringthe unsaturatedzone,thuscausingthereductionorcessation ofspeleothemgrowth. IncontrasttotheSoreqandJerusalemcavesincentral IsraelandPeqi'inCaveinthenorthernIsraelwhere speleothemdepositionwascontinuousduringthelast 250ka  ,speleothemdepositioninTzavoaCavewas episodicandmainlyoccurredduringtheglacialperiods: MIS-6andMIS-4,3and2,withveryshorthiatuses( Figs. 4and5 ).Duringinterglacials,speleothemdepositiononly occurredforrelativelyshortintervalsat 200 â€“ 190ka (MIS-7.1), 137 â€“ 123kaand 118ka(MIS-6.1 â€“ MIS5.5transitionandMIS-5.5), 96 â€“ 92ka(MIS-5.2),and 84 â€“ 77ka(MIS-5.1). Atpresent,intheMediterraneanclimatezonein northernandcentralIsraelonly 1/3oftheannualrainfall reachestheunsaturatedzone(i.e.,about200 â€“ 300mm)and theremaining2/3islosteitherbyevaporationand/orby runoff [43,49] .Despitethisloss,theamountofwater reachingtheunsaturatedzoneissufficienttoallow speleothemdepositiontodayandthroughouttheHolocene. TheareaaboveTzavoacavereceivesonly150 â€“ 160mm rainfall,mostlyfromthesamec yclones,butnospeleothem depositionoccurstodayoroccurredintheHolocene.We inferthatthislackofspeleot hemdepositioninTzavoaCave wasduetoanegativewaterbudgetintheunsaturatedzone causedbylowerrainfallamounts,evaporationandrunoff. Holocenespeleothemdepositionwasalsonotfound intheMa'ale â€“ EfrayimcavelocatedintheJordanValley rainshadow desert  ,althoughatpresentthis regionreceivesahigheramountofannualrainfall ( 250 â€“ 300mm)thanTzavoaCave( Fig.1 B).However,incavesfromtheMa'aleDragotquarry,located withinslightlymorehumidconditions( 300mmof annual rainfall),thereisminordepositionduringthe Holoceneasevident from athin3.8kaoldouterlayer coveringolderstalactites. Theseconsiderationsimplythatinpresent-dayand Holoceneclimateconditions,speleothemdeposition occursonlywhereannualrainfallexceeds 300 â€“ 350mm.Thus,theepisodicdepositionofspeleothems inthenorthernNegev,comparedtothecontinuous depositionintheMediterraneanclimateregionincentral andnorthernIsrael,suggeststhatsignificantinfluxof waterintotheunsaturatedzoneonlyoccurswhere rainfallexceeds 300mm. Thisdiscussionisstrictlyvalidonlyforclimatic conditionssimilartopresent-dayandHolocene.However,itispossiblethatduringcoolglacialperiodsthe minimumamountofrainfallrequiredforspeleothem depositionwaslowerthan300mm,becausetemperatures andevaporationrateswerelower,andthefrequencyofthe snoweventswashigher.Theminimumamountofrainfall requiredforspeleothemdepositioncouldbealsodifferent than 300mmifthetropicalrainfallarrivedfromthe southduringtheinterglacialepisodes. 6.3.Migrationofthedesertboundaryduringthehumid interglacialintervals SpeleothemdepositionintheTzavoacaveduringthe interglacialintervalsbetween200 â€“ 190ka,137 â€“ 123ka,393 A.Vaksetal./EarthandPlanetaryScienceLetters249(2006)384 â€“ 399
and84 â€“ 77ka,musthavebeenassociatedwithsignificantlymorerainfallthantheHolocene.Assumingthat temperaturesweresimilartotheseofpresent-day,the sourceofprecipitationwastheEMSeaandthata minimumof300 â€“ 350mmofannualrainfallwasrequired forspeleothemdeposition,theannualrainfallduringthese shortperiodswas 100%higherthanatpresent.Higher rainfallisinagreementwiththecontemporaneous presenceofsapropelsintheEMSea [23,24,50] . Supportiveevidencefortheincreaseinrainfall comesfromthe 13Cvalues( Figs.6and7 ).Vegetation typeisoneofthemostimportantfactorsaffectingthe 13Cvaluesofspeleothems [17,21,26,42,41,51,52] . SpeleothemsfromtheMediterraneanclimateregion, northernandcentralIsrael,usuallyhave 13Cvalues between 12 and 8 (average 10.5 ). TheserangesreflectdominantC3-Mediterraneantype vegetation.InthenorthernNegevthevaluesrange between 0 and 9.0 (average 6.0 ),and areconsistentwithC3 â€“ C4mixedvegetationofthe steppeandsemi-desert. 13Cvaluesareevenmore positivefurthertothesouth.Thisincreasein 13C valuesoflargenumbersofspeleothemsfromseveral cavesalongNâ€“ StransectfromMediterraneantoarid climatezonedemonstratesthatvegetationplaysamajor roleinthisregiononthespeleothem 13Cvalues. Atthebeginningofhumidinterglacialintervalbetween 200and196ka(MIS-7.1)thespeleothem 13Cvaluesin TzavoaCavedecreasedfrom 1 to 9.4 .Sucha shiftisindicativeofarapidchangefromdesertvegetation toMediterraneansteppe-forestwithsignificantfractionof C3plantspecies.Duringotherinterglacialhumidintervals (137 â€“ 123ka(MIS-6.1 â€“ MIS-5.5transitionandMIS-5.5), and78 â€“ 76ka(MIS-5.1))the 13Cvalueswere 6 to 7 ,consistentwithmixedC3 â€“ C4vegetationofsteppe proximatetotheborderofMediterraneanclimatezone. Itisconcludedthatthedesertboundarymigratedsouth ofTzavoaCaveduringtheseshortinterglacialtime intervals,( Fig.8 ).Theincreaseinrainfallinthenorthern Negevmostprobablyoccurredduetosouthwardshiftof theMediterraneancyclonetracks.Suchananaloguewe seeinthepresent-dayduringextremelyrainywinters  . Itisalsopossiblethattherewasagreatercontributionof moisturefromtropicalAtlanticandIndianOceans, becausethesimultaneousincreaseinmonsoonactivity hasbeendemonstratedinOmanspeleothems  ,bysharp decreasesinthesalinityoftheRedSea  ,theexistence ofhighlakestandsinsouthernSaharaDesert [6,7,10,12] , andtheformationofironores,travertinesandspeleothemsintheWesterndesert,Egypt  . NorthernNegevhumidinterglacialintervalscoincide withmonsoonmaximaat198ka,127ka,and83ka  . Humidconditionsalsooccurredinsouthernandcentral partsofSaharo â€“ ArabianDesertduringtheseperiods  .However,duringthemonsoonmaximaat105ka andtheHolocene(11ka),humidconditionsprevailedin southernandcentralSaharo â€“ ArabianDesert [8,56] , whereasnorthernNegevwasarid.Thissuggeststhat duringinterglacialsthereisnotalwaysalinkage betweenmonsoonactivityandhumidityinthenorthern boundaryoftheSaharo â€“ ArabianDesert. 6.4.Migrationofthedesertboundaryduringthe glacialperiods MostspeleothemformationintheTzavoaCave occurredduringthetwolastglacialperiods:speleothem depositionwascontinuousbetween190kaand150ka, andbetween76kaand13kaexceptforshorthiatus between25kaand23ka.Asimilarpatternofincreased humidityoccurredduringthetwolastglacialperiodsin Ma'ale â€“ Efrayimcave  .Speleothemdepositionin bothcavesprobablybecamepossiblebecauseof increasedprecipitationandloweringofthetemperatures by5 â€“ 10Cduringglacialperiods [28,45] ,resultingina higher P / E ratio.Thisindicatesthatduringcoolglacial intervalsthedesertboundarymigratedsouthward comparedwithitspresent-daypositionandcaused Mediterraneanclimateconditionstodominateinthe JordanValleyandnorthernNegev( Fig.8 ). Fig.8.Mapshowingthetwopositionsofthesouthernboundaryof speleothemdepositionidentifiedinthisstudy:Solidline:thedry episodesbetween150 â€“ 144ka,141 â€“ 140ka,117 â€“ 96ka,92 â€“ 85ka, 25 â€“ 23ka,and13 â€“ 0ka.Dottedline:humid(wet)periodsat200 â€“ 150ka,137 â€“ 123ka,96 â€“ 92ka,85 â€“ 25ka,23 â€“ 13ka.Othersymbols areasin Fig.1 C. 394 A.Vaksetal./EarthandPlanetaryScienceLetters249(2006)384 â€“ 399
Theaverage 13CvaluesofTzavoaCavespeleothems duringtheseperiodsareof 6 indicativeofmixedC3 â€“ C4(mostlyC4)steppevegetationproximatetotheborder Mediterraneanclimatezone( Fig.7 ).Lowervalues reaching 9.5 at 176ka,and 76kasuggest thatforshortintervalsMediterraneanC3typevegetation invadedfromthenorth.Veryhigh 13Cvaluesreaching 1 to 3 at 58ka,30 â€“ 25ka,and 21kaare reflectingscarcevegetationduetoaridconditions. Periodsofenhancedmonsoonalactivityinthe southernSaharo â€“ ArabianDesert  duringglacialsat 176ka,151ka,61kaand33ka  alsocoincidewith humidperiodsinthenorthernNegev.However,even whenthemonsoonalactivitywasminimal,thenorthern Negevregionwasusuallyhumid.Thus,duringglacials monsoonactivityandhumidityatthenorthernboundary oftheSaharoâ€“ ArabianDesertarenotalwayslinked, consistentwiththeMediterraneanSeabeingthemain humiditysourceinthisregion. 6.5.Comparisonofthespeleothemagefrequencies withthelevelsofthepaleo-lakeLisan LakeLisan(theprecursorofthepresent-dayDead-Sea) existedduringmostofthelastglacialfrom 70kato 15ka,andreceivedmostofitswaterfromnorthern IsraelviatheJordanriver.Duringmostofthisperiodthe lakelevelstandwas 265m,butdroppedto 325 and 320mat 46and40ka,respectivelyandreachedits maximumheightof 160mat26ka.From 25kathe lakelevelgraduallydecreasedto 400mformostofthe Holocene [57,58] .Theperiodofspeleothemdepositionin TzavoaCaveduringthelastglacialperiodcoincideswith theexistenceofLakeLisan.Themoststrikingcorrelation isthesharpdropinthelakelevelafter13ka,coinciding withthecessationofspeleothemdepositioninTzavoa Cave.Althoughthereisnosimplepositiveco-variance betweenthehighestrelativefrequenciesofspeleothem agesandthehighestlakelevels,thesimultaneous speleothemdepositionandoccurrenceofhighlevelsof LakeLisanindicatethattheMediterraneanrainsystems shiftedsouthwardduringthelastglacialperiod.Thisshift broughtaboutincreasedprecipitationassociatedwithcool temperatures,morefrequentsnowfallduringthewinter andlowevaporation,whichincreasedinfiltrationofwater tothecaves,aquifersandlakes. 6.6.Oxygenisotopeevolutionoftherainfallinnorthern Negevduringthelast200ka Comparisonofthe 18OvaluesoftheTzavoaand SoreqCavespeleothems( Fig.7 )revealssimilargeneral trends,butthatTzavoaCavespeleothemsareusually depletedbyapproximately 1 â€“ 2 relativetoSoreqCave samples,withanevenlargerdepletionof 4 between 173and177ka.Assumingthattheoriginofrainfalltoboth desertregionsistheEMSea,thelargedepletionin 18Oof TzavoaCavespeleothemsrequiresexplanation. Twopossibleexplanationsarise.OneisthatTzavoa speleothemsweredepositedathighertemperaturesthanat Soreqcave.Duringthelastglacialperiod,temperaturesin thenorthernNegevcouldhavebeenhigherthanincentral andnorthernIsraelowingtomuchsteepernorth â€“ south temperaturegradients.Thesecouldhavebeentheresultof extremelycoldtemperaturesinEuropeandTurkey,which wereunderPolarhighpressure.Contemporaneoushigh temperaturescouldhavedevelopedinSaharo â€“ Arabian Desertaffectedbysubtropicalhighpressure.Tworeasons partiallyargueagainstthisscenario:a)itdoesnotexplain whythe 18OvaluesofTzavoaCavespeleothemswere depletedbysimilarvaluesduringthehumidinterglacial intervals;b)basedontheoffsetinthe 18Ovalues,the north â€“ southtemperaturegradientwas6 â€“ 8Coverthe 100kmdistancebetweenSoreqandTzavoacaves, whichisveryunlikely. Asecondexplanationisthattherainabovethe TzavoaCavewasmoredepletedinheavyisotopesdue toRayleighfractionationeffects.Dansgaard  showedthatRayleighisotopicfractionationduring rainfallprecipitationresultsinprogressiveloweringof Dand 18Ovaluesofvaporastheresultofincreased rainoutasafunctionofdistancefromthesourceand coolingduetohigherelevation.Therainwater 18O depletionduetoelevationinMediterraneanbasinvaries between 0.1 and 0.6 /100m,whilethemost commonvaluesare 0.2 â€“ 0.25 /100m  . Asimplecalculationassessingtheviabilityofa Rayleighprocessasanexplanationofthe 18Ooffsetis madeusingthepresent-dayrainfallvalues.Present-day 18OofatmosphericvaporabovetheEMduringthe wintervariesfrom 11 to 18.6 ,withanaverage 18O 13.5  .Assumingcondensationtemperaturesof0C,5Cand10C,thecalculatedinitialrainfall 18OvaluesabovetheEMSeaare 1.9 , 2.3 and 2.7 ,respectively.Usingtheseinitialvaluesandthe RayleighfractionationequationsgivenbyCriss  , p.108,wethencalculatethe 18Ovaluesofwaterand vaporasafunctionof F ,thefractionofvaporremainingin theclouds.Theseareshownin Fig.9 .Estimatesofthe fractionofrainfallremovedfromthecloudsasthey advanceinlandareobtainedbyinsertingtherainfall 18O valuesatdifferentsitesinIsrael  .Adjacenttothe coastlineincentralIsrael,therainfall 18Ovaluesrange between 4.0 and 4.5 ,requiring 15%rainout.395 A.Vaksetal./EarthandPlanetaryScienceLetters249(2006)384 â€“ 399
40kmawayfromthecoastline, 400masl(Soreqcave area)theaveragerainfall 18Ois 5.0 to 5.5 , correspondingto20 â€“ 22%rainout,55 â€“ 60kmfromthe coastand700 â€“ 800masl(Jerusalem),therainfall 18Ois 6 to 6.5 giving 25 â€“ 30%rainout,andin Neve-Ativ,northernIsrael( Fig.1 B),45kmfromthecoast butatanelevationof1000masl, 18Ovaluesofrainfall are 7.0 to 7.5 ,requiring 35%rainout.Thus, thevariabilityofpresent-dayrainfall 18Ovaluesfrom varioussitesinIsraelcanbeattributedtomoderate Rayleighfractionation,andmainlytochangesin elevationwithalessmarkeddistanceaffect. Tzavoacaveislocated 80kmfromtheshore-line and550maslandisslightlymoreinlandandhigher thanattheSoreqCavesite(45kmand400m elevation).Thus,fromthepointofviewoftheRayleigh model,aslightdepletionwithrespecttotheSoreqcave istobeanticipated.Itisclear,however,thatthe Rayleigheffectcannotaccountforallthe1 â€“ 2 18O offset.Thus,atpresentitseemsthatacombinationof bothenhancedtemperaturegradientandRayleigh fractionationprovidesthebestexplanation,particularly forglacialintervals,wheretheoffsetismostmarked. Duringthebriefinterglacialintervals,anenhanced componentoflow 18Orainfalloftropicalorigin cannotberuledout. 7.Conclusions ThenorthernNegevislocatedinatransitionzone betweenMediterraneanclimateinthenorthandarid climateinthesouth,withaverageannualprecipitation of350to150mm(i.e.,semi-aridtoarid)withtypical vegetationchangingfromC3MediterraneansteppeforesttoC3+C4Irano â€“ Turaniansemi-desertvegetation.Duringthelast200kathisregionexperienced wetterconditionsbothduringinterglacialsandglacials at200 â€“ 150ka,137 â€“ 123ka,85 â€“ 25ka,and23 â€“ 13ka. Thedominantsourceofprecipitationthroughout thesewetintervalswastheEMSea,withapossible smallcontributionofthesoutherntropicalsources duringtheshorthumidinterglacialintervals.Whenthe rainfallduringtheinterglacialintervalswasofthefirst origin,theminimumannualrainfallmusthavebeen 300 â€“ 350mm,i.e.,aboutatwicethanatpresent. Duringmostofthelastglacialandpartsoftheprevious glacial,theregionwaswetter;however,theminimum amountofprecipitationrequiredforspeleothemdepositioncouldhavebeenlessthan300mmduetolower temperaturesresultinginmorefrequentsnowfalland lowerevaporation.Coincidingwiththeincreasein precipitation,MediterraneanC3typevegetationinvaded southwardforshortperiods,andtheclimateinthe northernNegevbecamemoreMediterranean,butstill remainedmixedC3+C4steppemostofthetime.The climatewasaridinthenorthernNegevduring interglacialandglacialintervalswhennospeleothem depositionoccurred:150 â€“ 144ka,141 â€“ 140ka,117 â€“ 96ka,92 â€“ 85ka,25 â€“ 23ka,and13katopresent-day. Therelativelydepleted 18OvaluesofTzavoaCave speleothemsrelativetoSoreqcavecouldinpartreflect Rayleighdistillationeffects,especiallyiftheywere enhancedbyamoremarkednorth â€“ southtemperature gradientduringglacialperiods.Thereisageneral correlationbetweentheperiodofintensivespeleothem depositionincaveslocatedinthenorthernNegevdesert andinthe rainshadow desert,withtheperiodof existenceofpaleo-lakeLisanduringthelastglacial. Humidperiodsfoundduringinterglacialsinthe northernNegevcoincidewithmonsoonmaximaand withhumidconditionsinsouthernandcentralpartsof Saharo â€“ ArabianDesertat198ka,176ka,151ka,127ka, and83ka,61kaand33ka.However,duringmonsoon maximaat105kaand11ka,humidconditionsoccurredin thesouthernandcentralSaharo â€“ ArabianDesert,butthe northernNegevwasdry.Duringtheglacialmonsoonal minimathenorthernNegevwasusuallyhumid,andthe southernpartsofSaharo â€“ ArabianDesertweredry.This impliesthatthereisnotalwayssynchroneitybetween Fig.9.Possiblechangesin 18O(SMOW) ofmeteoricwaterdueto Rayleighfractionationofrainfall(continuouslines)fromvapor(broken lines).Thechangesintheisotopiccompositionareshownfor condensationtemperaturesof0C,5Cand10C. F isthefractionof thewaterremaininginthecloudsastheyadvancefurtherfromthesource. The 18Odataofthepresent-dayrainfallisshownasfollows:Blackcircle EMSeacoast;opencircle SoreqCavearea;blackrectangle Jerusalem;openrectangle Neve â€“ Ativ(NA). 396 A.Vaksetal./EarthandPlanetaryScienceLetters249(2006)384 â€“ 399
monsoonactivityandhumidityatthenorthernboundary oftheSaharoArabianDesert,whichisconsistentwiththe majorsourceofhumidityintheregionbeingthe MediterraneanSea. Acknowledgements ThisresearchwassupportedbytheIsraelScience Foundation(grantNo.910/05).Wewouldliketothank: N.TepliakovandI.Segalforthehelpandadvicewith chemicalanalysesandwiththeMC-ICP-MS;E.Vaksand E.Reznik-Vaks,Z.Wiener,andtheBlochfamilyforthe helpwithrainwatersampling;A.SandlerfortheXRD analyses;M.DvorchekforguidanceinSEManalyses;S. Ashkenazi,E.Ram,andmembersoftheCaveResearch UnitattheHebrewUniversityforhelpwithfieldwork; theIsraeliNatureandParksAuthorityforpermissionto samplespeleothems.SpecialthankstoY.EnzelandB.Ziv forfruitfuldiscussionsandL.Laor,E.ElianiandU. Simchaifortheirhelpwithlaboratorywork. AppendixA.Supplementarydata Supplementarydataassociatedwiththisarticlecanbe found,intheonlineversion,at 10.1016/j.epsl.2006.07.009. ReferencesN.Roberts,H.E.Wright,Vegetational,lakelevel,andclimatic historyoftheNearEastandsouthwestAsia,GlobalClimate ScenesatLastGlacialMaximum,1993,pp.194 â€“ 220. T.M.L.Wigley,G.Farmer,ClimateofEasternMediterraneanand NearEast,Paleoclimates,PaleoenvironmentsandHuman CommunitiesintheEasternMediterraneanRegioninLater Prehistory,1982,pp.3 â€“ 37. M.Rindsberger,M.Magaritz,I.Carmi,D.Gilad,Therelation betweenairmasstrajectoriesandthewaterisotopecomposition ofrainintheMediterraneanSeaarea,Geophys.Res.Lett.10 (1983)43 â€“ 46. G.Eshel,Mediterraneanclimates,Isr.J.Earth-Sci.51(2002). M.Rossignol-Strick,MediterraneanQuaternarysapropels,an immediateresponseoftheAfricanmonsoontovariationofinsolation, Palaeogeogr.Palaeoclimatol.Palaeoecol.49(1985)237 â€“ 263. G.H.Miller,F.Wendorf,R.Ernst,R.Schild,A.E.Close,I. Friedman,H.P.Schwarcz,Datinglacustrineepisodesinthe easternSaharabytheepimerizationofisoleucineinostrich eggshells,Palaeogeogr.Palaeoclimatol.Palaeoecol.84(1991) 175 â€“ 189. B.J.Szabo,J.C.V.Haynes,T.A.Maxwell,AgesofQuaternary pluvialepisodesdeterminedbyuranium-seriesandradiocarbon datingoflacustrinedepositsofEasternSahara,Palaeogeogr. Palaeoclimatol.Palaeoecol.113(1995)227 â€“ 242. D.Fleitmann,S.J.Burns,U.Neff,A.Mangini,A.Matter, Changingmoisturesourcesoverthelast330,000yearsin NorthernOmanfromfluid-inclusionevidenceinspeleothems, Quat.Res.60(2003)223 â€“ 232. J.K.Osmond,A.A.Dabous,Timingandintensityofgroundwater movementduringEgyptianSaharapluvialperiodsbyU-series analysisofsecondaryUinoresandcarbonates,Quat.Res.61 (2004)85 â€“ 94. N.Petit-Maire,TheSaharaintheHolocene,Map,UNESCOand COMW,IGN(eds.),Aux-an-Provence,France(1993). M.K.Crombie,R.E.Arvidson,N.C.Sturchio,Z.ElAlfy,K.A. Zeid,AgeandisotopicconstraintsonPleistocenepluvial episodesintheWesternDesert,Egypt,Palaeogeogr.Palaeoclimatol.Palaeoecol.130(1997)337 â€“ 355. P.deMenocal,J.Ortiz,T.Guilderson,J.Adkins,M.Sarnthein,L. Baker,M.Yarusinsky,Abruptonsetandterminationofthe Africanhumidperiod:rapidclimateresponsestogradual insolationforcing,Quat.Sci.Rev.19(2000)347 â€“ 361. J.C.Fontes,F.Gasse,OntheagesofhumidHoloceneandlate PleistocenephasesinNorthAfrica;remarkson LateQuaternary climaticreconstructionfortheMaghreb(NorthAfrica) byP. Rognon,Palaeogeogr.Palaeoclimatol.Palaeoecol.70(1989) 393 â€“ 398. H.M.El-Asmar,Quaternaryisotopestratigraphyandpaleoclimateofcoralreefterraces,GulfofAqaba,SouthSinai,Egypt, Quat.Sci.Rev.16(1997)911 â€“ 924. C.Causse,B.Ghaleb,N.Chkir,K.Zouari,H.BenOuezdou,A. Mamou,HumiditychangesinsouthernTunisiaduringtheLate PleistoceneinferredfromU â€“ Thdatingofmolluscshells,Appl. Geochem.18(2003)1691 â€“ 1703. N.Petit-Maire,P.Sanlaville,A.Abed,S.Yasin,R.Bourrouilh,P. Carbonel,M.Fontugne,J.L.Reyss,NewdataforanEemian lacustrinephaseinsouthernJordan,Episodes25(2002) 279 â€“ 280. H.P.Schwarcz,B.Blackwell,P.Goldberg,A.E.Marks,Uranium seriesdatingoftravertinefromarchaeologicalsites,NahalZin, Israel,Nature277(1979)558 â€“ 560. A.Livnat,J.Kronfeld,PaleoclimaticimplicationsofU-series datesforlakesedimentsandtravertinesintheAravaRiftValley, Israel,Quat.Res.24(1985)164 â€“ 172. G.A.Goodfriend,M.Magaritz,PalaeosolsandlatePleistocene rainfallfluctuationsintheNegevDesert,Nature332(1988) 144 â€“ 146. A.Ayalon,M.Bar-Matthews,A.Kaufman,Climaticconditions duringmarineisotopestage6intheeasternMediterraneanregion fromtheisotopiccompositionoftheSoreqCave,Israel,Geology 30(2002)303â€“ 306. M.Bar-Matthews,A.Ayalon,A.Kaufman,LateQuaternary paleoclimateintheEasternMediterraneanregionfromstable isotopeanalysisofspeleothemsatSoreqCave,Israel,Quat.Res. 47(1997)155â€“ 168. M.Bar-Matthews,A.Ayalon,A.Kaufman,Paleoclimate evolutionintheEasternMediterraneanregionduringthelast 58,000 yearsasderivedfromthestableisotopesofspeleothems (SoreqCave,Israel),Isotope T echniquesintheStudyofthe EnvironmentalChange,1998,pp.673 â€“ 682. M.Bar-Matthews,A.Ayalon,A.Kaufman,Timingand hydrologicalconditionsofSapropeleventsintheEastern Mediterranean,asevidentfromspeleothems,Soreqcave,Israel, Chem.Geol.169(2000)145 â€“ 156. M.Bar-Matthews,A.Ayalon,M.Gilmour,A.Matthews,C.J. Hawkesworth,Sea-landoxygenisotopicrelationshipsfrom planktonicforaminiferaandspeleothemsintheEasternMediterraneanregionandtheirimplicationforpaleorainfallduring interglacialintervals,Geochim.Cosmochim.Acta67(2003) 3181 â€“ 3199. 397 A.Vaksetal./EarthandPlanetaryScienceLetters249(2006)384 â€“ 399
A.Frumkin,D.C.Ford,H.P.Schwarcz,Continentaloxygen isotopicrecordofthelast170,000yearsinJerusalem,Quat.Res. 51(1999)317â€“ 327. A.Frumkin,D.Ford,H.Schwarcz,Paleoclimateandvegetation ofthelastglacialcyclesinJerusalemfromaspeleothemrecord, GlobalBiochem.Cycles14(2000)863 â€“ 870. A.Matthews,A.Ayalon,M.Bar-Matthews,D/Hratiosoffluid inclusionsofSoreqcave(Israel)speleothemsasaguidetothe EasternMediterraneanMeteoricLinerelationshipsinthelast 120ky,Chem.Geol.166(2000)183 â€“ 191. S.McGarry,M.Bar-Matthews,A.Matthews,A.Vaks,B. Schilman,A.Ayalon,ConstraintsonhydrologicalandpaleotemperaturevariationsintheEasternMediterraneanregioninthe last140kagivenbythe[delta]Dvaluesofspeleothemfluid inclusions,Quat.Sci.Rev.23(2004)919 â€“ 934. A.Vaks,M.Bar-Matthews,A.Ayalon,B.Schilman,M.Gilmour, C.J.Hawkesworth,A.Frumkin,A.Kaufman,A.Matthews, Paleoclimatereconstructionbasedonthetimingofspeleothem growthandoxygenandcarbonisotopecompositioninacave locatedintherainshadowinIsrael,Quat.Res.59(2003)182 â€“ 193. J.C.Vogel,A.Fuls,A.Danin,Geographicalandenvironmental distributionofC3andC4grassesinSinai,Negev,andJudean deserts,Oecologia70(1986)258 â€“ 265. G.A.Goodfriend,RainfallintheNegevDesertduringthemiddle Holocene,basedon(super13)Coforganicmatterinlandsnail shells,Quat.Res.34(1990)186 â€“ 197. G.A.Goodfriend,TerrestrialstableisotoperecordsofLate QuaternarypaleoclimatesintheeasternMediterraneanregion, Quat.Sci.Rev.18(1999)501 â€“ 513. T.E.Cerling,J.Quade,Stablecarbonandoxygenisotopesinsoil carbonates,climatechangeincontinentalisotopicrecords, Geophys.Monogr.78(1993)217 â€“ 231. N.Feinbrun-Dothan,A.Danin,AnalyticalfloraofEretzâ€“ Israel, CANAPublishingHouse,Jerusalem,Israel,1998,pp.1â€“ 1008. U.Dayan,ClimatologyofbacktrajectoriesfromIsraelbasedon synopticanalysis,J.Clim.Appl.Meteorol.25(1986)591 â€“ 595. A.Zangvil,P.Druian,Upperairtroughaxisorientationandthe spatialdistributionofrainfalloverIsrael,Int.J.Climatol. IJCLEU10(1990)57 â€“ 62. Y.Itzhaki,Pleistoceneshore-linesinthecoastalplaneofIsrael, Geol.Surv.Isr.,Bull.32(1961)1 â€“ 9. A.Avital,GeologicalhistoryofthePlio-Pleistocene â€“ Holocene offshorebasedoncoresoffAshqelon,southernIsrael.,M.Sc. thesis,BenGurionUniversityoftheNegev,2003. F.McDermott,D.P.Mattey,C.Hawkesworth,Centennial-scale Holoceneclimatevariabilityrevealedbyahigh-resolution speleothemdelta(super18)OrecordfromSWIreland,Science 294(2001)1328 â€“ 1331. A.Kaufman,G.J.Wasserburg,D.Porcelli,M.Bar-Matthews,A. Ayalon,L.Halicz,Uâ€“ ThisotopesystematicsfromtheSoreq cave,Israelandclimaticcorrelations,EarthPlanet.Sci.Lett.156 (1998)141 â€“ 155. C.H.Hendy,Theisotopicgeochemistryofspeleothems-I.The calculationoftheeffectsofdifferentmodesofformationonthe isotopiccompositionofspeleothemsandtheirapplicabilityas palaeoclimaticindicators,Geochim.Cosmochim.Acta35(1971) 801 â€“ 824. M.Bar-Matthews,A.Ayalon,A.Matthews,E.Sass,L.Halicz, Carbonandoxygenisotopestudyoftheactivewater-carbonate systeminakarsticMediterraneancave:implicationsforpaleoclimateresearchinsemiaridregions,Geochim.Cosmochim.Acta60 (1996)337 â€“ 347. A.Ayalon,M.Bar-Matthews,E.Sass,Rainfall â€“ recharge relationshipswithinakarsticterrainintheEasternMediterranean semi-aridregion,Israel:[delta]18Oand[delta]Dcharacteristics, J.Hydrol.207(1998)18 â€“ 31. K.R.Ludwig,Isoplot3.00,AGeochronologicalToolkitfor MicrosoftExcel,BerkeleyGeochronologyCenter,Special Publication,vol.4,2003. K.-C.Emeis,T.Sakamoto,R.Wehausen,H.-J.Brumsack,The sapropelrecordoftheeasternMediterraneanSea resultsof OceanDrillingProgramLeg160,Palaeogeogr.Palaeoclimatol. Palaeoecol. 158(2000)371 â€“ 395. U.Dayan,R.Abramski,HeavyrainintheMiddleEast related to unusualjetstreamproperties,Bull.Am.Meteorol.Soc.64(1983) 1138 â€“ 1140. B.Ziv,Asubtropicalrainstormassociatedwithtropicalplume overAfricaandtheMiddle-East,Theor.Appl.Climatol.69 (2001)91 â€“ 102. L.K.Ayliffe,P.C.Marianelli,K.C.Moriarty,R.T.Wells,M.T. McCulloch,G.E.Mortimer,J.C.Hellstrum,500kaprecipitation recordfromsoutheasternAustralia:evidenceforinterglacial relativearidity,Geology26(1998)147â€“ 150. A.Ayalon,M.Bar-Matthwews,B.Schilman,Rainfallisotopic characteristicsatvariouscitesinIsraelandtherelationshipswith unsaturatedzonewater,IsraelGeologicalSurveyReportGSI/16/04, 2004,37pp. N.Kallel,J.C.Duplessy,L.Labeyrie,M.Fontugne,M.Paterne, M.Montacer,Mediterraneanpluvialperiodsandsapropel formationoverthelast200000years,Palaeogeogr.Palaeoclimatol.Palaeoecol.157(2000)45 â€“ 58. D.Genty,M.Massault,Carbontransferdynamicsfrombomb14Cand[delta]13Ctimeseriesofalaminatedstalagmite fromSWFrance modellingandcomparisonwithother stalagmiterecords,Geochim.Cosmochim.Acta63(1999) 1537 â€“ 1548. D.Genty,A.Baker,M.Massault,C.Proctor,M.Gilmour,E. Pons-Branchu,B.Hamelin,D eadcarboninstalagmites: carbonatebedrockpaleodissolutionvs.ageingofsoilorganic matter.Implicationsfor13Cvariationsinspeleothems,Geochim. Cosmochim.Acta65(2001)3443 â€“ 3457. Y.Enzel,R.Bookman,D.Sharon,H.Gvirtzman,U.Dayan,B. Ziv,M.Stein,LateHoloceneclimatesoftheNearEastdeduced fromDeadSealevelvariationsandmodernregionalwinter rainfall,Quat.Res.60(2003)263 â€“ 273. A.Almogi-Labin,C.Hemleben,D.Meischner,Carbonate preservationandclimaticchangesinthecentralRedSeaduring thelast380kyrasrecordedbypteropods,Mar.Micropaleontol. 33(1998)87 â€“ 107. M.Rossignol-Strick,M.Paterne,Asyntheticpollenrecordofthe easternMediterraneansapropelsofthelast1Ma:implicationsfor thetime-scaleandformationofsapropels,Mar.Geol.153(1999) 221 â€“ 237. K.Nicoll,Recentenvironmentalchangeandprehistorichuman activityinEgyptandNorthernSudan,Quat.Sci.Rev.23(2004) 561 â€“ 580. Y.Bartov,M.Stein,Y.Enzel,A.Agnon,Z.Reches,Lakelevels andsequencestratigraphyofLakeLisan,theLatePleistocene precursoroftheDeadSea,Quat.Res.57(2002)9 â€“ 21. Y.Bartov,S.L.Goldstein,M.Stein,Y.Enzel,Catastrophicarid episodesintheEasternMediterraneanlinkedwiththeNorth AtlanticHeinrichevents,Geology31(2003)439 â€“ 442. W.Dansgaard,Stableisotopesinprecipitation,Tellus16(1964) 438 â€“ 468. 398 A.Vaksetal./EarthandPlanetaryScienceLetters249(2006)384 â€“ 399
L.Gourcy,A.Aouad-Rizk,L.Araguas,A.Argiriou,P.Bono,M.F. Diaz-Teijeiro,A.Dirican,A.M.El-Asraq,J.Gat,R.Gonfiantini,N. Horvatincic,B.Ouda,P.Carreira-Paquete,D.Rank,O.Saighi,Y. Travi,P.Vreca,Isotopiccompositionofprecipitationinrelationto aircirculationpatternsintheMediterraneanbasin:preliminary results,InternationalWorkshopontheApplicationofIsotope TechniquesinHydrologicalandEnvironmentalStudies,2004. J.Gat,B.Klein,Y.Kushnir,W.Roether,H.Wernli,R.Yam,A. Shemesh,Isotopecompositionofairmoistureoverthe MediterraneanSea:anindexoftheair â€“ seainteractionpattern, Tellus55B(2003)953 â€“ 965. R.E.Criss,PrinciplesofStableIsotopeDistribution,Oxford UniwersityPress,NY,1999. J.K.Hall,V.A.Krashennikov,F.Hirsh,C.Benjamini,A.Flexer, GeologicalFrameworkoftheLevant,VolumeII:TheLevantine BasinandIsrael,2005,pp.1 â€“ 826. 399 A.Vaksetal./EarthandPlanetaryScienceLetters249(2006)384 â€“ 399
GEOLOGY, September 2007 831 Geology, September 2007; v. 35; no. 9; p. 831; doi: 10.1130/G23794A.1; 2 gures; Data Repository item 2007202. 2007 The Geological Society of America. For permission to copy, contact Copyright Permissions, GSA, or firstname.lastname@example.org.INTRODUCTION The African origin of the early modern humans ca. 200â€“ 150 ka is now well documented (McBrearty and Brooks, 2000; McDougall et al., 2005), but the routes of their expansion out of Africa are still debated (Vermeersch, 2001; Petraglia and Alsharekh, 2003). Fossil remains of early modern humans and sites of African Middle Stone Age industry in the eastern Sahara (McBrearty and Brooks, 2000; Smith et al., 2004a) and Arabian Peninsula (Petraglia and Alsharekh, 2003; Rose, 2004) suggest a migration route from tropical east Africa to the north and northeast. (The African terminology, the Middle Stone Age, is used in this paper to cover the same time period described in Europe as the Middle Paleolithic; Marean and Assefa, 2005.) Archaeological data from northern Israel indicate that one of the major waves of early modern human expansion out of the African continent occurred between 130 and 100 ka (Schwarcz et al., 1988; Valladas et al., 1988; Mercier et al., 1993; Bar-Yosef, 1998; Grun et al., 2005). All possible migration routes leading from Africa to the Levantine early modern human sites (Fig. 1A) converge in the arid to hyperarid Negev, Sinai, and southern Jordan Deserts (Fig. 1B), making it a key region for understanding climatic constraints on early modern human dispersal. Derricourt (2005) argued that the Sinai-Negev Desert route was the major (and possibly the only) way out of Africa, and that the passage through this â€œbottleneckâ€ region was dependent on the development of suitable climate conditions. The southern and central Saharan-Arabian Desert experienced increased monsoon precipitation during this period of early modern human emergence (Szabo et al., 1995; Rohling et al., 2002; Fleitmann et al., 2003; Osmond and Dabous, 2004), but it is not known if an increase in rainfall also occurred in the northern part of the migration corridor. Here, we present evidence for a period of enhanced rainfall activity between 140 and 110 ka in the central and southern Negev Desert, Israel, based on absolute U-Th dating of speleothems. The climate during this period consisted of episodic wet events that enabled the deserts of the northeastern Sahara, Sinai, and Negev to become more suitable for the movement of early modern humans. This period was preceded and followed by essentially unbroken arid conditions, which may indicate that climate change had a major limiting role in the timing of early modern human dispersal out of Africa. GEOLOGICAL CONTEXT The caves of the Negev Desert are dry today and speleothems do not form, but their presence in a number of caves clearly indicates that water reached the unsaturated zone in the past and that surface vegetation was present at times. Today, the mildly arid northern (300 mm/yr) and arid central parts of the Negev Desert (150 mm/yr) mainly receive their rainfall from Mediterranean cyclones (Dayan, 1986), as in the Mediterranean climate region of northern and central Israel. In contrast, the southern Negev is hyperarid (<50 mm/yr) (Amit et al., 2006) (Fig. 1B), and its rainfall mostly consists of rare thunderstorms bringing moisture of tropical origin (Kahana et al., 2002). The transition from the Mediterranean conditions in central Israel to the hyperarid desert in the southern Negev occurs over a distance of less than 150 km. The arid and hyperarid Sinai-Negev Deserts form the land bridge that links Africa to Asia. Speleothem deposition has been continuous during the last 200 k.y. in the Mediterranean climate region of northern and central Israel (Frumkin et al., 1999; Bar-Matthews et al., 2003), whereas in the mildly arid steppe zone in the Jordan Valley and the northern Negev, speleothem deposition has been more sporadic, with hiatuses during dry interglacial episodes (marine oxygen isotope stage [MIS] 5.3.2, and Holocene) and the two last glacial maxima (Vaks et al., 2006). Further south, where present-day precipitation is below 150 mm/yr (central and southern Negev), the cross sections of speleothems younger than 200 ka are reduced to 1 cm, compared with tens of centimeters in the northern Negev (Vaks et al., 2006). METHODS Eleven speleothems (stalactites, stalagmites, and owstone, see GSA Data Repository Table DR11) were collected from ve caves located along a north-south transect of the central and southern Negev Desert, from the present-day 150 mm isohyet to the 30 mm isohyet (Fig. 1B). The caves occur in Middle Cretaceous limestone and dolomite host rock. The speleothems were sectioned, and up to 500 mg of powder was drilled out from different laminae using 0.8.2-mm-diameter drill bits. Stable isotope tests (Hendy, 1971) performed on several of the speleo thems indicate equilibrium deposition. The procedures for accurate U-Th dating using 1GSA Data Repository item 2007202, Table DR1 and Figure DR1, is available online at www. geosociety.org/pubs/ft2007.htm, or on request from email@example.com or Documents Secretary, GSA, P.O. Box 9140, Boulder, CO 80301, USA. Desert speleothems reveal climatic window for African exodus of early modern humansAnton Vaks Institute of Earth Sciences, Hebrew University of Jerusalem, Jerusalem 91904, Israel, and Geological Survey of Israel, 30 Malkhei Israel St., Jerusalem 95501, IsraelMiryam Bar-Matthews Avner Ayalon Geological Survey of Israel, 30 Malkhei Israel St., Jerusalem 95501, IsraelAlan Matthews Institute of Earth Sciences, Hebrew University of Jerusalem, Jerusalem 91904, IsraelLudwik Halicz Geological Survey of Israel, 30 Malkhei Israel St., Jerusalem 95501, IsraelAmos Frumkin Department of Geography, Hebrew University of Jerusalem, Jerusalem 91905, Israel ABSTRACT One of the rst movements of early modern humans out of Africa occurred 130 thousand years ago (ka), when they migrated northward to the Levant region. The climatic conditions that accompanied this migration are still under debate. Using high-precision multicollectorâ€“ inductively coupled plasmaâ€“mass spectrometry (MC-ICP-MS) U-Th methods, we dated carbonate cave deposits (speleothems) from the central and southern Negev Desert of Israel, located at the northeastern margin of the Saharan-Arabian Desert. Speleothems grow only when rainwater enters the unsaturated zone, and this study reveals that a major cluster of wet episodes (the last recorded in the area) occurred between 140 and 110 ka. This episodic wet period coincided with increased monsoonal precipitation in the southern parts of the SaharanArabian Desert. The disappearance at this time of the desert barrier between central Africa and the Levant, and particularly in the Sinai-Negev land bridge between Africa and Asia, would have created a climatic â€œwindowâ€ for early modern human dispersion to the Levant. Keywords: Negev Desert, speleothems, U-Th dating, paleoclimate, out of Africa.
832 GEOLOGY, September 2007multicollectorâ€“inductively coupled plasmaâ€“mass spectrometry (MC-ICP-MS) at the Geological Survey of Israel are described in detail in Vaks et al. (2006). Thirty-three samples were dated (Fig. 2A based on data in Table DR1). Although some speleothems were 20 cm thick, only the outermost laminae of overall thickness of 5â€“ 30 mm were dateable (Fig. DR1, see footnote 1). Some laminae were less than 2 mm thick, which limited the sample size for dating to 10 mg. However, the increase in analytical uncertainty in these samples was minor due to the relatively high U concentrations (see Table DR1). A relative age-frequency diagram was calculated using Isoplot/Ex 3 software (Fig. 2B) (Ludwig, 2003). The higher age-frequency peaks indicate more intensive speleothem deposition. RESULTS AND DISCUSSION The ages of speleothem deposition in the ve caves mostly cluster between ca. 140 ka and 110 ka (interglacial MIS 5.5 and 5.4), with an additional minor deposition interval at ca. 88 ka (beginning of MIS 5.1) de ned by one speleo them 800 800 1000 600 600 400 200 Eastern Mediterranean Sea Syria Lebanon 400 400 600 600 600 50 km JordanIsraelEgypt31N 32N 30N 33N 200400200 Sinai Peninsula 100 100 100 Dead SeaHZ ASH ESID MMR KTO 50 50 50 SK QFGulf of Elat (Aqaba) Jerusalem Jo rdan Valley 1000 1000Negev Desert a b cSahara DesertArabian Desert EMH origin Africa N 35E B A Figure 1. Location maps. A: SaharanArabian Desert (gray shading indicates rainfall below 100 mm) and three possible routes for migration of early modern humans (EMH) toward Levant early modern human sites, passing through desert areas of Sinai, Negev, and southern Jordan: (a) along Nile River, (b) along African Red Sea coast, (c) along Arabian Red Sea coast after crossing Bab-el-Mandeb Straits (after Petraglia and Alsharekh, 2003; Rose, 2004). Rectangle indicates southern Levant area shown in Figure 1B. B: Southern Levant area showing location of ve studied cave sites in Negev Desert. Caves are denoted by black circles and labeled as follows: HZâ€”Hol-Zakh; ASHâ€”Ashalim; ESIDâ€” Even-Sid; MMRâ€”Maâ€™ale-ha-Meyshar; KTOâ€” Ktora Cracks. Open circles indicate caves with early modern human remains in northern Israel: Skhul (SK) and Qafzeh (QF). Lines represent rainfall isohyets (in mm/yr). Present-day Negev Desert can be divided into three subzones: northern mm to 150 mm isohyets; central mm to 50 mm isohyets; southern mm isohyet to Gulf of Elat (Aqaba). 70 160 Speleothem samples from caves of the Negev Desert with their dated laminaeMMR-7 ASH-11 KTO-1(1)ESID-2HZ-1 HZ-3(2) HZ-3(5) ASH-33 ASH-34 MMR-7(2)ESID-7(1) A1 A1 A2 A1.1 A1.2 A3.1 A1.1 A1+A2 A1 A2 B1 B2 A1 A2 B1 B2 B3 B4 A1.2 A1.1 A2.1 A2 A1 B1 B2 B1 A A1-I A1-II A2.1 A2.2 A3 A5 Relative frequency of the ages Ref. 1 Ref. 2 Ref. 3 8090100110120130140150Age (ka)Ref. 4SaharanArabian wet periods B C0.03 0.06 0.00 EMH presence in Levant A Figure 2. Speleothem ages and enhanced rainfall periods in Negev Desert determined in this study compared to data on occupation of caves in northern Israel by early modern humans (EMH) and periods of enhanced monsoonal rainfall in southern Saharan-Arabian Desert. A: U-Th ages (with 95% con dence [2 ] error bars) of speleothem samples of Negev Desert. Speleothem samples are identi ed along vertical axis, and data for each speleothem are separated from one another by thin horizontal lines. Bold horizontal lines separate data for different caves (cave locations appear in Fig. 1B). Stratigraphic order of laminae in each individual speleothem is from bottom to top of row. Dark-gray rectangle shows range of ages of wet period indicated by speleothems. Bold gray bar below upper horizontal axis indicates archaeological period associated with early modern human remains in Skhul and Qafzeh Caves in northern Israel (Schwarcz et al., 1988; Valladas et al., 1988; Mercier et al., 1993; Bar-Yosef, 1998; Grun et al., 2005). B: Relative age-frequency curve calculated from age data, showing fraction of samples that formed in same age interval at 95% con dence level. C: Periods of enhanced monsoonal activity in southern and central parts of Saharan-Arabian Desert indicated by horizontal bars: Ref. 1â€”Osmond and Dabous (2004); Ref. 2â€”Szabo et al. (1995); Ref. 3â€”Fleitmann et al. (2003), and Ref. 4â€”Smith et al. (2004b). Figure shows that early modern human presence in Levant started at peak wet phase of the desert.
GEOLOGY, September 2007 833age (Fig. 2A). The speleothem age-frequency distri butions (Fig. 2B) show that the major depositional period was between 133 and 122 ka. No speleothem deposition was found between 185 and ca. 140 ka (MIS 6) (Fig. DR1, between ca. 110 and ca. 90 ka (MIS 5.4.2), and after ca. 85 ka (i.e., during the most of the interglacial MIS 5.1, the last glacial period, and the Holocene). We therefore conclude that the central and southern Negev Desert was arid to hyperarid during these periods. The lack of deposition during the glacial periods contrasts with the continuous deposition of speleothems in the Mediterranean climate zone (in northern and central Israel; BarMatthews et al., 2003), and with the speleothem deposition during glacial periods in northern Negev (Vaks et al., 2006). Hol-Zakh, the northernmost cave studied in this work (Fig. 1B), and the northern Negev Tzavoa Cave studied by Vaks et al. (2006) are located only 10 km apart, yet the two caves are characterized by a sharp present-day precipitation gradient (150 mm/yr at Tzavoa to 90 mm/yr at Hol-Zakh), which is marked by thinning of the speleothem laminae at Hol-Zakh and a different chronology of speleothem deposition events. Speleothem deposition in an otherwise arid to hyperarid region indicates an increase of rainfall above the caves, leading to recharge of water in the unsaturated zone. The increase in precipitation in the Negev Desert could have occurred because of a southern shift of the Mediterranean cyclones and/or an increase in their intensity. However, the periods of enhanced rainfall most probably would have been short in duration and alternated with long droughts. The data supporting this hypothesis of episodic rainfall are: (1) only thin, 5 mm speleothem laminae were deposited in the caves, and (2) the absence of calcite horizons in middle-late Quaternary gypsic-salic soils of the southern Negev indicates that precipitation levels greater than 80 mm/yr could not have persisted for periods longer than tens to hundreds of years (Amit et al., 2006). Thus, the wet episodes were too short to cause the development of calcite horizons in the soils but long enough to allow the formation of thin speleothem laminae in the caves. Although the wet episodes were short, they would have had a signi cant environmental impact. The caves are located at shallow depths (2 m) below the tops of the hills, and thus their catchments are small compared to those of wadi channels, which would have seen signi cantly greater amounts of water and increased vegetation cover. Pollen records show a signi cant contemporaneous development of vegetation cover over the entire region (WeinsteinEvron, 1987), and increased spring discharge is marked by travertine deposition in the hyperarid southeastern parts of the Negev Desert (Livnat and Kronfeld, 1985). Higher rainfall is also indicated by the Eemian record of perennial lakes in basins occupied today by playas in southern Jordan (Petit-Maire et al., 2002). Figure 2 shows that the cluster of humid episodes in the Negev Desert between 140 and 110 ka was synchronous with increased monsoonal intensity in the southern Arabian Peninsula (Fleitmann et al., 2003) and the Saharan Desert (Szabo et al., 1995; Rohling et al., 2002; Osmond and Dabous, 2004; Smith et al., 2004b). This would have sustained habitable conditions for early modern humans over the entire desert belt, and not just in oases and the Nile Valley (Smith et al., 2004b). The northward movement of early modern humans through the Sinai-Negev land bridge would have been facilitated under these conditions, allowing them to reach the more hospitable climates of Levant ca. 130 ka, as documented by dated early modern human remains from the Skhul and Qafzeh Caves in northern Israel (Figs. 1B and 2) (Schwarcz et al., 1988; Valladas et al., 1988; Mercier et al., 1993; Bar-Yosef, 1998; Grun et al., 2005). Earlier prehistoric sites outside of Africa have not yielded unequivocal early modern human skeletal remains (Stringer, 2002; Barkai et al., 2003; Grun et al., 2005). The faunal assemblages of the archaeological layer in Qafzeh Cave (QF in Fig. 1B) show that Palearctic fauna, which dominated in the caves of southern Levant during MIS 6, disappeared at the onset of interglacial MIS 5, and African and Saharan-Arabian elements became the highest since 780 ka. These latter elements include species of micromammals, hartebeest, equids, and ostrich egg shells, all of which indicate an abundance of savannah species in the area (Rabinovich and Tchernov, 1995; Tchernov, 1992, 1996). The faunal assemblage of Qafzeh Cave thus provides supportive evidence for the northward expansion of African and Saharan-Arabian biotic zones during the period when early modern humans are rst recorded in the area. Middle Stone Age sites are found in the Negev and southern Jordan. Rink et al. (2003) suggested an age range of 90 ka or older for these Middle Stone Age sites. However, the absence of the skeletal remains makes it impossible to determine the hominid type. The Nile River corridor provides a potential path to the Mediterranean Sea through the SaharanArabian Desert (route a in Fig. 1A). Wetter conditions in Sinai and Negev Deserts at this time would have made migration from the Nile Delta to the Levant easier. The increased monsoonal precipitation over the tropics and southern and central Saharan-Arabian Desert at 140 ka also would have enhanced the Nile River ow, made this route more usable, and opened new migration routes across the desert. Stringer (2000) also argued that prehistoric human expansions mainly occurred along the coasts, which provided the rst and fastest migration routes. The possible coastal migration routes from Ethiopia to the Levant follow: (1) the African Red Sea coast (route b in Fig. 1A), or (2) the Arabian Red Sea coast (route c in Fig. 1A), if early modern humans crossed the Bab-elMandeb Straits (Rose, 2004). The Red Sea coastal routes were extremely arid before 140 ka and after 115 ka (Almogi-Labin et al., 2004). Archaeological evidence supports the proposition of Red Sea coastal migration routes. Walter et al. (2000) found that early modern humans occupied the arid Red Sea coast of Eritrea at 125 ka; Van Peer et al. (1996) reported evidence of early modern human occupation of Red Sea coastal mountains of southeastern Egypt during approximately the same period. Petraglia and Alsharekh (2003) found that Middle Stone Age sites of early modern humans (although not radiometrically dated) are concentrated in the western parts of the Arabian Peninsula and some of them are close to the Red Sea coast. Whereas the signi cant increase in wet episodes over northern Africa during the 140â€“ 110 ka period could have removed the climatic barriers to human and animal migration to the north, the aridization of the northern parts of Saharan-Arabian Desert at ca. 110 ka (expressed by the cessation of speleothem deposition in the Negev caves) would have suppressed the return route from the Levant to Africa for at least 20,000 yr. The highest frequency of speleothem ages (indicating highest speleothem deposition rate) occurs in the earlier half of the 140 ka period, followed by a gradual decrease in deposition as conditions moved to aridity (Fig. 2B). Thus, this period of expanding aridity in the Saharan-Arabian Desert may have encouraged early modern humans to move further into the Mediterranean climate zone of the Levant and possibly to the other parts of Eurasia. Wet interglacial events have been suggested to be crucial to the occurrence of Pleistocene migrations â€œout of Africaâ€ (Derricourt, 2005). We therefore suggest that the episodic humid phase between 140 and 110 ka in the northern Saharan-Arabian Desert opened an important interglacial climatic window that allowed early modern humans to disperse out of Africa to other parts of the world.ACKNOWLEDGMENTS This research was supported by the Israel Science Foundation (grant 910/05) and by the Ring Foundation. We would like to thank N. Tepliakov, I. Segal, O. Yoffe, S. Ashkenazi, B. Schilman, and E. Eliani from Geological Survey of Israel, and the team of the Cave Research Unit at Hebrew University, Jerusalem, for their assistance in eld and laboratory work. We would also like to thank L. Grosman, A. Almogi-Labin, R. Amit, Y. Enzel, E. Hovers, N. Goren-Inbar, and R. Rabinovich for advice, and T. Esat, C. Marean, and the two anonymous reviewers for the fruitful comments, which led to considerable improvement of the manuscript.
834 GEOLOGY, September 2007 REFERENCES CITED Almogi-Labin, A., Bar-Matthews, M., and Ayalon, A., 2004, Climate variability in the Levant and northeast Africa during the late Quaternary based on marine and land records, in GorenInbar, N., and Speth, D., eds., Human Paleoecology in the Levantine Corridor: Oxford, UK, Oxbow Books, p. 117. Amit, R., Enzel, Y., and Sharon, D., 2006, Permanent Quaternary hyperaridity in the Negev, Israel, resulting from regional tectonics blocking Mediterranean frontal systems: Geology, v. 34, p. 509, doi: 10.1130/G2254.1. Barkai, R., Gopher, A., Lauritzen, S.E., and Frumkin, A., 2003, Uranium series dating from Qesem Cave, Israel, and the end of the Lower Paleolithic: Nature, v. 423, p. 977, doi: 10.1038/nature01718. Bar-Matthews, M., Ayalon, A., Gilmour, M., Matthews, A., and Hawkesworth, C.J., 2003, Sea-land oxygen isotopic relationships from planktonic foraminifera and speleothems in the Eastern Mediterranean region and their implication for paleorainfall during interglacial intervals: Geochimica et Cosmochimica Acta, v. 67, p. 3181, doi: 10.1016/S0016 (02)01031. Bar-Yosef, O., 1998, On the nature of transitions: The Middle to Upper Palaeolithic and the Neolithic revolution: Cambridge Archaeological Journal, v. 8, p. 141. Dayan, U., 1986, Climatology of back trajectories from Israel based on synoptic analysis: Journal of Climate and Applied Meteorology, v. 25, p. 591, doi: 10.1175/1520(1986) 025<0591:COBTFI>2.0.CO;2. Derricourt, R., 2005, Getting â€œout of Africaâ€: Sea crossings, land crossings: Journal of World Prehistory, v. 19, p. 119, doi: 10.1007/ s10963-z. Fleitmann, D., Burns, S.J., Neff, U., Mangini, A., and Matter, A., 2003, Changing moisture sources over the last 330,000 years in northern Oman from uid-inclusion evidence in speleothems: Quaternary Research, v. 60, p. 223, doi: 10.1016/S0033(03)00086. Frumkin, A., Ford, D.C., and Schwarcz, H.P., 1999, Continental oxygen isotopic record of the last 170,000 years in Jerusalem: Quaternary Research, v. 51, p. 317, doi: 10.1029/1999GB001245. Grun, R., Stringer, C., McDermott, F., Nathan, R., Porat, N., Robertson, S., Taylor, L., Mortimer, G., Eggins, S., and McCulloch, M., 2005, U-series and ESR analyses of bones and teeth relating to the human burials from Skhul: Journal of Human Evolution, v. 49, p. 316, doi: 10.1016/j.jhevol.2005.04.006. Hendy, C.H., 1971, The isotopic geochemistry of speleothems: I. The calculation of the effects of different modes of formation on the isotopic composition of speleothems and their applicability as palaeoclimatic indicators: Geochimica et Cosmochimica Acta, v. 35, p. 801. Kahana, R., Ziv, B., Enzel, Y., and Dayan, U., 2002, Synoptic climatology of major oods in the Negev Desert, Israel: International Journal of Climatology, v. 22, p. 867, doi: 10.1002/ joc.766. Livnat, A., and Kronfeld, J., 1985, Paleoclimatic implications of U-series dates for lake sediments and travertines in the Arava Rift Valley, Israel: Quaternary Research, v. 24, p. 164, doi: 10.1016/0033(85)90003. Ludwig, K.R., 2003, Isoplot 3.00: A Geochronological Toolkit for Microsoft Excel: Berkeley, California, University of California, Special Publication No. 4, 70 p. Marean, C.W., and Assefa, Z., 2005, The Middle and Upper Pleistocene African Record for the Biological and Behavioral Origins of Modern Humans, in Stahl A., ed., African Archaeology: A Critical Introduction; Blackwell Studies in Global Archaeology: Malden, Massachusetts, Blackwell Publishing, p. 93. McBrearty, S., and Brooks, A.S., 2000, The revolution that wasnâ€™t: A new interpretation of the origin of modern human behavior: Journal of Human Evolution, v. 39, p. 453, doi: 10.1006/jhev.2000.0435. McDougall, I., Brown, F.H., and Fleagle, J.G., 2005, Stratigraphic placement and age of modern humans from Kibish, Ethiopia: Nature, v. 433, p. 733, doi: 10.1038/nature03258. Mercier, N., Valladas, H., Bar-Yosef, O., Vandermeersch, B., Stringer, C., and Joron, J.L., 1993, Thermoluminescence date for the Mousterian burial site of Es-Skhul, Mt. Carmel: Journal of Archaeological Science, v. 20, p. 169, doi: 10.1006/jasc.1993.1012. Osmond, J.K., and Dabous, A.A., 2004, Timing and intensity of groundwater movement during Egyptian Sahara pluvial periods by U-series analysis of secondary U in ores and carbonates: Quaternary Research, v. 61, p. 85â€“ 94, doi: 10.1016/j.yqres.2003.09.004. Petit-Maire, N., Sanlaville, P., Abed, A., Yasin, S., Bourrouilh, R., Carbonel, P., Fontugne, M., and Reyss, J.L., 2002, New data for an Eemian lacustrine phase in southern Jordan: Episodes, v. 25, p. 279. Petraglia, M.D., and Alsharekh, A., 2003, The Middle Palaeolithic of Arabia: Implications for modern human origins, behaviour and dispersals: Antiquity, v. 77, p. 671. Rabinovich, R., and Tchernov, E., 1995, Chronological, paleoecological and taphonomical aspects of the Middle Paleolithic site of Qafzeh, Israel, in Buitenhuis, H., and Uerpmann, H.P., eds., Archaeozoology of the Near East: Leiden, Netherlands, Backhuys Publishers, p. 4. Rink, W.J., Richter, D., Schwarcz, H.P., Marks, A.E., Monigal, K., and Kaufman, D., 2003, Age of the Middle Palaeolithic site of Rosh Mor, central Negev, Israel: Implications for the age range of the early Levantine Mousterian of the Levantine corridor: Journal of Archaeological Science, v. 30, p. 195, doi: 10.1006/jasc.2002.0831. Rohling, E.J., Cane, T.R., Cooke, S., Sprovieri, M., Bouloubassi, I., Emeis, K.C., Schiebel, R., Kroon, D., Jorissen, F.J., Lorre, A., and Kemp, A.E.S., 2002, African monsoon variability during the previous interglacial maximum: Earth and Planetary Science Letters, v. 202, p. 61, doi: 10.1016/S0012X(02)00775. Rose, J.I., 2004, The question of Upper Pleistocene connections between east Africa and south Arabia: Current Anthropology, v. 45, p. 551â€“ 555, doi: 10.1086/423500. Schwarcz, H.P., Grun, R.V.B., Bar-Yosef, O., Valladas, H., and Tchernov, E., 1988, ESR dates from the hominid burial site of Kafzeh in Israel: Journal of Human Evolution, v. 17, p. 733, doi: 10.1016/0047(88)90063. Smith, J.R., Giegengack, R., and Schwarcz, H.P., 2004a, Constraints on Pleistocene pluvial climates through stable-isotope analysis of fossilspring tufas and associated gastropods, Kharga Oasis, Egypt: Palaeogeography, Palaeoclimatology, Palaeoecology, v. 206, p. 157, doi: 10.1016/j.palaeo.2004.01.021. Smith, J.R., Giegengack, R., Schwarcz, H.P., McDonald, M.M.A., Kleindienst, M.R., Hawkins, A.L., and Churcher, C.S., 2004b, A reconstruction of Quaternary pluvial environments and human occupations using stratigraphy and geochronology of fossil-spring tufas, Kharga Oasis, Egypt.: Geoarchaeology: International Journal (Toronto, Ontario), v. 19, p. 407, doi: 10.1002/gea.20004. Stringer, C., 2000, Coasting out of Africa: Nature, v. 405, p. 24, doi: 10.1038/35011166. Stringer, C., 2002, Modern human origins: Progress and prospects: Philosophical Transactions of the Royal Society of London, v. 357, p. 569â€“ 579, doi: 10.1098/rstb.2001.1057. Szabo, B.J., Haynes, J.C.V., and Maxwell, T.A., 1995, Ages of Quaternary pluvial episodes determined by uranium-series and radiocarbon dating of lacustrine deposits of eastern Sahara: Palaeogeography, Palaeoclimatology, Palaeoecology, v. 113, p. 227, doi: 10.1016/0031(95)00052-N. Tchernov, E., 1992, The Afro-Arabian component in the Levantine mammalian faunaâ€”A short biogeographical review: Israel Journal of Zoology, v. 38, p. 155. Tchernov, E., 1996, Rodent faunas, chronostratigraphy and paleobiogeography of the southern Levant during the Quaternary: Acta Zoologica Cracoviensia, v. 39, p. 513. Vaks, A., Bar-Matthews, M., Ayalon, A., Matthews, A., Frumkin, A., Dayan, U., Halicz, L., AlmogiLabin, A., and Schilman, B., 2006, Paleoclimate and location of the border between Mediterranean climate region and the Saharo-Arabian Desert as revealed by speleothems from the northern Negev Desert, Israel: Earth and Planetary Science Letters, v. 249, p. 384, doi: 10.1016/j.epsl.2006.07.009. Valladas, H., Reyss, J.L., Joron, J.L., Valladas, G., Bar-Yosef, O., and Vandermeersch, B., 1988, Thermoluminescence dating of Mousterian â€˜proto-Cro-Magnonâ€™ remains from Israel and the origin of the modern man: Nature, v. 331, p. 614, doi: 10.1038/331614a0. Van Peer, P., Vermeersch, P.M., Moeyersons, J., and Van Neer, W., 1996, Paleolithic sequences of Sodmein Cave, Red Sea Mountains, Egypt, in Pviti, G., and Soper, R., eds., Aspects of African Archaeology: Harare, Zimbabwe, University of Zimbabwe, p. 149. Vermeersch, P.M., 2001, â€˜Out of Africaâ€™ from an Egyptian point of view: Quaternary International, v. 75, p. 103, doi: 10.1016/ S1040(00)00082. Walter, C., Buf er, R.T., Bruggemann, J.H., Guillaume, M.M.M., Berhe, S.M., Negassi, B., Libsekal, Y., Cheng, H., Edwards, R.L., von Cosel, R., Neraundeau, D., and Gagnon, M., 2000, Early human occupation of the Red Sea coast of Eritrea during the last interglacial: Nature, v. 405, p. 65, doi: 10.1038/35011048. Weinstein-Evron, M., 1987, Palynology of Pleistocene travertines from the Arava Valley, Israel: Quaternary Research, v. 27, p. 82, doi: 10.1016/0033(87)90051. Manuscript received 12 February 2007 Revised manuscript received 30 April 2007 Manuscript accepted 2 May 2007 Printed in USA
8. Paleoclimate of northern Saharan-Arabian Desert from speleothem record in central and southern Negev and Judea Desert, Israel 8.1 Introduction This chapter reconstructs paleoclimate of the arid and hyper-arid central and southern Negev Desert and "rain-shadow" Judea desert, addressing issues that are not discussed in chapters 1-3. The subjects discussed are: sp eleothems as recorders of humid periods during the Pliocene and Quaternary, the minimum precipitation (rain and snow) amounts necessary for the speleothem deposition during glacial and interglacial periods, sources of rainfall, the vegetation cover, and correlation of the data to global climate changes. This chapter compares the paleoclimate records of the central and southern Negev with those of rain shadow Judea Desert and mildly arid zones of northern Negev and Jordan Valley, and attempts to synthesize these records to provide an overall picture of the paleoclimate evolution. Additional issues discussed are the role of humid periods older than 180 ka in the migration of hominids and animals out of Africa, speleothem deposition as a record of relative erosion rate s and Sr isotopes in desert speleothems as tracers of dust origin and supply. The background and methods for this chapter are describe d in Introduction (1) and Methodology (4) chapters of this thesis. 8.2 Results 8.2.1 Field and petrographic characterization of the speleothems and the cave sediments 184.108.40.206 Vadose speleothems and preliminary U-Pb dating results Most speleothems from the central and southern Negev Caves (#5-#14 in Fig.1C and Appendix 2) are vadose, consisting of stalag mites, stalactites and flowstones, composed
of low Mg calcite whose morphology shows the direction of the gravitational water flow. Their cross sections range from a few millimet ers to ~40 cm thick. The speleothems have a similar stratigraphy in all caves located south of 150 mm presentday isohyet, and can be divided into three main stratigraphic memb ers, based on their field appearance, color, laminae width, and crystal size: the Basal member, the Intermediate member, and the thin Young member (Figs. 3, 4). The massive Basal Member is composed of large (up to few cm in length) columnar low Mg calcite crystals, mostly light transmittin g (Figs. 3(1-2), 4A). The entire thickness of this member is 5-40 cm and it is the thickest of the three stratigraphic members. In Izzim, Ashalim, Mitzpe-Ramon, Wadi Lotz an d Ktora Cracks Caves, the Basal member usually comprises more than 80-95% of th e speleothem volume. The columnar, light transmitting, low-Mg calcite crystals usually show continuous growth across this stratigraphic member (Figs. 3(1), 4A), suggesting that Basal Member was deposited from continuously slowly dripping water. Rarely, main ly in stalactites, the growth of columnar calcite crystals is interrupted by the thin laminae of milky-white or beige, microcrystalline calcite, or by brown or beige lamin ae containing detrital material (Fig. 3(5)). The Basal member terminates at its top by gray, white, beige or red colored lamina, 1-8 mm thick, mainly containing microcrystalline calcite, evaporite minerals and detritus (Figs. 3 (1-4); 4A, B). This possibly suggests a hiatus (growth break) between the Basal member and the Intermediate member. The Basal member is older than ~550 ka , which is the limit of the U-Th dating method. Preliminary U-Pb ages of two laminae, ASH-15-D and ASH-15-E, from the top of the Basal Member in a flowstone from Ashalim Cave (Fig. 3(1)) (performed in University of Melbourne, Australia) are 3.045 0.021 Ma and 3.048 0.014 Ma respectively (Jon Woodhead, personal comm unication, Table 1). Sample ASH-15-D+E was also dated by U-Pb method at the Univers ity of Leeds, UK, giving a preliminary age of 3.015 0.205 Ma (Robert Cliff, personal comm unication). Laminae KTO(1)-1-J and
KTO(1)-1-K from the bottom of the Basal Me mber in a flowstone in Ktora Cracks (Fig. 4A), gave a preliminary U-Pb ages of 2.713 0.060Ma and 3.317 0.063 Ma, respectively (Jon Woodhead, personal commun ication, Table 1). Thus , deposition of the speleothems of the Basal member occurr ed probably 3.3-2.7 Ma ago in the Middle Pliocene. The Intermediate member is 2-15 cm thick, and comprises ~ 4-20% of the speleothem volume in most caves. The member is layere d and composed of a lternating black, brown, beige, gray or colorless laminae of columnar, light transmitting, low-Mg calcite crystals (Fig. 3, 4). The width of each lamina ranges fr om a few mm to 5 cm. Several fine grained milky-white or beige, laminae of micro-crystalline calcite cut through the columnar crystalline structure suggesting hi atuses (Fig. 3(1, 5), 4B). This member is also older than 550 ka but was deposited after the Basal member. The lamina ASH-15-C from the base of the Intermediate member of a flowstone ASH-15 from Ashalim Cave (Fig. 3(1)) gave a preliminary U-Pb age of 1.269 0.013Ma, which is Early Pl eistocene (Jon Woodhead, personal communication, Table 1). The Young member is 0.5-2 cm thick, compos ed of thinly laminated, columnar, lowMg crystalline calcite, and comprises less than 1% of the speleothem volume. At some localities it was possible to accurately date the uppermost 2 to 6 calcite laminae by the UTh method (chapter 8.2.2). The light transmitti ng crystalline calcite laminae are several mm thick and separated from one another by fine (<1 mm), opaque, milk-white, red, brown grey and beige fine laminae (Fig. 3(6), 4B). The latter are composed of microcrystalline calcite, gypsum, halite, clay minerals, quartz, and minor amounts of celestine, barite, apatite, and minerals of Fe, Mn, Ti, and Zn (Fig. 5; 6; 7). This composition is similar to present day â€œpa tinaâ€ accumulating on speleothem surfaces in arid and hyper-arid environments, suggest ing that these laminae formed due to condensation corrosion and/or accumulation of evaporite minerals or detrital material where conditions were dry. Condensation corrosion occurs when tiny water droplets
undersaturated with respect to calcium carbonate condensate on dry speleothem surface, and corrode the crystalline calcite forming fine, milky-white, opaque, and sometimes porous lamina of micro-crystalline calcite on the speleothem surface (Auler and Smart, (2004); Dreybrodt et al., (2005) and references within). The detrital material accumulated after it was washed into the cave by rare rain events; drying out left a residue of clays and evaporite minerals on the speleothem surface. In Kanaim Cave, Judea Desert, speleothem s show laminated structure similar to speleothems in mildly-arid zones of Jord an Valley (Vaks et al., 2003) and northern Negev (Vaks et al., 2006), with the difference th at they are composed of pure calcite with very small amounts of detrital material. Stalact ite sections dateable by the U-Th method in this cave reach more than 10 cm in lengt h. These stalactites are composed of beige or gray, 2-15 mm thick laminae of columnar, li ght transmitting, low-Mg calcite crystals, alternating with fine (<1 mm), opaque, milk-w hite laminae of micro-crystalline calcite with some halite and gypsum (Fig. 8). The la tter represent hiatuses in calcite growth, during which condensation corrosion and depos ition of evaporite minerals took place. The laminae older than the dating limit show ed no change in petrography relative to younger ones. The three-member speleothem stratigraphy of the central and southern Negev was not observed in Kanaim Cave. 220.127.116.11 Phreatic speleothems Phreatic speleothems are found in Ma'ale Dragot Cave in northern Negev, Makhteshha-Qatan Cave, the lowermost passages of the Ashalim Cave, Ma'ale-ha-Meyshar Cave and some small caves nearby. These speleo thems were deposited below the groundwater table, when the caves were submerged. In all caves, with the exception of Makhtesh-haQatan, the phreatic speleothems are older th an the 550 ka. Phreatic speleothems, as evident of any apparent gravitational control on their morphology, form 1-20 cm calcite overgrowth layers on the caves floor, walls and ceilings (Figs. 9A, B). In Ma'ale-ha-
Meyshar Cave and the nearby small caves, calc ite "rafts" formed on ancient pool water surface and some of them sank later to the floor, forming a cover of up to 1 m thick on the caves floor (Fig. 9C). The formation of the cave â€œraftsâ€ probably postdates the calcite overgrowth layer on the walls and ceiling of the caves. Cave "rafts" are the precise marker for the location of the ancient groundw ater level (Polyak a nd Asmerom, 2005). 18.104.22.168 Cave sediments and paleosols washed into the caves Three major types of soil are found above the caves: 1) In areas of the northern Negev and Jordan Valley where annual precipi tation exceeds 250-300 mm, brown Terra-Rosa soil is found; this type of soil becomes more reddish when climate becomes more humid Mediterranean; 2) beige Loess soils are found in northern and cen tral Negev and the Judea Desert, where present-day precipitati on is between 250 and 80 mm; 3) beige and gray Hamada soils, typical of hyper-arid desert where present-day annual precipitation is 80 mm or lower, are found in the central and southern Negev, in Dead-Sea and Arava valleys (Dan et al., 1975; Rabikovitz, 1992). The cave sediments found in most caves resemble the Loess or Hamada soils above the caves. In caves with the natural openings, some animal related organic matter and bat guano were found. In Even-Sid-Ramon and ad jacent caves, and the Maâ€™ale-ha-Meyshar Caves, cave sediments differ from present-da y soils due to their reddish color and high contents of iron oxide. They are possibly pale osols that were washed to the caves. In Maâ€™ale-ha-Meyshar Caves, these sediment s clearly postdate the phreatic calcite overgrowths (Fig. 9A), and were probably washed into the caves as the same time as the formation of the cave â€œraftsâ€ and vadose speleothems older than 550 ka. In the vadose speleothems the laminae of the columnar cal cite are separated by fine detrital laminae that have a reddish color similar to the paleosol color (Fig. 6A).
8.2.2 Uranium concentrations of centra l and southern Negev speleothems Uranium concentrations in central and southern Negev speleothems (Table 1) range between 0.04 ppm and ~20 ppm. In speleothem s located within Turonian limestone host rock in the southern Negev (S hizafon mini-caves and Ktora Cracks), the range is 0.18 1 ppm, whereas in speleothems of the central Negev in Turonian host rock the range is 0.7 20 ppm. In speleothems located in Cenom anian dolomite and limestone host rocks uranium concentrations are slightly lower, 0.04 2 ppm. In a speleothem from Wadi-Lotz Cave located in Eocene limestone, uranium concentration is ~0.4 ppm. Corresponding uranium concentrations in the host rocks are as follows: 1 3.5 ppm in Turonian limestone from the southern Negev, 0.1 0.6 ppm in Turonian limestones from the central Negev, 0.35 ppm in Cenomanian limestones, 2 3.2 ppm in Cenomanian dolomites, and ~0.8 ppm in Eocene limestone. Ur anium concentrations in the soils above the caves vary from 1.7 to 3.8 ppm. No direct link was found between the uranium concentrations in speleothems and the host ro ck and/or the soil, s uggesting that uranium concentration in speleothems depends largely on migration routes of the water reaching a cave. Some speleothems of from three caves in Turonian host rock, Kanaim, Hol-Zakh and Ashalim, contain uranium concentrations ranging from ~4 ppm to ~20 ppm, which are much higher than the soil and the host roc k. It is suggested that the source water migrated through Senonian phosphorites that co ntain up to 120 ppm uranium (Soudry et al., 2002), and are located in vicinity of Kanaim and Hol-Zakh Caves. Senonian phosphorites were not found above Ashalim Cave, but they are located above the Turonian host rocks 2.5 km to the north. It is thus possible that phosphorites were located in vicinity to Ashalim Cave in the past, but have been eroded.
8.2.3 U-Th ages and depositional periods 22.214.171.124 Kanaim Cave, Judea Desert and the modifi ed record of Ma'a le-Efrayim Cave, Jordan Valley U-Th ages were determined on 14 speleothem samples from Kanaim Cave (Table 1). The relative age frequency curve, calculated using Isoplot/Ex3 so ftware (Ludwig, 2003) (Fig. 10), shows that speleothem deposition clus ters at the following time intervals: 370340 ka, 335-315 ka, 270-240 ka, 211-204 ka, 157-154 ka, 149-144 ka, 143-140 ka, 133130 ka, ~38 ka, and 32-31 ka. These results indicate that speleothem deposition occurred only in short episodes during the 370 ka time span. U-Th ages were determined on 15 speleo them samples from Ma'ale-Efrayim Cave using MC-ICP-MS (Table 1) in order to add more ages to the existing record from TIMS dating (Vaks et al., 2003). The relative age frequency curve, which includes both new and previous age results, calculated using Isoplot/Ex3 software (Fig. 10), shows that speleothem deposition in Jordan Valley clusters between 32 0-255 ka, 230-205 ka, 185170 ka, 162-145 ka, ~135 ka, 75-24 ka, 17-16 ka , and 13.5-11 ka. This record expands the earlier work of Vaks et al . (2003) to 320 ka, eliminates an erroneous 78 ka deposition episode, but includes a short depositional episode at 12.1 1.3 ka (sample ME-12-out at the top of the stalagmite ME-12). 126.96.36.199 Caves from the central and southern Negev Desert The dating study was performed on the outer most laminae of the Young member, 0.22 cm thick, which were found to be dateab le by the U-Th method. One hundred twenty three U-Th age determinations using MC-ICPMS were carried out (Table 1, Fig. 11A), and the relative age frequency curve was calc ulated using Isoplot/E x3 software (Figs. 11B, 12A, B, C). The results show that sp eleothem deposition was highly discontinuous, with clusters at the following time inte rvals: 520-440 ka, 3 80-285 ka, 255-240 ka, 225185 ka, 160-155 ka, 142-109 ka, and 89-86 ka.
The age uncertainties for the time interv al between 550 ka and 350 ka are large and overlap one another. Thus, it is difficult to precisely de termine the periods of speleothem deposition. The relative age frequencies were higher at 520-440 ka and 380-350 ka, and very low between 440 ka and 380 ka (Fig. 12C). Speleothem deposition during the last 350 ka is classified here into two groups: 1) periods with relatively intensive deposition, defined by at least two ages from tw o different caves; 2) episodes with limited deposition defined by single ages from single caves. Periods of intensive speleothem deposition occurred during the three major time intervals: 350-290 ka, 220-190 ka, and 142-109 ka. Limited speleothem deposition occurre d at 290-285 ka, 255-240 ka, 225-220 ka, 190-185 ka, 160-155 ka and 89-86 ka. No speleothem deposition is obser ved at 285-255 ka, 240225 ka, 185-160 ka, 155-142 ka, 109-89 ka, 86-0 ka. There are more speleothem deposition periods in the central Negev compared to the southern Negev (to the south of 50 mm is ohyet). Intensive speleothem deposition in southern Negev only occurred during the pe riod between 350 ka and 310 ka. Limited speleothem deposition occurred at ~137 ka, whereas other periods of speleothem deposition are missing, including the periods at 310-290 ka and 220-190 ka during which many speleothems formed in central Negev (Fig. 11). Even-Sid-Ramon mini-caves are the southern most caves were speleoth em deposition aged 310-290 ka period is found. Ma'ale-ha-Meyshar Cave is the southern mo st cave in which speleothems aged 220-190 ka are found. During the three periods of intensive spel eothem formation, peaks occur at 343 ka, 320 ka, 295 (Fig. 12C); 218 ka, 212 ka, 205 ka , 200 ka, 193 ka (Fig. 12B) and 138 ka, 133-131 ka, 126 ka, 117 ka, and 112-111 ka (Fig. 12A). During the periods between 220190 ka and 142-109 ka, the average time sp an between adjacent peaks is 6.4 1.2 ka. Most ages follow the stratigraphic order with the outermost laminae being the youngest. However, in some cases laminae adjacent to hiatuses or next to the present day
surface show an age reversals, e.g. stalact ite KN-8 from Kanaim Cave and flowstone ASH-33 from the Ashalim Cave (Table 1; Fig. 8, Fig. 13). In KN-8, the measured age of out er lamina A is older (182.6 2.3 ka) than that of lamina B (146.5 2.8 ka). This may be due to condensation corrosion processes that affected lamina A and resulted in radionuclide remobilization (Auler and Smart, 2004). Lamina B most probably represents the true age because of its clear columnar calcite crystals which are typical for a closed system (Fig. 8A, B). In ASH-33, the ages of upper 13 sub-lamin ae (A1 to E2.1) follo w the stratigraphic order from 218.0.0 ka to 116.7.1 ka, but the 14th sub-lamina E2.2 above the hiatus (between laminae E and F (>500 ka)), has an age of 199.5.5 (Table 1; Fig. 11A). The hiatus between laminae E and F represents a >300 ka interval with no speleothem deposition, during which detrital material (clay minerals, and host rock fragments) accumulated on speleothem surface (Fig. 13 (1), (2 ), (3), (4)). On re newal of speleothem deposition at about 220 ka, calcite laminae E 2.1-D1 overgrew the detrital fragments, but porosity remained in between the new lami nae and the old surface, causing open system conditions, that lead to deposition of the younge r secondary calcite, and the age reversal in lamina E2.2 (Figs. 13(4), (5)). The fabric of lamina E2.2 differs from the columnar crystal fabric of laminae A1 to E2.1, consis tent with this hypoth esis (Fig. 13(3)). 8.2.4 Oxygen and carbon isotopic compositio ns of the central and southern Negev speleothems 18O and 13C ( PDB) profiles were measured on 12 speleothems from six caves. Eleven speleothems have sections of 350 ka and younger, and four speleothems include Basal and Intermediate members, several la minae of the Young member older than 550 ka and one lamina dated by U-Th to 516 51 ka.
188.8.131.52 Oxygen isotopic compositions 18O values of speleothems older than 550 ka were measured in Ashalim Cave across the entire cross section of flowstone ASH-15, which comprises the Basal and Intermediate members, the bottom laminae of the Young member from stalagmite ASH34 (Ashalim Cave), stalactite ESID-2 (Even-Sid-Ramon mini-caves) and flowstone MMR-7(2) (Maâ€™ale-ha-Meyshar Cave). 18O values of the Basal member vary slightly from -9.4 and -9.7 on the bottom ~ 3 cm, than they then decrease to -10.1 -11.8 across the remaining major part of the Basal member. 18O values of the Intermediate member range between 9.3 to -10. At the bottom of the Young member, the 18O values vary between -7.1 to -10.4 . In Even-Sid-Ramon minicaves (ESID-2, lamina B) an episode of sp eleothem deposition was dated to 516 51 ka gave 18O values decreasing from -0.3 to -3.5 (Fig. 14) 18O profiles for the period from 350 ka to 315 ka are repres ented by flowstone SHACH-1 from the Shizafon mini-caves; a nd for the 305-285 ka period â€“ by phreatic speleothem MKTC-5 from Makhtesh-ha-Qatan Cave and by stalac tite ESID-2 from Even-Sid-Ramon mini-caves. During the former interval, the 18O values range between -11.3 and -10.2 and are the lowest values found in the last 350 ka. During the latter period 18O values decrease from -1.6 to -3.7 in the Makhtesh-ha-Qatan Cave and then vary between -4 and -2.8 in Even-S id-Ramon mini-caves (Fig. 15A, D, E). One of the most important periods of sp eleothem deposition in the central Negev Desert occurred between ~220 ka and 190 ka. 18O profiles were measured on five speleothems from four caves: stalactite HZ-3(2) from Ho l-Zakh Cave; flowstone ASH-33 and stalagmite ASH-34 from Ashalim Cave ; stalactite ESID-2 from Even-Sid-Ramon mini-caves; and flowstone MMR-8 from Ma'a le-ha-Meyshar Cave, covering the periods between 221 ka and 219 ka, and between 213 ka and 197 ka. 18O values range between -6.9 and -10.5. During the 221-219 ka period 18O values measured on one flowstone from the Ashalim Cave show an increase from -7.7 to -7.2 and a decrease
from -6.9 to -7.7 between 213 ka and 208 ka . In contrast, five speleothems formed between 208 ka and 197 ka give 18O values between -7.2 and -10.5. Although these differences in 18O values exist between the caves, the overall 18O trends are similar and the calculated running average curves indicate three peaks of nega tive values -9.6, -9.65 and -10.1 at 206.5 ka, 204 ka and 199 ka, respectively. Two peaks of high values of -7.0 and -7.2 o ccur at 214 ka and 203 ka respec tively (Figs. 15A, 16 C, D). 18O profile of speleothem deposited during the short episode at ~157 ka is represented by lamina C of a phreatic speleothem MKTC -5 from Makhtesh-ha-Qatan Cave, having 18O values varying between -3 and -4.7 (Fig. 15C). The major speleothem deposit ion period in the central Negev Desert during the last 550 ka occurred between 142 ka and 109 ka, and is represented by fourteen speleothems from six caves. 18O profiles were measured on seve n speleothems from four caves covering time span from 133 ka to 115 ka. Th e studied speleothems are flowstone HZ-1 and stalactite HZ-3(2) from Hol-Zakh Cave ; stalactite ASH-11, flowstone ASH-33 and stalagmite ASH-34 from Ashalim Cave; stal actite ESID-7 from Even-Sid-Ramon minicaves and flowstone MMR-7(2) fr om Ma'ale-ha-Meyshar Cave. Based on 18O values, it is possible to divide th e 133-115 ka period into two parts. The first sub-period occurred between 133 ka and 127 ka, represented by seven speleothems from the four caves. During this period the 18O values range between -7.9 and -10.2. Although the 18O values differ between the caves, the trends are similar. The running average of 18O shows four peaks of negativ e values of -9.5, -9.4, -9.8 and -9.6 at 131.5 ka, 128.5 ka, 127 ka, and 126.5 ka respectively. Three peaks of positive values of -8.6, -8.2 and -8.3 occur at 132 ka, 130 ka and 127.5 ka respectively. The second sub-period between 127 ka and 117 ka is represented by four speleothems from two caves. Three spel eothems from the Ashalim Cave have 18O values between -7.4 to -8.5 between 127 ka and 118 ka with a de crease to -9.8 at 117 ka. The speleothem from Ma'ale-ha-Meyshar Cave which grew between 127 ka and
119 ka, have 18O values ranging from -10.4 to -10.7. Due to the large difference in 18O values and trends a running average was not calculated for this period (Fig. 15 A, 16A, B). The youngest period of speleothem deposition at 88-87 ka is characterized by 18O values vary between -7.7 and -8.1 (Fig. 15A, B). 184.108.40.206 Carbon isotopic compositions Speleothem 13C values vary in the Basal member vary between -2 and -3.5; in the Intermediate member the values decrease continuously from +2.1 at the base to -6.1 at the top. In the bottom laminae of the Young member 13C values largely vary from +1.35 to -8.5. 13C values decrease from -2.0 to -3.1 during speleothem deposition episode at 516 51 ka (Fig. 17). Oscillations between high and more negative 13C values are characteristic of all the ensuing periods. Between 350 ka and 315 ka, 13C values range between +1 and -1.8, compared to the 305-285 ka period where they decrease from -3.2 to -7.5 (Fig. 18A, D, E). Between 221 ka and 219 ka, 13C values range between +1.5 and +0.7, which are the highest values during the last 350 ka. Between 213 ka and 208 ka, 13C values vary from -0.2 to -0.6, a nd from -0.5 to -6.8 between 208 ka and 197 ka, with the highest values being found for the Ashalim Cave and the lowest for Even-Sid-Ramon mini-caves. The running average 13C values emphasize two major peaks of negative values of -5.0 and -4.3, at 205 ka 198 ka, respectively. Three major peaks of positive values of +1, 0.3, and -0.9 are identified at 220 ka, 208 ka and 202 ka respectively (Fig. 18A, 19C, D). 13C values typical to the ~157 ka deposition episode in Makhtesh-ha-Qatan Cave range from -7 to -8.5, increasing to -6 in the end (Fig. 18A, C). During the 133 ka 115 ka period the 13C values range between -0.5 and -7.7 (Fig. 18A; 19A, B). Between 133 ka and 127 ka, the 13C values are between -0.5 and -7.6. Although the 13C values differ between the caves , generally decr easing trends
mark this period with the most significan t decreases being found in the northernmost caves. The running average shows four negative peaks of -3.5, -5.0, -5.6 and -5.8, at 131 ka, 130 ka, 128.5 ka and 126.5 ka, respectively, and three positive peaks of -1.6, -2.9, and -3.2 occur at 132 ka, 130.5 ka, 128 ka respectively (Fig. 19A, B). Following this, 13C values sharply increase and during the 126-118 ka period the 13C values range between -1.4 and -2.8 (running average ranges between -1.8 and -2.3). After 118 ka the 13C values decrease sharply reaching -7.7 at 116 ka. Speleothems of Ashalim and Ma'ale-ha-Meysha r caves have similar trends during this period (Fig. 19A, B). During the youngest sp eleothem deposition event at 88-87 ka (Even-Sid-Ramon Cave), the 13C values vary from -5.6 to -4.1 (Fig. 18A, B). 8.2.5 Present-day rainfall in the Negev Desert: amounts and hydrogen and oxygen isotopic compositions Rainfall data: amounts, D and 18O values for the northern, central and southern Negev Desert measured durin g the winters of 2004-2005 (incl uding data of Vaks et al., (2006)) and 2005-2006 are given in Table 2. Rain fall amounts vary as follows: in Be'erSheva from 207 mm to 389 mm; in Arad from 53 mm to 161 mm; in Makhtesh-ha-Qatan from 25 mm to 122 mm; in Mitzpe-Ramon from 48 mm to 50 mm and in Neot-Smadar from 9 mm to 19 mm. D and 18O values of rainfall in the northern Negev (Be'er-Sheva, Arad) are between +37.4 and -68.3 and between +6.4 and -11 .8, respectively with d-excess values between +35 and -13 (Table 2, Fig. 20). Average 18O and D compositions during the years 2004-2006 are respectively -4.7 and -17.6 (d-excess = +20.2) in Be'erSheva, and -4.1 and -13.8 (d-excess = +19.1) in Arad (Table 2; Fig. 20A, B; Fig. 21C). D and 18O values of rainfall in central and southern Ne gev are between +39.0 and -44.0 and between +5.1 and -8.0, respectiv ely with d-excess values ranging
between +41 and -12 (Table 2; Fig. 21). All events with d-excess values below +1 are smaller than 2 mm. Average rainfall 18O and D values in Makhtesh-ha-Qatan for 2002-2006 are -2.4 and -6.1 respectively (d-e xcess = +13.4) (including data of Kuperman (2005) for 2002-2004); -3.3 and 11.1 (d-excess = +15.1) in MitzpeRamon, and -2.0 and -7.5 (d-excess = +8.8) in Neot-Smadar (Table 2; Fig. 21A, B, C). The average 18O and D values of the Ein-Netafim spring water are -3.9 and -18.4 respectively, with d-excess of +13.2 (Table 2). 8.2.6 D values of speleothem fluid incl usions from northern, central and southern Negev Corrected D values of speleothem fluid incl usions (FI) from Tzavoa, Hol-Zakh, Ashalim and Ma'ale-ha-Meyshar Caves, range between -23 and -58 (Table 3, and Fig. 22A, B). These values are in the range of the present day rainfall, but significantly lower than the average (Table 2, Fig. 20, 21). D values of the FI water of Ashalim Cave speleothems older than the 550 ka, are between -26 and -53. The speleothems of the Basal member in Ashalim Cave have very low FI water content (<0.1 l for 200 mg calcite) and only one sample was successfully analyzed, giving a D of -47.3. Speleothems of the Intermediate member have higher water contents and 7 samples from this member gave D values between -26 and -53 (Fig. 22B). D values of speleothem FI deposited dur ing the last 220 ka range between -23.0 and -58. Between 220 ka and 190 ka, they vary from -43.8 to -56.9, with no significant differences between the caves. Between 177 ka and 172 ka, D values range between -43.6 and -48.7, and between 161 ka and 157 ka they in crease to -34.0. Values for the 132-117 ka period vary from -55.9 to -31.7, with the lowest values in Ma'ale-ha-Meyshar Cave and the highest valu es in Tzavoa Cave. Between 67 ka and 65
ka, FI D values rise to -23.0 and are the highe st in the Negev speleothems and high values of -31.1 of was also found between 35 ka and 32 ka. 8.2.7 Sr concentrations and 87Sr/86Sr ratios in speleothems, cave host rocks and soils for the Jordan Valley and the Negev Desert Sr concentrations in vadose speleothem s vary largely between 41 ppm and 1270 ppm (Table 4; Fig. 23A-D). Speleothem Sr concentrati ons in central and southern Negev range between 100 ppm and 326 ppm in the Basal member, and between 41 ppm and 110 ppm in the Intermediate member. Sr concentrations in speleothems of the Young member vary from 43 ppm to 557 ppm (Fig. 23A). High Sr concentrations (252-557 ppm) are found in Ashalim and Hol-Zakh Caves, whereas low Sr concentra tions (43-203 ppm) occur in Even-Sid-Ramon and Ma'ale-ha-Meyshar caves, Ktora Cr acks and Shizafon mini-caves. In younger speleothems the following ranges are observe d: at 350-290 ka betw een 43 and 200 ppm; at 225-187 ka between 111 and 557 ppm; at 135 -110 ka from 58 to 510 ppm, and lamina deposited at 89-86 ka episode, 203 ppm (Table 4; Fig. 23D). Sr concentrations in northern Negev range from 57 ppm to 527 ppm with 1270 ppm being observed in one speleothem lamina. A deta iled record of Sr c oncentrations during the last 200 ka derived from Tzavoa Cave speleothems shows si gnificant changes. Between 199 ka and 197 ka Sr concentrations are low (6678 ppm), increase to 166-173 ppm between 176 ka and 173 ka, and 354 ppm at ~157 ka. Between 140 ka and 123 ka, Sr concentrations decrease from 151 ppm to 57 ppm. Between 79 ka and 14 ka the Sr concentrations are high 257-527 ppm, with the highest valu es at 78 ka and 70 ka (335402 ppm), 38 ka (433 ppm), 21 ka (527 ppm ), 18-17 ka (450-462 ppm) and 14 ka (473 ppm). Sr concentrations of Ma'ale-Dragot Cave speleothems vary between 152 ppm (lamina dated to ~3.8 ka) to 1270 ppm (lamina dated to ~65 ka) (Table 4; Fig. 23C).
Sr concentrations in Jordan Valley speleo thems (Maâ€™ale-Efrayim) range between 100 ppm and 350 ppm, the lower values (115-154 ppm) are typical to th e period from 310 ka to 205 ka, and higher values (generally >150 pp m) from 160 ka to the 16 ka (Table 4; Fig. 23B). Sr concentrations in the phreatic speleothems range between 160 ppm and 2140 ppm. The value of 2140 ppm for speleothem in Makht esh-ha-Qatan (MKTC-5) is the highest in the Negev speleothems. Sr concentrations in host rocks vary between 78 ppm and 950 ppm. In Terra-Rosa soils above Ma'ale-Dragot and in Ma'ale-Efr ayim, Sr concentrations range between 121 ppm and 186 ppm, whereas in the Loess and Ha mada soils Sr concentrations are higher, between 314 ppm and 408 ppm (Table 4). 87Sr/86Sr ratios in Ma'ale-Efrayim Cave speleothems vary from 0.7082 to 0.7085. Between 310 ka and 205 ka 87Sr/86Sr ratios vary from 0.7082 to 0.7084. This is followed by generally steady line increase from 0.7083 to 0.7085 from 200 ka to 72 ka. Between 72 ka and 16 ka the 87Sr/86Sr ratios oscillate between 0.7083 and 0.7085, with highest values of 0.7085 at 72 ka, 63 ka, 38 ka, and 25 ka respectively (Fig. 24A). 87Sr/86Sr ratios in the vadose speleothems from Negev caves vary from 0.7077 to 0.7085. In Basal member 87Sr/86Sr ratios increase from 0.7077 in the base to 0.7079 at the top, in Intermediate member the values vary between 0.7081 and 0.7085, and in the Young member â€“ between 0.7079 and 0.7085. Low values of 0.7079-0.7080 at 350-310 ka rise to 0.7082-0.7083 between 310 ka and 170 ka, increase to 0.7085 at 160-155 ka and decrease to 0.7082 at 132 ka. Between 132 ka and 110 ka 87Sr/86Sr ratios oscillate from 0.7082 to 0.7084 and between 110 ka and 13 ka show gradua te but irregular increase from 0.7082 to 0.7085, with high peaks of 0.7085 at 63 ka, and 16 ka. In ~3.8 ka stalactite lamina MD(1)-8-A1, Ma'ale-Dragot caves, the 87Sr/86Sr ratio drops to 0.7081 (Table 4, Fig. 25A, B).
87Sr/86Sr ratios in phreatic speleothems from Ma'ale-ha-Meyshar Cave range between 0.7077 and 0.7078, and in Makhtesh-ha-Qatan Cave (MKTC-5-C) the 87Sr/86Sr ratio is 0.7079 (Table 4). 87Sr/86Sr ratios of the cave host rocks range between 0.7074 and 0.7078 (Table 4 and Figs. 24B, 26) and typical of Cretaceous and Eocene carbonate rocks (Koepnick et al., 1985). 87Sr/86Sr ratios in the bulk soils above the caves vary from 0.7084 to 0.7110. The 87Sr/86Sr ratios in the bulk Loess and Hamada desert soils range between 0.7084 and 0.7091. The 87Sr/86Sr ratios in bulk Te rra-Rosa soils are higher, 0.7098 and 0.7110, with 0.7098-0.7100 above Ma'ale-Dragot Caves in northern Negev, and 0.7110 above the Ma'ale-Efrayim Cave in Jordan Valley. 87Sr/86Sr ratios in silicate fraction of soils range between 0.7101 and 0.7123. In silicate fraction of the Negev soils: Loess, Hamada soils and Terra-Rosa soil above Ma'aleDragot Cave 87Sr/86Sr ratios range between 0.7101 a nd 7115, whereas higher value of 0.7123 is found in silicate fractio n of the Terra-Rosa soil in Jordan Valley (Ma'aleEfrayim Cave) (Table 4; Figs. 24B, 26). The si licate fraction of paleosols washed to the caves ranges between 0.7098 and 0.7140 (Table 4). 8.3 Discussion 8.3.1 Humid episodes in central /sou thern Negev, and Judea deserts 220.127.116.11 Speleothem deposition as recorder of effective precipitation Periods of speleothem deposition in arid environments are indicative of water availability in unsaturated zone (Holmgren et al., 1995; Fleitmann et al., 2003; Vaks et al., 2003, 2006 and 2007). In this study the term â€œhumid/wet" period is used to define time intervals with a positive effective precipit ation, i.e. in which sufficient water enters the unsaturated zone for speleothem de position to occur. Correspondingly, the term "arid/dry" period refers to a negative effective precipitation where insufficient water
enters the unsaturated zone, for of speleoth em growth to occur. During the "humid/wet" period the infiltration coefficients (the precipitation fraction entering the groundwater) are greater than zero, whereas during the "arid/dry" period they are close to zero. Humid and arid episodes have been correlat ed with the glacial/interglacial global climate change framework of Marine Isotopi c Stages (MIS) as defined by Shackleton and Opdyke (1973), Martinson et al. (1987), and Fontugne and Calvert (1992). At present, in the Mediterranean climate zone in northern and central Israel ~1/3 of the annual rainfall reaches the unsaturated zone and the remaining 2/3, i.e., about 300-350 mm in Soreq Cave area is lost either by ev aporation and/or by r unoff (Ayalon et al., 1998, 2004). Despite this loss, the amount of water reaching the unsaturated zone is sufficient to allow speleothem deposition today and throughout the Holocene (BarMatthews et al., 2003-a). The southernmost Holocene (3.8 ka) speleothem deposition occurred in Ma'ale-Dragot Caves located at the present-da y 280-300 mm isohyet (Fig. 1C) (Vaks et al., 2006). No present-day speleothem deposition occurs in semi-desert and desert regions of Israel that rece ive less than ~300 mm of annual rainfall. Thus, 300-350 mm isohyets can be defined as present-day limit of speleothem deposition in Israel, i.e. the zone with positive effectiv e precipitation is lo cated to the north of this boundary, and negative effective precipitation (infiltrati on coefficient of zero) to the south. During past interglacial periods, regional te mperatures were similar to those of the Holocene, as recorded by alkenone and G. ruber paleo-temperatures (Emeis et al., 1998, 2003; Kallel et al., 2000) and speleothem fluid inclusions (McGarry et al., 2004). Based on this evidence is possible to conclude that speleothem deposition during interglacial periods required at least 300-350 mm of annual rainfall (Vaks et al., 2006). During the glacial periods, regi onal temperatures were 5-10 C lower than at present (Emeis et al., 2003; McGarry et al., 2004). Thus, a main question is the minimum precipitation amount that was necessary for speleothem deposition during the glacial periods compared to the interglacial periods. At present, most infiltration to the
unsaturated zone aquifers occurs during mid-winter heavy rainstorms when air temperature is below 10C (Ayalon et al., 1998; 2004). A study made by Bonacci (2001) in Dinaric karst area, Istria Peninsula, Croatia, showed th at during the winter months when average temperature is ~3.5 C, infiltration coefficients are 0.9-0.95, i.e. almost all the rain and snow infiltrate to the aquifer. During the April and October when the temperatures are 10 C and 12C respectively (similar to pr esent day average winter temperatures in Israel), the infiltration co efficients are ~0.6, a nd during the July and August when average temperature is ~21 C, the infiltration coeffi cients drop to ~0.2. In the above region the annual precipitation is ~1000 mm and falls throughout the year. During years when most of the rainfall occu rred during the hot season, the mean annual infiltration coefficient was 0.36, whereas in ye ars where most of the precipitation fell in the cold season the mean coefficient wa s 0.76. Similarly, the study of Shuman and Donnelly (2006) in Massachusetts, US, showed that during years when most rainfall and snow falls in winter, the lake levels increase significantly relative to years when most of rainfall falls in summer. In the Oslo region, Norway most precipi tation during winter months occurs as snowfall (~200mm) and ra infall of ~600 mm occurs mainly during the warmer season (Norwegian Meteorological institute data http://www.worldweather.org/008/c00021.htm ). The infiltration is ~1/2 (i.e., 400 mm) of the total precipitation, and French and Binley (2004) (and references within) argue that 50% of the total recharge is from snowmelt, implying that additional 200 mm of recharge is from rainfall, thus emphasizing the importance of infiltration from snowmelt. The data of Bonacci (2001) suggest that the infiltration coefficients can increase by ~50-60% when the temper atures drop from 10-12C to 3-4 C. A similar ~50-60% increase in infiltration coefficients in the Mediterranean climate of Israel would reduce the amounts of water lost due to evapor ation/runoff by 75-100 mm and imply that 200275 mm compared to 300-350 mm in interglaci al periods, is the minimum amount of annual rainfall and snow required for speleothems deposition in glacial periods.
In desert locations vast bedrock surfaces can be found, and during the rain events bedrock surfaces collect runoff from larger areas. The runoff infiltrates to small soil pockets or cracks in the rock, leading to existence of limited local areas receiving an equivalent of 4-5 times higher rainfall amount (Yair, 1990). In cav es like Ktora Cracks, Maâ€™ale-ha-Meyshar, Even-Sid-Ramon or Makhtes h-ha-Qatan Cave this effect can cause increased water infiltration. However, all these caves are dry today and their major speleothem deposition periods (and major hiat uses) do not significantly differ from those of Ashalim Cave, above which bedrock surf aces contributing to runoff are not found. Thus, although it is possible that runoff infiltration from the bedrock surfaces is able to cause formation of speleothems even for minor increases in precipitation, it is likely that such events will be recorded in single caves, rather than in multiple caves. 18.104.22.168 Humid periods in central and southern Negev Desert Based on the above rough estimates of th e minimum precipitation necessary for speleothem deposition during glacials and in terglacials, it becomes possible to semi quantitatively assess variations in precipita tion amounts in central and southern Negev Desert and the "rain shadow" a nd Judea Desert during the glacial-interglacial cycles that occurred in the last 350 ka. As noted previously, it is possible to describe speleothem deposition events in the central and southern Negev De sert as either periods of intensive speleothem deposition, characterized by two or more speleothem ages from two or more caves in the region; or periods of limited speleothem deposition, characteri zed by a single speleothem age. Periods of intensive speleothem deposition are assumed to have occurred when regional precipitation amounts were higher than the minimum precipitation value, leading to positive effective precipitation in the entire region and the formation of numerous speleothems. Limited speleothem deposit ion occurred when the precipitation amounts were lower in the wider regi on, but high enough to enable local speleothem deposition.
Such a scenario could occur when runoff from vast bedrock surface contributes water to the small area (Yair, 1990), as de scribed in section 22.214.171.124. Based on the numbers of speleothem ages for various caves, the timing of deposition, and the correlation with global climate changes, five groups of rainfall/speleothem formation regimes during the last 350 ka in th e central and southern Negev Desert can be defined: 1) Intensive speleothem deposition periods dur ing previous interg lacial periods, defining the most humid events with annual precipitation exceeding 300 mm. These events occurred between 350 ka and 310 ka (i nterglacial MIS-9), be tween 220 and 190 ka (interglacial MIS-7.3, 7.2 and 7.1), and betw een 142 and 109 ka (including glacial MIS6.1 â€“ interglacial MIS-5.5 transition (Termination II) (Usami et al., 1998), interglacial MIS-5.5, 5.4 and the beginning of MIS5.3) (Figs. 11A, B; 12A, B, C). 2) Intensive speleothem deposition during pr evious glacial periods, most probably reflecting at least 200-275 mm of annual precipitation. This is represented by speleothem growth periods between 310 ka and 290 ka (gl acial MIS-8.5) (Figs. 11A, B; 12A, B, C). 3) Episodes of limited speleothem deposition during interglacials, possibly representing events with precipitation higher compared to the present, but not as high as in the case 1. These episodes occurred at 247.4 10.8 ka (probably at the transition from glacial MIS-8.1 to the interglacial MIS-7.5), 225-220 ka (beginning of the interglacial MIS-7.3), 190-185 ka (end of the interglacial MIS-7.1), and 87.7 1.4 ka (transition from interglacial MIS-5.2 to MI S-5.1) (Figs. 11A, B). 4) Limited speleothem deposition during glacial periods with less precipitation than in cases 1, 2 and 3, but possibl y more that at present. These humid events occurred between 290 ka and 285 ka (transition from glacial MIS-8.5 to MIS-8.4), and at 157.2 3.8 ka (glacial MIS-6.2) (Figs. 11A, B).
5) Periods with no speleothem deposition (hiatuses) due to dry conditions. These hiatuses occurred during the following periods: 285-255 ka, 240-225 ka, 185-160 ka, 155-142 ka, 109-89 ka, 86-0 ka. Types 1 and 2 are defined in this study as â€œNegev Humid Periodsâ€ (NHP), as follows: NHP-4 from 350 ka to 310 ka; NHP-3 from 310 ka to 290 ka; NHP-2 from 220 ka to 190 ka, and NHP-1 â€“ from 142 ka to 109 ka. During two older humid periods NHP-4 and NHP-3, the speleothem ages have large analyt ical errors. Their sub-division is based on the fact that NHP-3 speleothem age frequency is the highest at ~295 ka, i.e. glacial MIS8.5, whereas the NHP-4 age frequency is th e highest at ~320 ka and ~342 ka, i.e. interglacial MIS-9 (Fig. 12C). During peri ods older than 350 ka large analytical uncertainties of speleothem ages do not en able to determine wh ether the speleothem deposition was intensive or limited. Based on the age frequencies, number of lami nae, their thickness, and distribution in several caves it is possible to determine whic h of these four humid periods were the more humid. Speleothems from NHP-1 (Vaks et al., 2007) are found in six caves and the number of laminae and the age frequencies are the highest, their laminae are relatively thicker, suggesting they repres ent the wettest period. However, their cross-section is not continuous and alternate with short hiatuses suggesting clusters of short wet episodes with long droughts in between them. Speleoth ems from NHP-2 are f ound in four caves, their age frequencies are lower, and their cross section is <1.5 cm, possibly suggesting conditions were slightly drier than duri ng NHP-1. Speleothems fr om NHP-3 and NHP-4 are found in two and four caves respectivel y, but their laminae are thin, and age frequencies are low, suggesting drier condi tions than during NHP-2 and NHP-1 (Fig. 11A, B; 12A, B, C. NHP-4 was at its most intense in the southern Negev and more limited in the central Negev. This contrast s with NHP-3, 2 and 1, which were most intense in the central Negev, but limited or absent in the southern Negev. The short durations of the NHPs are is in accord with th e absence of calcite horizons in the regolith
soils of the southern Negev of Middle-Late Quaternary (Amit et al ., 2006; Enzel et al., 2007). 126.96.36.199 Humid periods in Judea Desert 13 speleothem samples were dated in Kanaim Cave, Judea Desert. Their ages cluster at: 354 19 ka (MIS-9-10), 324 10 ka (MIS-9), four ages cluster between 270 ka and 240 ka (MIS-8), 207.5 3.8 ka (MIS-7), thre e ages cluster between 160 ka and 140 ka (MIS-6), 131.4 2.1 ka (Termination II and MI S-5.5), and two ages are 37.90 0.25 ka and 31.46 0.22 ka (late MIS-3) (Fig. 10A). Thus, most of the speleothem deposition in Judea Desert occurred during glacial periods, similar to the speleothem deposition in the northern Negev (Vaks et al., 2006) and Jordan Valley (Vaks et al., 2003). However, their cross sections are thinner, suggesting that the paleo-climate in the Judea Desert was slightly drier compared with the northern Negev and the Jordan Valley. Speleothems deposited during interglacial periods in Kanaim Cave (Fig. 10A) overlap with NHP-4, NHP-2 and NHP-1 from the central and southern Negev, with speleothem deposition periods of 200-196 ka and 137-123 ka from th e northern Negev, and with speleothem deposition periods of 225-205 ka and ~135 ka in Jordan Valley, as found by (Vaks et al., 2003, 2006) and in this study (Table 1, Figs. 10B, 11). 8.3.2 Origin of precipitation in the Nege v Desert under presen t-day conditions and during the past 550 ka 188.8.131.52 Present-day rainfall The origin of present-day rainfall in Israel is determined by D 18O) relationship (Gat and Carmi, 1987; Ayalon et al., 1998). Prec ipitation from rain clouds forming above the EM Sea follows the local Mediterranean Meteoric Water Line (MMWL) with average d-excess values of +22 (Gat and Carmi, 1987; Ayalon et al., 1998), and sometimes even >+40 (Bar-Matthews et al., 2003-a; K uperman, 2005) but with the same slope of
the global Meteoric Water Line (MWL) ( D=8 18O+10) (Craig, 1961) with d-excess values of +10. D-excess in rainfall from tr opical origin (African monsoon) follows the MWL (Njitchoua et al., 1999), and the 18O values in rainfall south of Israel (Eastern Africa and Arabia) vary from -2 to +2 ( GNIP database, Aggarwal et al., (2007)), thus isotopical signature of the rainfall from tr opical origin is differe nt from precipitation originated in EM Sea. Evaporation below th e clouds causes the d-excess to be lower and the slope differs from the MMWL and the MWL (these are referred as evaporation lines). This is typical to present-day rainfall in the Jordan Valley (Vaks et al., 2003). Most rain events in the northern Negev (Be' er-Sheva and Arad), including the majority of rain events of >10 mm, have d-excess between +16 and +34, and follow the MMWL. Only a small fraction of the rain events follow the MWL, or fall on evaporation lines with d-excess values as low as -13 (Table 2 and Fig. 20). This includes rain events with <10 mm, but also some rain even ts of >10 mm (in this study, 5 events). The ranges of 18O and D values are such that it is impossible to differentiate between rainfall from tropical origin and events that underwent evaporation. Accordingly, data on the origin of each synoptic system was obtaine d from Israel Meteorological Service (ref: http://www.ims.gov.il/ ), Israel Weather (ref: www.israelweather.co.il ) and the Skiron Forecasts of University of Athens (ref: http://forecast.uoa.gr/forecastamg.html ). Three events with >10 mm were of tropical origin (Fig. 20), and two events associated with Atlantic-Mediterranean fronts o ccurred during the warmer months of October and April and their d-excess thus reflects evaporation. Particular events at Be'er-Sheva station gave the most negative d-excess values (-8 to -13), with the data defining an evaporation line with an equation: D=4.918O+4.0 (Figs. 20, 21C). The average (weighted mean (Ayalon et al., 1998)) isotopic compositions of Be'erSheva and Arad rains give d-excess valu es of +20.2 and +19.4 respectively (Figs. 20, 21C). These values are close to MMWL but slightly lower than average d-excess of
rains from central and northern Israel, possibly because of the combination of evaporation effects and the inclusion of several events of tropical origin. In the central and southern Negev Desert, rain events have d-excess between +41 to -12.6 (this study and Kuperman (2005)) (Table 2, Fig. 21-A, B), with averages of +15.1 in Mitzpe-Ramon, +13.4 in Makhtesh-ha-Qatan (this study and Kuperman (2005)) , + 8.8 in Neot-Smadar, and +13.2 in EinNetafim spring water near Elat. The average isotopic compositions of the rainfa ll in Mitzpe-Ramon, Makhtesh-ha-Qatan and Neot-Smadar follow the evaporation line defined for Be'er-Sheva rainfall (Fig. 21C), which makes it difficult to differentiate between rainfall of tropical or igin and evaporated rain. Based on data of Israel Meteorological Service, Israel Weather and Skiron Forecasts of University of Athens in Mitzpe-Ramon more than half of the rainfall originated in EM Sea, whereas in Makhtesh-ha-Qatan and in Neot-Smadar ~60% of the rainfall was from tropical origin. However, Kuperman (2005) reported that during th e winter 2002-2003 at Makhtesh-ha-Qatan the average d-excess of rainfall was +27, indicating that significant amounts of EM rainfall can reach this region. The dexcess value of EinNetafim spring water near Elat is +13.2, probably suggesting mixing between EM and tropical origin rainfall. This spring, like the speleothems, represents relatively large rain events that infiltrate into the ground and not undergo significant evaporation. 184.108.40.206 Origin of precipitation in the past Speleothem FI 18OD relationships can be used as a major proxy to define the precipitation source (Genty et al., 2002; Matthews et al., 2000; McGarry et al., 2004). Since only the D of FI was measured in this study, the 18O of the FI water were calculated from the measured 18O of the speleothem calcite, using the paleo-temperature equation of (O'Neil et al., 1969): (2) 18Ocalc 18Ow = 2.78 106 T-2 2.89
Where 18Ocalc is 18O of the speleothem, 18Ow is 18O of the paleo-cave water, and the T is cave temperature in Kelvin. Cave temperatures approximate average annual surface temperatures (Poulson and White, 1969; Gascoyne, 1992). The paleotemperatures used in the 18Ow calculation were taken from sources such as EM Sea surface temperatures (SST) (Emeis et al., 1998, 2003; Kallel et al., 2000; Kolosovsky, 2003), which approximate land temp eratures, and terrestrial pa leo-temperatures estimated from FI in caves from central and northern Israel and the Jordan Valley (McGarry et al., 2004). Correction for different topographic elevations of cave sites are made by subtracting 1 C for each 150 m change asl. For the time period between 161 ka and 157 ka there are no studies of the paleo-temperatur e of the EM Sea or other caves in Israel, and the calculated temperature from Western Mediterranean were used (Martrat et al., 2004) assuming that SST in western Mediterranean were ~4-5 C lower than in EM. With this assumption, paleo-temperatures from this period were 3-4 C lower than present. The calculated paleo-temperatures for each cave are summarized in Table 3. Because of inherent inaccuracy in the temperature calculations a 2 C range of temperature was taken for each time period (McGarry et al., 2004). The results are summarized in Fig 27A-D. 18O and D of the paleo-cave waters show a wi de range, with d-ex cess values above the MMWL and values below the MWL (Figs. 27A, 28). However, it is possible to characterize these relationshi ps for each time period and for individual caves. For glacial periods only Tzavoa Cave was studied. All samples from Tza voa Cave either fall above the MMWL (between 177ka and 172 ka), or follow the MMWL during the last glacial (67-65 ka, 64-61 ka and 35-32 ka) (Figs 27B , 28). NHP-1 was studied in Tzavoa, HolZakh, Ashalim and Maâ€™ale-ha-Meyshar showing th at in Tzavoa also during interglacial NHP-1, similar to the glacial periods, the 18OD relationships follow the MMWL, indicating an EM source. NHP-1 samples in Ashalim and Maâ€™ale-ha-Meyshar caves either fall above or close to the MMWL, also implying an EM source. One sample in
Hol-Zakh aged 132-131 ka falls between the two lines (Figs. 27C, 28). During NHP-2, Tzavoa and Maâ€™ale-ha-Meyshar caves fall n ear the MMWL whereas Ashalim cave shows completely different pattern with all samp les falling on and below the MWL (Figs. 27D, 28). This could imply a tropical source or evaporation. The main difference between the FI from Ashalim Cave during NHP-2 and thos e from this period at Tzavoa and Ma'aleha-Meyshar caves is that 18O values of the first are ~ 1.5 to 3 higher (-7.3 -8.1), (Fig. 27D). It is not clear why the Ashalim speleothems are 18O enriched, since evaporation will also lead to significant increase in their D values. Also, Ashalim Caveâ€™s location between Tzavoa Cave and Ma'ale-ha-Meyshar Cave (Fig. 1C) make it difficult to suggest different origin for ra infall. A possible explanation is cooler temperatures. The entrance to the Ashalim Cave has a funnel shape up to ~6 m depth and the cave chambers are located below the f unnel. This cave morphol ogy is known to trap cold air during winter (Forbes, 1998) and probably caused the 18O values of Ashalim speleothems to be 1-2 more positive than in other caves during NHP-1. It was impossible to measure D of fluid inclusions of speleothems between 220 ka and 550 ka owing to their very thin laminae. Another important evidence fo r a dominant EM Sea precipitation source comes from the petrography of NHP-1 and NHP-2 speleoth ems, which show a decrease in their volume along a north-south transect (Fig. 29AI). To the south of Maâ€™ale-ha-Meyshar Cave, laminae of NHP-2 and NHP-3 disappear from the caves and the cross sections of speleothems from NHP-1 decrease from >1.5 cm to a few mm. Only the NHP-4 speleothems are more common in southern Negev than in cen tral Negev, possibly indicating an increased supply of moisture from the tropical source (Fig. 29-J). Thinning of speleothem cross sections from the north to the south is a part of a regional scale phenomena in which the length of the depositio nal periods (Fig. 29-J) and the thickness of entire cross sections of speleothems from Israeli caves decrease drastically along the north-south transect from central Israel to the central and southern Negev. In Soreq Cave
the height of largest stalagmites (from base to the top) exceeds 8 m, whereas in the northern Negev (Ma'ale-Dragot, Tzavoa and Iz zim Caves), the largest stalagmites are ~22.5 m in height. Further to the south large stalagmites are ab sent from the caves and the widest cross sections of stalagmites, flowst one and stalactites from the bedrock to the youngest laminae do not exceed 0.5 m. Other criteria for the origin of rainfall in the Negev speleothems come from the comparison of their 18O profiles with those of speleo thems from central and northern Israel (Bar-Matthews et al., 2003-a) and the northern Negev (Vaks et al., 2006) . In central and southern Negev there is a large differe nce between the glacial (-1.5 to -5) and interglacial (-6.8 and -11.4) 18O values. The variations between glacial and interglacial 18O values are similar or even higher than in speleothems from central and northern Israel (Bar-Matthews et al., 2003-a) and from the northern Negev (Vaks et al., 2006) (Figs. 15A, 30A). This im plies that the cave water 18O values reflect the glacialinterglacial variability originating in the same rain source: the EM Sea. On the other hand, during NHP-2 and NHP-1 18O profiles of speleothems from the central Negev have lower variability compared with speleo thems from central and northern Israel and the northern Negev (Fig. 30). In the latt er, there is a sharp change during the Termination-II at 135 ka, and at 127 ka, with minimum values between 128-121 ka whereas in the central and southern Negev, there is very little change throughout the entire period from 133-115 ka (Fig. 30). Moreover, the 18O values of speleothems from central Negev for this period are lower by 1-3 compared to those from central and northern Israel, but similar to t hose of Tzavoa Cave in the no rthern Negev (Figs 30 B, C). The lower variability of 18O values of the speleothems from the central and southern Negev could may imply that the rainfall source (at least at the beginning and the end of NHP-1 and NHP-2) was of tropical origin, and during the peak of these humid periods the EM source became dominant. This patter n is not evident from the FI data, most probably because of its lower resolution, but also contradicts the speleothem thickness
trend from north to south. A more likely possi bility is that Raylei gh fractionation isotopic processes (Vaks et al., 2006) and amount e ffect (Dansgaard, 1964; Bar-Matthews et al., 1997, 2003-a) mask the fluctuation in the 18O of rainfall as source indicator. A correlation of the lowest 18O values with peak speleothem age frequencies (Fig. 30 B, C) at 208-197 ka and at 132 ka , 126 ka and 117 ka, supports increased rainfall amounts during these time intervals. The decrease in 18O of speleothems from central Israel to the Negev during NHP-1 and NHP-2 contrasts with the north-south increasing 18O trend observed for present-day rainfall. This contrast can be explained by th e very small present-day rainfall amounts, which enable evaporation processes to obscure possible Rayleigh fractionation isotopic effects. Rayleigh fractionation isotopic processes with increasing distance from the EM Sea coast / elevation occur in central and northern Israel, where precipitation amounts are large (Ayalon et al., 200 2; Vaks et al., 2006). To conclude the speleothem record indicates that EM Sea was the major source for the precipitation in the central Negev, with a po ssible minor contribution of rainfall from tropical sources in the southern Negev. 8.3.3 Vegetation in the central and south ern Negev Desert determined from the 13C of the speleothems and its relati on to regional climatic change Speleothem 13C values in central and southern Negev Desert vary between 1.5 and -8.4 and in most cases are much higher than those of the northern Negev (Vaks et al., 2006), Jordan Valley (Vaks et al., 2003), and ce ntral and northern Is rael (Bar-Matthews et al., 2003-a) (Figs. 18, 19, 31). 13C values of speleothems from central and northern Israel were interpreted to reflect both the vegetation type and soil-water rock interaction (Bar-Matthews et al., 1996, 1997, 2003-a; Frumkin et al., 1999; 2000). 13C values between -9 to -13 were interpreted to us ually reflect C3 type vegetation and higher values were attributed to an increase in the proportion of C4 type vegetation. However,
(Bar-Matthews et al., 2000, 2003-a) also argue that during certain time intervals that correlate with periods of sa propel formation in the EM Se a (sapropel 1 between 9-7 ka, and sapropel 5, during MIS-5.5 between 127-120 ka), the combination of lowest 18O values with highest 13C values (-7 to -9 and 1 to -6 respectively), is indicative of deluge periods with increased 13C contribution from host rock and atmospheric CO2. Frumkin et al. (2000) on the other hand relate this combination during sapropel 5 to very dry conditions. In the northern Negev, 13C values range from ~0 to -9.5 with an average of -6 (Vaks et al., 2006) . The higher 13C values from 1.5 to -8.5 in central and southern Negev (Figs. 18, 19, 31) are considered to reflect either of the dominance of scarce C4 vegeta tion when the values are in the range of -2.5 to -8.5 , and/or higher contribution of host rock and atmospheric CO2 when 13C values are isotopically heavier. The highest speleothem 13C values usually occur at the beginning of each NHP and decrease during the period to the lowest valu es at its end (Figs 18, 19, 31). This change correlates well with the highe st speleothem age frequenc ies occurring at 208-197 ka, 133125 ka, 118-116 ka (Fig 31B,C), suggesting that the lowering in 13C is indicative of increased precipitation and the development of more vegetation above the cave. As the frequency of speleothem deposition decreases between 126 ka and 118 ka, 13C shows a corresponding increase. This period is one of the warmes t in the region during the last 220 ka with EM average SST of 22-24 C and annual mean of 17-22 C in Soreq Cave area (Emeis et al., 2003; McGarry et al ., 2004); this is sugge stive of increased evaporation, and hence lower e ffective precipitation, relative to the periods at 133-126 ka and 118-116 ka. A decrease in speleothem 13C between 205 ka and 197 ka and between 128 ka and 126 ka in northern and central Negev is associated with the sharp increases in speleothem 13C (from -11.5 to -6.5 and from ~-13 to 1, respectively) in Soreq and Peqi'in Caves from central and northern Israel (Fig. 31B, C), and in Jerusalem Cave (Frumkin et al., 2000). As noted earlier, Frumki n et al. (2000) interpreted the latter
increase in speleothem 13C in the Jerusalem Cave to be a result of aridif ication, reduced vegetation, forest fires and the stripping of the soil cover. However, the fact that speleothems grew in the northern and central Negev, indicate that during this period the Negev Desert was relatively we t. Moreover, speleothem flui d inclusions from Tzavoa Cave during this period have d-excess rang ing between 53 to 34, and in the more southern caves d-excess values range betw een 37 and 14 (Fig. 28), indicating that the EM Sea was the major moisture source for the Negev. The fact that 13C values in Tzavoa and Hol-Zakh decreased from -2 to -7.5 and further to the south the 13C values increased from ~ -3.8 to ~ -2 may su ggest a relative decrease in rainfall from the northern Negev, which is closer to the EM Sea compared to the central Negev. Thus the speleothem record from the Negev desert probably indicates an increase in intensity of the Atlantic-Mediterran ean (Cyprus) cyclones resulted in extremely wet conditions above the EM Sea as eviden t also from the formation of EM sapropels which are suggested to reflect increase precipitation over the entire Medi terranean Sea (K allel et al., 1997, 2000). Bar-Matthews et al., (2000, 2003-a) interpre ted the increase in 13C values to reflect deluge events in central and northern Israel that resulted in fast water infiltration to the cave with little interaction with soil CO2. In northern Negev the amounts of rainfall also increased but were lower, with no deluge effects, resulting in increase in vegetation cover and decrease in speleothem 13C values. The increase in rainfall in the region is also supported by Rossignol-Strick and Patern e (1999) who recorded an increase in deciduous oak pollen in sapropel sediments indicating precipitation amounts of 650 mm or higher. Only short episodes of speleothems formati on occur in the centra l Negev in Makhteshha-Qatan Cave at ~157 ka (glacial MIS-6.2), with 13C values of -8.4 (Fig. 18, 31A). These are the lowest values known for the central Negev speleothems. This speleothem is phreatic and the cave was probably located be low the stream channel of Wadi Hazera. Therefore it is possible that the low speleothem 13C is indicates vegetation that locally
developed in the wadi channel and does not reflect the entire region vegetation. Another short event of speleothem deposition in the central Negev occurred between 89-86 ka; the speleothems 13C values are ~ -4.2 to -5.6, suggesting the dominance of C4 semidesert type vegetation. 8.3.4 Correlation between the Negev Humid Periods and the regional and global climate changes Good correlation was found during the NHP-1 between the timing of 18O decrease between MIS-6.1 and MIS-5.5 at 132 ka in the Soreq Cave speleothems and the timing of maximum speleothem deposition in the Ne gev. The second high peak of speleothem deposition in the Negev at 127126 ka correlates with the 18O decrease to the minimal values at the climax of MIS-5.5 in Soreq Cave. There is also a good correlation between one of the negative peaks of the 18O in Peqiâ€™in Cave at 200 ka and the highest peak of the speleothem deposition in the Negev during the NHP-2 (F ig. 30). Bar-Matthews et al. (1997, 2000, 2003-a) interpreted these decreases in 18O values as associated with increased rainfall in central and northern Israel. As was shown in chapter 220.127.116.11, the precipitation in the Negev Desert during the NHPs originated at EM Sea, indicating that strongest humid episodes above the EM Sea brought the precipitation to the Negev Desert. Not all humid periods described by Bar-Matthews et al., (1997, 2000, 2003-a) reached the Negev Desert because of its southern position. Episodes of speleothem deposition NHP-s 4, 3, 2 and 1 in the cen tral and southern Negev Desert are very short and coincide with: periods of high amplitude oscillations in the solar energy in the northern and southern hemispheres (B erger, 1978) (Fig. 32), peaks of northern Hemisphere insolation, an African monsoon index of 51 cal/cm2 day (Rossignol-Strick (1983), for e xplanation see caption of Fig. 33) and the formation of the Mediterranean sapropels (Rossignol-S trick and Paterne, 1999) (Fig. 33).
NHP-4, 3, 2 and 1 occurred contemporaneously with speleothem deposition in Hoti Cave, northern Oman, at 330-300 ka, 200-180 ka , and 135-120 ka (Fig. 11C â€“ Ref. 1). Fleitmann et al. (2003) suggested that speleothem deposition coincided with the northward migration of the ITCZ and intens ification of the Indian Monsoon. Comparison between the timing of NHPs and humid periods from other parts of the southern SaharanArabian Desert such as north-western Sudan and central and southern Egypt, shows that wet periods in the southern part of the Sahara desert coincide with wetter conditions in the central and southern Negev (Fig. 11, Refs . 2-4) ( Szabo et al., 1995; Crombie et al., 1997; Osmond and Dabous, 2004); however the larg e age uncertainties of the latter does not enable a precise correlation with NHP-s. Since this study demonstrates that the EM was the major precipitation source to the Negev Desert, this link between maxima of the African monsoon and increased precipitation from Mediterrane an sources suggests that the intensity of AtlanticMediterranean Cyclones must have increased contemporaneously with the African Monsoon. This was also suggested by Almogi-L abin et al. (2004), Bar-Matthews et al. (2000, 2003-a), based on the timing of the anomalously low 18O and anomalously high 13C values of speleothems from the central an d northern Israel (Fi g. 34). However, all the NHP commenced several thousand years befo re the peaks of maximum insolation and the beginning of the sapropel formation, and they ended several thousands year afterward (Figs. 34, 35). This suggests that the NHP were triggere d before the monsoon maxima. A change in the isotopic composition of spel eothems from central and northern Israel during Termination II and the transition from MIS 7-2 to MIS 7-1, and the sharp increase in the age frequencies during NHP-1 and 2, at 204 ka and 135 ka (Fig. 34), coincides with rapid global warming recorded by speleo them deposition in Alps (Spannagel Cave, Austria) (Spotl et al., 2002), the rise of the global sea level (Henderson and Slowey, 2000; Gallup et al., 2002; Antonioli et al., 2004) , the start of speleothem deposition in
Oman (Fleitmann et al., 2003), and the timing of speleothem deposition in northern England (Baker et al., 1993). The contemporaneous tropical and mid-la titude humid periods can have a common cause. During higher northern hemisphere su mmer insolation periods, the SST in subtropical Atlantic Ocean were higher causing the Azorean high pressure cell to weaken, enabling the intensiv e monsoon rainfall in the sout hern Sahara during the summer (deMenocal, 1995; 2004). The weaker Az orean high pressure cell potentially can lead to a more negative NAO i ndex during winter, resulting in increasing rainfall above the Mediterranean Sea (Hurrel, 1995; Krichak and Alpert , 2005) and consequently above the Negev Desert. During the Holocene an d MIS-11 (430-400 ka) no speleothems formed in the central and southern Negev Desert. It could be explained by lower northern Hemisphere insolation, lower African Mons oon Index (Fig. 32, 33) and stronger Azorean high pressure cell, which prevented from Atlant ic-Mediterranean cyclones to reach as far to the south, as during the NHPs. 8.3.5 "North â€“ south paradox" of the Negev paleoclimate The speleothem record of northern Negev, J udea Desert and Jordan Valley shows that speleothem deposition was inte nsive during most of the two last glacial periods MIS-6, and MIS-4-2. These glacial speleothems are thick, indicating high effective precipitation. They were deposited at the time of high le vels of Lake Lisan between 70-17 ka (Neev and Emery, 1967; Begin et al., 1974; Bartov et al. 2002, 2003; Bookm an et al., 2006) and Lake Amora, the precursor of modern Dead Sea, which existed during MIS-6 (Waldman et al., 2007). Speleothems were also deposited during interglacials, mainly at 200-186 ka, 137-123 ka, ~118 ka and 85-76 ka, but their deposition was episodic compared to glacial speleothems, and their cross sections are thi nner, indicating lower effective precipitation. During the MIS-5, the levels of the Lake Samra (the precur sor of the modern Dead Sea
during MIS-5) were also lower, apart from a minor increase between 130-120 ka (Waldman et al., 2007), compared to the levels of glacial Lake Lisan. In contrast to the northern Negev, Judea De sert and Jordan Valley, most deposition in the central and southern Nege v Desert occurred during interglacials, with almost no deposition during the two last glacial periods. The cross sections of these interglacial speleothems are even thinner than those of the interglacial speleothems from northern Negev, indicating that the central and sout hern Negev speleothems were deposited under lower effective precipitation; nevertheless their deposition occurred throughout the entire Negev Desert. The absence of glacial speleo thems from the central and southern Negev, when the climate was apparently wetter in the northern Negev and "rain shadow" desert, and the corresponding formation of thin inte rglacial speleothems in the central and southern Negev Desert, when c onditions in the north were appa rently drier, is considered as the â€œnorth-south paradoxâ€. This apparent paradox is emphasized by the fact that the major source of precipitation of the entire region was the EM Sea. One of the factors limiting the penetration of rainfall and snow from the EM Sea to the central and southern Negev during the glacial periods could be the northern shift of the Sinai Mediterranean coast due to glacial sea level decrease (Fig. 36, En zel et al. (2007)). This however does not explain the penetratio n of rainfall to the entire Negev during interglacial NHP-s. An increase in precip itation to at least 300-350 mm associated with Cyprus cyclones could explain speleothem de position in the entire Negev Desert during the interglacials. The intensity of these cy clones during the glacial periods could have been lower, bringing less th an 200 mm of annual precipitation to the ce ntral and southern Negev, not enough for speleothem deposition. Th is in turn implies that the amount of precipitation during glacials in the northern Negev, Judea Desert a nd Jordan Valley was at least higher than 200 mm (as was di scussed in chapter 8.3.1), enabling water infiltration to the caves.
Low temperatures would result in higher infiltration coefficients during glacial periods, giving a higher influx of water to caves and aquifers, and allowing deposition of thick speleothems in the northern Negev, Jud ea Desert and the Jordan Valley. The high lake levels of the two last glacial precursors of the modern Dead Sea, Lake Lisan and Lake Amora (Waldman et al., 2007), were probably the result of the high infiltration coefficients and possible snowmelt runoff into the Dead Sea Rift. Low infiltration coefficients during interglacial periods could also explain th e low levels of Lake Samra during MIS-5 (Waldman et al., 2007). Kolodny et al. (2005) calculated the relativ e humidity above the Lake Lisan during the last glacial period to be ~0.88 compared with values from 0.46 to 0.625 calculated for the present day Dead Sea (Gat, 1984; Krumgalz et al., 2000). The higher relative humidity during glacial most probably led to less ev aporation of rainfall and from soil surface, resulting in increased amounts of water infiltrating to the caves and aquifers. This most probably led to a rise of the regional water table and higher lake levels. An increase in infiltration of the rainwa ter occurred also in the Negev highlands during the last glacial period, as evident by travertine deposition in the central Negev springs (Schwarcz et al., 1979). However, this increase in infiltration evidently was not enough to form speleothems, possibly because of the small recharge areas of the caves, compared to the large catc hments of the springs. Thus it is suggested that Ne gev "north-south paradoxâ€ can be explained as follows. During the last glacial periods Cyprus cyclones brought more than 200 mm of annual precipitation to the area of the present da y 150-160 mm isohyet (Tzavoa Cave). This amount was enough for the development of steppe vegetation above Tzavoa Cave. The fact that 13C values of the northern Negev speleoth ems are always higher (Vaks et al., 2006) than those of contemporaneous speleo thems from central and northern Israel, indicates that the amount of precipitation reaching the northern Negev during glacial periods must have been lower. In contrast, the 13C and 18O of the speleothems in the
Jordan Valley, are similar to those of central and northern Israel (Vaks et al., 2003) implying higher precipitation than in the northern Ne gev. In both regions low temperatures and high infiltration coefficients enabled water to enter the caves and form thick speleothems. In the drier Judea desert (Kanaim Cave) precipitation reached more than 200 mm only for short time episodes during the two last glacial periods, as evident from the relatively thin laminae. In the cen tral and southern Ne gev precipitation must have been below ~200 mm as evident from the lack of speleothem deposition apart from NHP-3. During the interglacials NHP-1, 2 and 4, the in tensity of Cyprus cyclones must have been higher compared with glacial periods , bringing precipitation of at least 300-350 mm as far south as present day 50 mm isohyet (Ma'ale-ha-Meyshar Cave). However, evidently temperatures were high and infiltration coefficients were low, thus bringing only small amounts of water to the caves and resulting in the formation of only thin speleothem laminae. 8.3.6 Pliocene-Pleistocene humid periods in the central and southern Negev Desert as evident from speleothems older than 550 ka The Negev speleothems of the Basal member with four preliminary U-Pb ages of 3.32.7 Ma (see chapter 18.104.22.168) must have been deposited under humid conditions, they are massive (up to 40 cm in width), contain large crystals with very little signs of hiatuses in between laminae. In the Intermediate a nd Young members the number of hiatuses increases and the laminae become thinner. The change in the appearance and the petrography towards the younger part are indicative of drying since middle Pliocene (Fig 3, 4). The drying-out of the Negev Desert during the last ~ 2.7 Ma occu rred together with the desertification of the Sa hara (deMenocal, 1995; 2004). The Basal Member was probably deposited contemporaneously with the Pliocene Arava Formation lakes and (Avni, 1997; Avni et al., 2001-b) suggested that these lakes
formed at 4 2 Ma. The lake sediments c ontain lacustrine marls including abundant freshwater fauna, and they c overed area larger than 1300 km2 in the southern Negev and Arava Valley. These lakes were formed under humid conditions except for one hypersaline event (Avni, 1997). This is very si milar to the massive appearance of the Basal member showing very few signs of hiatuses. The lamina ASH-15-C from the base of the In termediate member of Ashalim Cave has a preliminary U-Pb age of 1.269 0.013 Ma (see chapter 22.214.171.124) which suggests that these speleothems formed contemporaneously with lakes of Early Pleistocene Zhiha formation (Ginat et al., 2003). The major lake (Lake Zihor) formed in the southern Negev at ~1.4 Ma and covered area of 18 km2. The lake sediments show 3 periods of high stand with abundant freshwater fauna and periods of desiccation in between. The latter were marked by the formation of reddish so ils containing carbonate horizons. Annual precipitation varied from 150 mm to 500 mm dur ing the deposition of the lake sediments (Ginat et al., 2003). This is similar to the discontinuous formation of the Intermediate member, which shows more evidence of hiatuses than the Basal member. Middle-Late Pleistocene sediments in the s outhern Negev include series of alluvial terraces up to 20 m above the present day wadi channels. Gypsic-salic soils typical to present day hyper arid climate formed on this terraces. (Avni et al ., 2001-a) suggests that their formation was controlled mainly by clima tic variations. Thus it is possible that the Young member of the speleoth ems formed during these phases of terraces formation. The fact that several growth / hiatus cycl es can be observed in single speleothem over >3 Ma shows that in the Negev Desert en vironment the water c onduit system from the surface to the caves (fractures) seem to remain open over long times. It is the reflection of absence of prior precipitation of calcite in the cracks due to the arid conditions. 18O values of the Pliocene Basal member sp eleothems in Ashalim Cave range from -9.4 to -11.7 and 18O values of the Intermediate member in Ashalim Cave and Even-Sid-Ramon mini-caves are -9.3 -10. All fluid inclusion D values range
between -32 and -55. These values are sim ilar or slightly lower than those of the speleothems deposited during NHP-4, 2 and 1. The lack of independent temperature estimates for these periods make it impossible to reconstruct the sour ce of precipitation. 8.3.7 Phreatic speleothems in Makhtesh-ha-Qatan as mark ers of paleo-groundwater levels Phreatic speleothems of Makhtesh-ha-Qatan Cave were dated at ~290 ka and older, with one lamina deposited at ~157 ka. At present the cave is located 4 m above the present day channel of Wadi Hazera (an ephe meral stream). In the past, before the incision of the Wadi Hazera channel, the cave was probably located below the ancient channel. Speleothems deposition suggests high er ground water level, or periods when Wadi Hazera was a perennial stream and the cave was flooded. Kronfeld and Livnat (1987) found that tufa was deposited at Makhtesh-ha-Qatan area, about 1.5 km to the west, during the last gl acial period (between 73 ka and 30 ka) and during older episodes. However, no speleoth ems were deposited in Makhtesh-ha-Qatan Cave during the last glacial, probably becau se the precipitation amounts were too low to form vadose speleothems (like in central Negev, section 8.3.5) , and groundwater levels in Judea Group aquifer were too low to form phreatic speleothems. Because of large analytical uncertainties of Kronfeld and Livnat (1987) ages it is impossible to compare their older tufa ages to the speleothem ages determined in Makhtesh-ha-Qatan Cave. 8.3.8 â€“ Humid periods in the northern Sa haran-Arabian Desert as climatic â€œwindowsâ€ for Out of Africa disp ersals of hominids and animals Humid episodes in the Saharan-Arabian Desert were an important control of hominid migration from the African continent to the other parts of the world. This was discussed in detail as part of this thes is in the article published in Geology (Vaks et al., 2007). In this paper, the emphasis was on the critical importance of NHP-1. Similarly, all other
NHP-s and older humid periods could have pl ayed very important role as a â€œclimatic windowsâ€ controlling the dispersion of ho minids and animals out of the African continent. Derricourt, (2005) argues that SinaiNegev land bridge was the major and probably the only route out of Africa before the early modern humans developed the ability of seafaring during the last glacial period. The stud y of Fernandes et al. (2006) argues for the absence of post-Miocene land bridges across the Red Sea, with the exception of the Sinai-Negev Desert. Thus, we t climate conditions in this "bottleneck" region were of great importance during the Late Pliocene and Pleistocene. Today the Nile River provides a way across the Sahara, but the usability of this route by hominids and animals in the past was highly dependant on the inte nsity of the African monsoon. Interglacial phases of high monsoona l intensity increased the Nile flow, and other Sahara-crossing rivers possibly formed (Rohling et al., 2002). During the phases of low monsoon intensity, mainly during glacial periods, the flow of the Nile decreased, becoming discontinuous in extremely arid c onditions similar to Last Glacial Maximum (Almogi-Labin et al., 2004), when southward expansion of Sahara Desert reached 5-11 N (Roche et al., 2007). Said (1993) argues that the sources of River Nile did not reach as far as tropical Africa until about 800 ka. Thus, th e Saharan-Arabian Desert that began to form at ~ 2.7 Ma (deMenocal, 2004) could have been an even more significant barrier for migration out of Africa for the ~2 millions years, prior to the establishment of the connection between the Nile a nd the tropical Africa. This emphasizes the importance of wet phases in the Sahara and Sinai-Negev la nd bridge for the out of Africa dispersal during the Early Pleistocene. Hand axes, similar to those found in the Ubei diya site in northern Jordan Valley and dated at ~1.4 Ma, were also found by Ginat et al. (2003) (and references within) in sediments of the Early Pleistocene Lake Zihor in the southern Negev, identifying a probable route between the southern Ne gev Desert and the Jordan Valley.
NHP-3, 4 and earlier humid periods in the Negev also could be associated with migrations of hominids and animals. During the NHP-2 there is data showing that between 200 ka and 195 ka, humid conditions prevailed both in the northern and southern parts of Saharan-Arabian Desert (Szabo et al., 1995; Crombie et al., 1997; Fleitmann et al., 2003; Osmond and Dabous, 2004; and this study). This potenti ally opened the passage through Sahara to the Levant. Recently the early Mousterian site in Rosh-Mor, central Negev was dated to ~200 ka, suppor ting this idea (Rink et al., 2003). 8.3.9 Sr isotope ratios from Israeli desert speleothems as tracers of the origin of dust and host rock weathering Sr substitutes for Ca in th e speleothem calcite lattice (Fairchild et al., 2001). Three major sources potentially supply Sr to th e groundwater reaching the cave system: host rock, the soil above the cave and sea spray (Ayalon et al., 1999; Bar-Matthews et al., 1999; Frumkin and Stein, 2004). Sr contribution from sea spray (87Sr/86Sr = 0.7092) is most probably the highest near the coast and decreases inland. Sr isotopic ratios in Negev rainfall vary between 0.7079 and 0.7084, co mpared to 0.7078-0.7092 in coastal and northern parts of Israel (Herut et al., 1993), suggesting a very low contribution of the sea spray to the Negev rainwater relative to the nor thern and coastal region. Soils in Israel are dust-born, originating from distal Saharan-Arabian sources and from the proximal source of northern and central Sinai Peninsula (Yaalon, 1971; Yaalon and Ganor, 1973; Yaalon and Dan, 1974; Amit and Gerson, 1986; Tsoar and Pye, 1987). (Crouvi et al., 2007; Enzel et al., 2007) showed that dust that the forms the Negev loess soils mainly originates from wind blown sediments of the marine shelf of the Nile Delta and the northern and central Sinai Peninsula. 87Sr/86Sr ratios in the Negev desert speleo thems range from 0.7083 to 0.7085 for two last glacial periods at 1 60-155 ka and 75-13 ka; is lowe r and ranges between 0.7081 and 0.7084 during the glacial intervals between 310-285 ka and 176-172 ka and during the
interglacial periods 350-310 ka, 220-185 ka, 142-76 ka. Speleothem 87Sr/86Sr ratios in the Jordan Valley during two last glacial periods and interglacial intervals are similar to the Negev values. (Table 4, Fig. 24-25). In Sore q and Jerusalem Caves, central Israel the 87Sr/86Sr ratios in glacial and interglacial sp eleothems are very similar to each other (Ayalon et al., 1999; Bar-Ma tthews et al., 1999; Frumkin and Stein, 2004); during the two last glacial periods they range between 0.7083-0.7085 (similar to the desert speleothems), whereas during the penultimate and the Holocene interglacials the speleothem 87Sr/86Sr ratios are 0.7078-0.7080 (Fig. 37 ). The glacial-interglacial fluctuations of 87Sr/86Sr ratios in speleothems from central Israel are mached by similar trends in Sr concentrations. During glacial in tervals the Sr concentrations range between 55 and 200 ppm in the Jerusalem Cave (Frumk in and Stein, 2004) and 45 and 220 ppm in the Soreq Cave (Bar-Matthews and Ayalon, 2001), but during the interglacials, the Sr concentrations are lower betw een 7 and 55 ppm in Jerusale m Cave (Frumkin and Stein, 2004), and 50 and 170 ppm in Soreq Cave (B ar-Matthews and Ayalon, 2001) (Fig. 38). Similar trends, but with higher concentratio ns, are observed in the Negev speleothems. During the two last glacial periods concentrations range from 120 to 340 ppm in Ma'aleEfrayim Cave, and 255 to 525 ppm in Tza voa cave, whereas during the two last interglacials in Ma'ale-Efrayim Cave values are 100 to 260 ppm and Tzavoa Cave values are 60 to 150 ppm (Table 4, Fig. 23B,C, Fig. 38). Speleothem 87Sr/86Sr ratios of the Intermedia te and Young members (0.7079-0.7085) lie between the host rock ratios (0.7074-0.7078) and the bulk soils values (0.70840.7110) (Fig. 39A). 87Sr/86Sr ratios of Ma'ale-Efrayim and Tzavoa speleothems slightly approach the host rock composition during the interglacial periods but the soil compositions during glacials (Fig. 39C). Frumkin and Stein (2004) interpreted th e high glacial Sr concentrations and 87Sr/86Sr values to intensive dust fluxes from Sahara to the EM region due to the hyper arid conditions in the Saharan areas. Reduced speleothem Sr concentrations and 87Sr/86Sr
ratios during the penultimate interglacial a nd the early-middle Holocene, were interpreted to reflect reduced dust fluxes from Sahara du e to the intensive monsoon precipitation. In contrast, Ayalon et al. (1999) and Bar-Matthews et al. (1999) interpre ted the same trends to reflect increase in chemical erosion of the host rock during the interglacial humid periods, whereas during glacials weathering wa s less intensive and the contribution from exogenic sources increased. High interglacial 87Sr/86Sr ratios and Sr concentrati ons of the Negev speleothems relative to speleothems from caves in central Israel can be attributed to the fact that precipitation amounts in the Negev were always significantly lower. Because of the lower precipitation, chemical erosion of host ro ck was less significant, and the high Sr concentration and Sr isotopic values mainly originated from dust-born soil. Supportive evidence for low relative rates of chemical erosion in the central and southern Negev comes from the fact that speleothem cross-section become thinner southward by ~ 2 orders of the magnitude (see section 126.96.36.199), whic h is indicative of reduced rates of hostrock dissolution above the caves. It further supported by the study of Haviv et al. (2006) showing that the erosion rates in the arid a nd mildly arid zones in Israel (100-400 mm rainfall) are 1-4 m/Ma, whereas in Mediterran ean climate zone, central Israel (500-600 mm), the erosion rates increase to 15-40 m/Ma. The similar speleothem 87Sr/86Sr in caves from the Nege v Desert, Jordan Valley and central Israel during the glaci al periods (Fig. 37) suggests common sources. Frumkin and Stein (2004) suggest the Sahara Desert. However, other more proximal sources such as the Nile delta, and northern and central Sina i are possible (Crouvi et al., 2007; Enzel et al., 2007). The contribution of these sources is evident from the dust fall map of Ganor and Foner (1996; 2001) which shows an increa se in dust fall above the north-eastern Sinai and northern and central Negev relativ e to north-western Sinai (Fig. 40). During glacial periods vast parts of continental sh elf and the Nile Delta were exposed due to lower glacial sea level, and probably increased the dust contribution to the region during
the winter storms (Enzel et al., 20 07) (Fig. 36). The similarity between 87Sr/86Sr of present-day rainfall in the Negev Desert (0.7079-0.7084) (Herut et al., 1993) and the 87Sr/86Sr of glacial speleothems from all Is raeli caves (0.7083-0.7085) suggest that the main source remained the same. During the interglacial periods there is a marked difference in Sr isotopic composition between Negev speleothems and central Israel speleothems. Whereas in central Israel the 87Sr/86Sr values are closer to that of the host rock, in the Negev they only slightly decrease relative to the glacial values (Fig. 37, 39). This supports a higher dust contribution of dust during interglacials to the Negev. Even during periods when Saharan dust source was possibly shut down (Fru mkin and Stein, 2004), the Negev Desert continued to receive dust, most probably from the proximal Sina i source (Fig. 36, 40). The reason why such Sinai dust was not transported to the north is because it is composed of coarse silt particles (>20 m) traveling in low altitude, whereas dust from the SaharanArabian source is composed of fine particle s traveling in high alt itude in atmosphere (Frumkin and Stein, 2004; Enzel et al., 2007). During interglacial pe riods the coarse dust was transported from Sinai to the Negev by st rong south-westerly and westerly winds of Cyprus cyclones, but washed down by rain before reaching the central Israel. This is very similar to present day situation, where dust fall is maximal in the Negev, but sharply decreases to the north (Ganor and Foner, 2001; Enzel et al., 2007) (Fig. 40). Supportive evidence for this process is the present-day 87Sr/86Sr ratios of the Jordan Valley bulk soils and of their silicate compone nts, which are 0.7110 and 0.7123 respectively, and nearer to the Saharan dust end member (0.7165-0.7200) (K rom et al., 1999) than the Negev soils, with bulk ratios of 0.70840.7100 and 0.7101-0.7115 in silicate components (Figs. 24, 26, 39A, Table 4). This difference may indicate northward enrichment in the Saharan dust source, whereas in the sout h the Sinai proximal dust sour ce is more dominant. The 87Sr/86Sr and 1/Sr of the bulk so ils define a mixing line be tween the two dust sources, starting at ~0.7084, which is also the averag e composition of most of the speleothems
(Fig. 39A). Consequently, th is is probably an average 87Sr/86Sr of the Sinai proximal source. 87Sr/86Sr ratios of the Base Member spel eothems formed at 3.3-2.7 Ma are 0.70770.7079, similar to the host-rock va lues (Fig. 39B), whereas in the Intermediate (~ 1.3 Ma) and Young member the ratios are 0.7079-0.7085 (Fig. 39A-C). (deMenocal, 1995; 2004) suggested that the formation of Saharan-Arabian desert be lt started at ~ 2.7 Ma. Since Intermediate and Young members were deposited later, it is possi ble that there was higher contribution of dust from Saha ra-Sinai deserts du ring their deposition. 9. Summary and conclusions Speleothem deposition in present-day arid regions of Israel indicates that humid climatic conditions (i.e. periods with positiv e effective precipitation/infiltration index) occurred in the past. Effective precipitation te nds to decrease with increasing temperature and evaporation. The minimum precipitation amounts necessary to deposit speleothems are estimated in this stu dy to be: 200-275 mm/year duri ng glacial periods and 300-350 mm/year during interg lacial periods. In the Jordan Valley Maâ€™ale-Efrayim Cave, speleothem deposition mainly occurred during the three last glacial periods, with minor deposition during Termination II (~135 ka) and MIS-7 interglacial (225-205 ka), and no speleothem deposition during the Holocene, the peak of last glacial maximum (~19 ka) and the previous interglacial MIS5(1-4). This contrasts with the continuous speleothem growth that occurred in the Mediterranean climate zone on the western side of the Central Mountain Ridge (CMR) during the last 240 ka. Comp arison of the present-day 18O and D values of the cave and rain water on the western and the eastern side s of the CMR shows that at present, higher temperatures and greater evaporation on the easte rn flank and in the Jordan Valley are the major influences on the isotopic com positions and lack of rainfall. The 18O and 13C profiles of speleothems depos ited between 67 ka and 25 ka in Ma'ale-Efrayim Cave
match the general trends of the isotopic prof iles of Soreq Cave speleothems, suggesting the similar precipitation sour ce (EM Sea) and similar clim atic conditions. Thus, during the last glacial the desert boundary in the Jord an Valley effectively migrated further south from the present day location, whereas during the Holocene and the last interglacial the desert boundary was in its present-day pos ition. The decrease in temperature and consequent increase in effective precipitation in the Jordan Valley are probably the major factors controlling the decay of the "rain shadow" effect during glacial periods. Most of the speleothem deposition in th e Tzavoa Cave, northern Negev, occurred during the glacial periods (as in Jordan Valle y), whereas during the interglacial periods and glacial maxima speleothem deposition was limited and episodic. No speleothem deposition occurred at MIS-5(3-2), Younger Dr yas and Holocene, indicating that climate in these periods was similar to present or even more arid. Holocene speleothems were only found in Maâ€™ale-Dragot cav es located on the present-day 280-300 mm isohyets. These observations are consistent with the proposal that the precipitation that caused speleothem deposition during th e interglacials in Tzavoa Cave was 300-350 mm (i. e. twice than the present day), whereas duri ng the glacial periods , precipitation amounts were 200-275 mm, owing to cooler temp eratures and reduced evaporation. 13C values of the speleothems in the northern Negev during glacial periods (-3 -9) show that vegetation was usually C3+C4 steppe, whereas the 13C values in Jordan Valley speleothems (-8 -11) formed at the same time indicate a higher percentage of C3 Mediterranean vegetation, and wetter co nditions. Both in the Jordan Valley and northern Negev the 18O values of the speleothems generally show the same overall trends as the Soreq Cave in the Mediterranean climate, pointing to the EM Sea as the precipitation source. D18O relations of speleothem fluid inclusions also indicate the EM precipitation source in northern Negev and Jordan Valley. The 18O values of speleothems in the northern Negev are, however, systematically lower than in contemporaneous speleothems from the Jord an Valley and central and northern Israel.
This in part is attributed to the increased rainout of the heavy isotopes by the Rayleigh fractionation processes, possibly due to fu rther distance of these caves from the Mediterranean Sea. The study shows that the maximum southward shift of the border between the Mediterranean climate and the se mi-desert zone during the humid periods in the northern Negev was 20-25 km. Major humid periods in the central and southern Negev Desert (Negev Humid Periods â€“ NHP) during the last 350 ka occurred at: 350-310 ka (NHP-4), 310-290 ka (NHP-3), 220-190 ka (NHP-2), and 142-109 ka (NHP-1). NHP-4, NHP-2 and NHP-1 occurred during interglacial conditions (MIS-9, MIS-7(3-1) and Te rmination II + MIS-5(5-4) respectively), whereas NHP-3 is associated with glacial period (MIS-8). D18O relations in the paleo-cave waters trapped in fl uid inclusions indicate that the EM Sea was the major source of the rainfall during the last 220 ka in Jordan Valley, northern and central Negev. Another important evidence for major EM Sea rain source during the NHP-1, 2 and 3 is the decrease in speleothem amounts from north to the south. Intensive speleothem deposition occurred in the centr al Negev during the NHP-1, 2 and 3, whereas speleothem deposition was limited during NH P-4. In the southern Negev intensive speleothem deposition occurred only durin g the NHP-4, with limited speleothem deposition occurring during the NHP-1 and no speleothems were found from NHP-2 and 3. This comparison indicates that the major source of precipitati on during NHP-1, 2 and 3 was the EM Sea, whereas tropical rain source probably became a prominent source during NHP-4. However, the record of the central and southern Negev is in contrast with northern Negev, Jordan Valley and Judea Desert, where the most of the speleothems were deposited during the glacial periods. Glacial speleothem deposition in the latter occurred due to precipitation amounts of 200 mm or higher associated with high infiltration coefficients during the cold glacial periods, whereas central and southern Negev remained dry. During the humid interglacial episodes th e intensities of the Cyprus cyclones were higher than during the glacials, bringing pr ecipitation high as 300-350 mm as far south as
the present-day ~50 mm isohyet. However, th e higher temperatures and low effective precipitation caused only limited water infiltration to the caves leading only to deposition of thin speleothem laminae. The major decreases of speleothem 18O in Soreq Cave at 132 ka and 127-126 ka and in Peqiâ€™in Cave at 200 ka occurred simultaneous ly with the highest peaks of speleothem deposition in the Negev Desert during th e NHP-1 and NHP-2. These events were probably associated with most intensiv e humid episodes above the EM Sea. NHP-s were contemporaneous with periods of monsoon index of 51 (cal/cm2 day) and formation of the sapropels in the Mediterranean Sea. Such simultaneous intensification of the monsoon and Atlantic-Mediterra nean cyclones is probably related to the weakening of the high pre ssure cell above sub-tropical Atlantic Ocean. Absence of speleothem deposition in the Negev duri ng the Holocene and MIS-11 (430-400 ka) interglacials could be explained by lower Northern Hemisphere insolation and lower African Monsoon index than during the NHPs. Thus, the high pressure cell above subtropical Atlantic Ocean was probably stronger, leading to mo re northern trajectories of Atlantic-Mediterranean cyclones and preven ting from Mediterranean precipitation to reach central and southern Negev. Speleothem 87Sr/86Sr ratios in the speleothems in the â€œrain shadowâ€ and the Negev deserts usually vary between 0.7082 and 0.7085, and only slightly change from the glacial to interglacial conditions, compared to central Israel. This could be due to lower host rock dissolution rates, or higher dust supply. The dust supply to the Negev Desert was higher than to central Israel because it is located near the proximal SinaiNile Delta dust source. Speleothem and soil 87Sr/86Sr ratios probably indicate ~ 0.7084 as the 87Sr/86Sr ratio for this source. Most of the speleothems older than the uranium series dating limit deposited in the Negev belong to the Basal member and have preliminary U-Pb ages ranging between 3.3 Ma and 2.7 Ma, simultaneous with formation of the lakes of the Arava formation. A few
speleothems (Intermediate member) were deposited at ~1.3 Ma, probably simultaneously with the occurrence of the lakes of Zhiha Formation. A reduction in the rate of speleothem deposition, together with an increase in speleothem 87Sr/86Sr ratios between Basal and Intermediate-Young members, probably indicates increased desertification of this region and of entire Saharan-Arabian Desert after ~2.7 Ma, which caused increased dust supply from the Sinai and Saharan-Arabian sources. Multiple growth / hiatus cycles in a single speleothem over >3 Ma show that in arid environment the fractures conducting water from the surface to the caves remain open over long times. The Sinai-Negev land bridge was the major, and possibly the only land-bridge, connecting the Africa with Asia since the Miocene, and the Saharan-Arabian Desert was significant obstacle for hominid and animal migrations from Africa to other parts of the world. This study suggests that humid phases in the Sinai-Negev land bridge that occurred contemporaneously with humid phases in southern and central Saharan-Arabian Desert probably opened the climatic â€œwindowsâ€ for migration of hominids and animals out of the African continent. Migration of the early modern humans from Africa to the Levant was probably associated with NHP-1 at 142-109 ka.
Bibliography: Aggarwal, P. K., Alduchov, O., Araguas, L., Dogramaci, S., Katzlberger, G., Kriz, K., Kulkarni, K. M., Kurttas, T., Newman, B. D., and Pucher, A. (2007). New capabilities for studies using isotopes in the water cycle. EOS, Transections, American Geophysical Union 88, 537-538. Almogi-Labin, A., Bar-Matthews, M., and Ayalon, A. (2004). Climate variability in the Levant and northeast Africa during the Late Quaternary based on marine and land records. In: (Goren-Inbar, N. and Speth, D. Eds) Human paleoecology in the Levantine corridor , p. 117-134. Amit, R., Enzel, Y., and Sharon, D. (2006). Permanent Quaternary hyperaridity in the Negev, Israel, resulting from regional tectonics blocking Mediterranean frontal systems. Geology 34, 509-512. Amit, R., and Gerson, R. (1986). The evolution of Holocene Reg (gravelly) soils in the deserts; an example from Dead Sea region Catena 13, 59-79. Antonioli, F., Bard, E., Potter, E.-K., Silenzi, S., and Improta, S. (2004). 215-ka history of sea level oscillations from marine and continental layers in Argentarola Cave speleothems (Italy). Global and Planetary Change 43, 57-78. Auler, A. S., and Smart, P. L. (2004). Rates of condensation corrosion in speleothems. Speleogenesis and Evolution of Karst Aquifers www.speleogenesis.info 2, p.2 (4 pages) Avni, Y. (1991). "The geology, paleogeography and landscape evolution in Central Negev Highlands and the western Ramon structure." M.Sc. (in Hebrew, English Abstract), Hebrew University. Avni, Y. (1997). Geological evolution of the central and southern Negev as an indicator of the evolution of the Dead Sea Transform western margin in the late Neogene and Quaternary. Ph.D. Thesis (in Hebrew, English Abstract) , 240 pp. Avni, Y., Bartov, Y., Garfunkel, Z., and Ginat, H. (2001-a). Evolution of the Paran drainage basin and its relation to the Plio-Pleistocene history of the Arava Rift western margin, Israel. Israel Journal of Earth Sciences 49, 215-238. Avni, Y., Bartov, Y., Garfunkel, Z., and Ginat, H. (2001-b). The Arava formation A Pliocene sequence in the Arava Valley and its western margin, southern Israel. Israel Journal of Earth Sciences 50, 101-120. Avni, Y., Porat, N., Plakht, J., and Avni, G. (2006). Geomorphic changes leading to natural desertification versus anthropogenic land conservation in an arid environment, the Negev Highlands, Israel. Geomorphology 82, 177-200. Ayalon, A., Bar-Matthews, M., and Sass, E. (1998). Rainfall-recharge relationships within a karstic terrain in the Eastern Mediterranean semi-arid region, Israel: [delta] 18O and [delta]D characteristics. Journal of Hydrology 207, 18-31. Ayalon, A., Bar-Matthews, M., and Kaufman, A. (1999). Petrography, strontium, barium and uranium isotope ratios in speleothems as paleoclimatic proxies: Soreq Cave, Israel. The Holocene 9, 715-722. Ayalon, A., Bar-Matthews, M., and Kaufman, A. (2002). Climatic conditions during marine isotope stage 6 in the eastern Mediterranean region from the isotopic composition of the Soreq Cave, Israel. Geology 30, 303-306. Ayalon, A., Bar-Matthwews, M., Schilman, B. (2004). Rainfall isotopic characteristics at various cites in Israel and the relationships with unsuturated zone water. Israel Geological Survey Report GSI/16/04, 37pp.
Baker, A., Smart, P. L., and Ford, D. C. (1993). North-west European paleoclimate as indicated by growth frequency variations of secondary calcite deposits. Palaeogeography, Palaeoclimatology, Palaeoecology 100, 291-301. Baker, A., and Smart, P. L. (1995). Recent flowstone growth rates: field measurements in comparison with theoretical predictions. Chemical Geology 122, 121-128. Banner, J. L. (1995). Application of the trace element and isotope geochemistry of strontium to studies of carbonate diagenesis. Sedimentology 42, 805-824. Bar-Matthews, M., Matthews, A., and Ayalon, A. (1991). Environmental controls of speleothems mineralogy in a karstic dolomitic terrain (Soreq Cave, Israel). Journal of Geology 99, 189-207. Bar-Matthews, M., Ayalon, A., Matthews, A., Sass, E., and Halicz, L. (1996). Carbon and oxygen isotope study of the active water-carbonate system in a karstic Mediterranean cave: Implications for paleoclimate research in semiarid regions. Geochimica et Cosmochimica Acta 60, 337-347. Bar-Matthews, M., Ayalon, A., and Kaufman, A. (1997). Late Quaternary Paleoclimate in the Eastern Mediterranean Region from Stable Isotope Analysis of Speleothems at Soreq Cave, Israel. Quaternary Research 47, 155-168. Bar-Matthews, M., Ayalon, A., Kaufman, A., and Wasserburg, G. J. (1999). The Eastern Mediterranean paleoclimate as a reflection of regional events: Soreq cave, Israel. Earth and Planetary Science Letters 166, 85. Bar-Matthews, M., Ayalon, A., and Kaufman, A. (2000). Timing and hydrological conditions of Sapropel events in the Eastern Mediterranean, as evident from speleothems, Soreq cave, Israel. Chemical Geology 169, 145-156. Bar-Matthews, M., Ayalon, A., Matthews, A., and Frumkin, A. (2000-b). Eastern Mediterranean Paleoclimate during the last 250,000 years as derived from the petrography, mineralogy, trace element and isotopic composition of cave deposits (speleothems), Israel. Israel Geological Survey Report TR-GSI/10/2000, 17 pp. Bar-Matthews, M., and Ayalon, A. (2001). Eastern Mediterranean paleoclimate during the last 250,000 years as derived from the petrography, mineralogy, trace element and isotopic composition of cave deposits (speleothems), Israel, pp. 73. Report GSI/41/01, Geological Survey of Israel, The Ministry of National Infrastructures, Jerusalem, Israel. Bar-Matthews, M., Ayalon, A., Gilmour, M., Matthews, A., and Hawkesworth, C. J. (2003-a). Sea-land oxygen isotopic relationships from planktonic foraminifera and speleothems in the Eastern Mediterranean region and their implication for paleorainfall during interglacial intervals. Geochimica et Cosmochimica Acta 67, 3181-3199. Bar-Matthews, M., Almogi-Labin, A., Ayalon, A., Schilman, B., Kolosovsky, E., Aizenshtat, Z., McGarry, S., and Matthews, A. (2003-b). The relationship between the marine and terrestrial (speleothems) climatic record during the Quaternary in the Eastern Mediterranean region. Israel Geological Survey Report GSI/10/2003, 31 pp. Bartov, Y., Stein, M., Enzel, Y., Agnon, A., and Reches, Z. (2002). Lake Levels and Sequence Stratigraphy of Lake Lisan, the Late Pleistocene Precursor of the Dead Sea. Quaternary Research 57, 9-21. Bartov, Y., Goldstein, S. L., Stein, M., Enzel, Y. (2003). Catastrophic arid episodes in the Eastern Mediterranean linked with the North Atlantic Heinrich events. Geology 31, 439-442. Begin, Z. B., Ehrlich, A., and Nathan, Y. (1974). Lake Lisan, the Pleistocene precursor of the Dead Sea. Bulletin of the Geological Survey of Israel 63, 30 pp.
Berger, A. (1978). Long-term variations of caloric insolation resulting from the Earth 's orbital elements. Quaternary Research 9, 139-167. Bonacci, O. (2001). Monthly and annual effective infiltration coefficients in Dinaric karst: example of the Gradole karst spring catchment. Hydrological SciencesJournal-des Sciences Hydrologiques 46 . Bookman, R., Bartov, Y., Enzel, Y., and Stein, M. (2006). The levels of the Quaternary lakes in the Dead Sea basin: two centuries of research. In: Enzel, Y., Agnon, A. Stein, M. (Eds), New frontiers in Dead Sea paleoenvironmental research., Geological Society of America Special Paper 401, 155-170. Brook, G. A., Embabi, N. S., Ashour, M. M., Edwards, R. L., Cheng, H., Cowart, J. B., and Dabous, A. A. (2002). Djara Cave in the Western Desert of Egypt: morphology and evidence of Quaternary climatic change. Cave and Karst Science: The transactions of the British Cave Research Association 29, 57-66. Buchmann, D., and Dreybrodt, W. (1985). The kinetics of calcite dissolution and precipitation in geologically relevant situations of karst areas (2. Closed system). Chemical Geology 53, 109-124. Burns, S. J., Fleitmann, D., Matter, A., Kramers, J., and Al-Subbary, A. (2003). Indian ocean climate and an absolute chronology over Daansgard/Oeschger events 9 and 13. Science 301, 1365-1367. Calvo, R. (2002). "Stratigraphy and petrology of the Hazeva Formation in the Arava and the Negev: Implications for the developments of the sedimentary basins and morphotectonics of the Dead Sea Rift Valley." Ph.D. thesis (in Hebrew, English Abstract) Hebrew University. Causse, C., Ghaleb, B., Chkir, N., Zouari, K., Ben Ouezdou, H., and Mamou, A. (2003). Humidity changes in southern Tunisia during the Late Pleistocene inferred from U-Th dating of mollusc shells. Applied Geochemistry 18, 1691-1703. Cerling, T. E., Quade, J. (1993). Stable carbon and oxygen isotopes in soil carbonates. Climate Change in continental Isotopic Records, Geophysical Monograph 78, 217-231. Craig, H. (1961). Isotopic variations in Meteoric Waters. Science 133, 1702-1703. Crombie, M. K., Arvidson, R. E., Sturchio, N. C., El Alfy, Z., and Zeid, K. A. (1997). Age and isotopic constraints on Pleistocene pluvial episodes in the Western Desert, Egypt. Palaeogeography, Palaeoclimatology, Palaeoecology 130, 337355. Crouvi, O., Amit, R., Enzel, Y., Porat, N., and MCDonald, E. V. (2007). Regional pedosedimentary stratigraphic units in the primary eolian loess of the Negev Desert, Israel. In: Crovi, O., Malik, U., Oren, O., Shalev, E., Friedman, V., and Yechieli, Y., eds., Israel Geological Society, Annual Meeting: Neve Zohar, The Dead Sea. , p.23. Dan, Y., Yaalon, D., Koymdjisky, H., and Raz, Z. (1975). The Soils of Israel (in Hebrew, with 1:500,000 soil map of Israel). In "Bulletin 168 of The Volcanic Institute of Agricultural Research, ." pp. 28, Beit Dagan, Israel. Dansgaard, W. (1964). Stable isotopes in precipitation. Tellus 16, 438-468. Dayan, U. (1986). Climatology of Back Trajectories from Israel Based on Synoptic Analysis. Journal of climate and applied meteorology 25, 591-595. deMenocal, P. (1995). Plio-Pleistocene African Climate. Science 270, 53-59. deMenocal, P. B. (2004). African climate change and faunal evolution during the Pliocene-Pleistocene. E arth and Planeta r y Science Letters 220, 3.
Dennis, P. F., Rowe, P. J., and Atkinson, T. C. (2001). The recovery and isotopic measurement of water from fluid inclusions in speleothems. Geochimica et Cosmochimica Acta 65, 871-884. Derricourt, R. (2005). Getting "Out of Africa": sea crossings, land crossings. Journal of World Prehistory 19, 119-132. Drake, N., and Bristow, C. (2006). Shorelines in Sahara: geomorphological evidence for an enhanced monsoon from paleolake Megachad. The Holocene 16, 901-911. Dreybrodt, W., Lauckner, J., Liu, Z., Svensson, U., and Buchmann, D. (1996). The kinetics of the reaction CO2 + H2O >> H+ + HCO3 as one of the rate limiting steps of the dissolution of calcite in the system H2O CO2 CaCO3 Geochimica et Cosmochimica Acta 60, 3375-3381. Dreybrodt, W., Gabrovek, F., and Perne, M. (2005). Condensation corrosion: a theoretical approach. Acta Carsologica 34, 317-348. Ehrlich, S., Gavrieli, I., Ben-Dor, L., and Halicz, L. (2001). Direct high-precision measurements of the 87Sr/86Sr isotope ratio in natural water, carbonates and related materials by multiple collector inductively coupled plasma mass spectrometry (MC-ICP-MS). Journal of Analytic Atomic Spectrometry 16, 13891392. Emeis, K. C., Schulz, H. M., Struck, U., Sakamoto, T., Doose, H., Erlenkeuser, H., Howell, M., Kroon, D., and Paterne, M. (1998). Stable isotope and alkenone temperature records of sapropels from sites 964 and 967: constraining the physical environment of sapropel formation in the eastern Mediterranean Sea. Proceedings of the Ocean Drilling Program, Scientific Results 160, 309-331. Emeis, K.-C., Struck, U., Schulz, H.-M., Rosenberg, R., Bernasconi, S., Erlenkeuser, H., Sakamoto, T., and Martinez-Ruiz, F. (2000). Temperature and salinity variations of Mediterranean Sea surface waters over the last 16,000 years from records of planktonic stable oxygen isotopes and alkenone unsaturation ratios. Palaeogeography, Palaeoclimatology, Palaeoecology 158, 259. Emeis, K. C., Schulz, H. M., Struck, U., Rossignol-Strick, M., Sakamoto, T., Erlenkeuser, H., Howell, M., Kroon, D., Mackensen, A., Ishizuka, S., Oba, T., Sakamoto, T., and Koizumi, I. (2003). Eastern Mediterranean surface water temperatures and 18O composition during deposition of sapropels in the late Quaternary. Palaeogeography 18 (1), doi:10.1029/2000PA000617. Enmar, L. (1999). "The travertines in the northern and central Arava, petrography and geochemistry." M.Sc. (in hebrew, english abstract), Hebrew University. Enzel, Y., Amit, R., Dayan, U., Crouvi, O., Kahana, R., Ziv, B., and Sharon, D. (2007). The climatic and physiographic controls of the eastern Mediterranean over the late Pleistocene climates in the southern Levant and its neighboring deserts. Global and Planetary Change in press . Fairchild, I. J., Baker, A., Borsato, A., Frisia, S., Hinton, R., W., McDermott, F., and Tooth, A. (2001). Annual to sub-annual resolution of multiple trace-element trends in speleothems. Journal of the Geological Society 158, 831-841. Feinbrun-Dothan, N., Danin, A. (1998). "Analytical flora of Eretz-Israel." CANA Publishing House Ltd., Jerusalem, Israel. Fernandes, C. A., J., R. E., and Siddall, M. (2006). Absence of post-Miocene Red Sea land bridges: biogeographic implications. Journal of Biogeography 33, 961-966. Fleitmann, D., Burns, S. J., Neff, U., Mangini, A., and Matter, A. (2003). Changing moisture sources over the last 330,000 years in Northern Oman from fluidinclusion evidence in speleothems. Quaternary Research 60, 223-232.
Fontugne, M., and Calvert, S. E. (1992). Late Pleistocene variability of the carbon isotopic composition of organic matter in the eastern Mediterranean: monitor of changes in carbon sources and atmospheric CO2 concentrations. Palaeoceanograpy 7, 1-20. Forbes, J. (1998). Air temperature and relative humidity study: Torgac Cave, New Mexico. Cave and Karst Studies 60, 27-32. Ford, D., and Williams, P. W. (1989). "Karst Geomorphology and Hydrology." Unwin Hyman, London. French, H., and Binley, A. (2004). Snowmelt infiltration: monitoring temporal and spatial variability using time-lapse electrical resistivity. Journal of Hydrology 297, 174186. Frisia, S., Borsato, A., Fairchild, I. J., and McDermott, F. (2000). Calcite fabric, growth mechanisms, and environments of formation in speleothems from Italian Alps and southwestern Ireland. Journal of Sedimentary Research 70, 1183-1196. Frisia, S., Borsato, A., Fairchild, I. J., McDermott, F., and Selmo, E. M. (2002). Aragonite-calcite relations in speleothems (Grotte de Clamouse, France): Environment, fabrics and carbonate geochemistry. Journal of Sedimentary Research 72, 687-699. Frumkin, A., Ford, D. C., and Schwarcz, H. P. (1999). Continental Oxygen Isotopic Record of the Last 170,000 Years in Jerusalem. Quaternary Research 51, 317327. Frumkin, A., Ford, D., and Schwarcz, H. (2000). Paleoclimate and vegetation of the last glacial cycles in Jerusalem from a speleothem record. Global Biochemical Cycles 14, 863-870. Frumkin, A., and Stein, M. (2004). The Sahara-East Mediterranean dust and climate connection revealed by strontium and uranium isotopes in a Jerusalem speleothem. Earth and Planetary Science Letters 217, 451-464. Frumkin, A., and Fischhendler, I. (2005). Morphometry and distribution of isolated caves as a guide for phreatic and confined paleohydrological conditions. Geomorphology 67, 457-471. Gallup, C. D., Cheng, H., Taylor, F. W., and Edwards, L. R. (2002). Direct determination of the timing of sea level change during Termination II. Science 295, 310-313. Ganor, E., and Foner, H. A. (1996). The mineralogical and chemical properties and the behavior of the eolian Saharan dust over Israel. In: Guerzoni, S. and Chester, R. (Eds.) The impact of desert dust across the Mediterranean pp. 163-172. Ganor, E., and Foner, H. A. (2001). Mineral dust concentrations, deposition fluxes and deposition velocities in dust episodes over Israel. Journal of Geophysical Research 106, 18431-18437. Garfunkel, Z. (1970). The tectonics of the western margin of the southern Arava. Ph.D. Thesis (in Hebrew, English Abstract) , 204 pp. Garfunkel, Z. (1978). The Negev regional synthesis of sedimentary basin. International Association of Sedimentology (IAS). Tenth International Congress of Sedimentology. Part 1: Precongress, Israel. , pp. 35-110. Garfunkel, Z. (1982). The structures of the Syrian Arc in Israel. Israel Geological Society Annual Meeting , Elat, p. 23. Garfunkel, Z. (1988-a). The history of the Dead Sea Basin. Israel Geological Society Annual Meeting , En-Boqeq, p. 32. Garfunkel, Z. (1988-b). Relations between continental rifting and uplifting: evidence from the Suez rift and northern Red Sea. Tectonophysics 150 , 33-49.
Garfunkel, Z., and Horowitz, A. (1966). The Upper Tertiary and Quaternary morphology of the Negev. Israel Journal of Earth Sciences 15, 101-117. Gascoyne, M. (1992). Paleoclimate determination from the cave calcite deposits. Quaternary Science Reviews 11, 609-632. Gat, J. R. (1984). The stable isotope composition of the Dead Sea waters. Earth and Planetary Science Letters 71, 361-376. Gat, J. R., and Carmi, I. (1987). Effect of climate changes on the precipitation patterns and isotopic composition of water in a climatic transition zone:case of the Eastern Mediterranean Sea area. In "The influence of Climatic Change and Climatic Variability on the Hydrologic Regime and Water Resources." pp. 513-523. IANS 168 Symp. Proc. Genty, D., Plagnes, V., Causse, C., Cattani, O., Stievenard, M., Falourd, S., Blamart, D., Ouahdi, R., and Van-Exter, S. (2002). Fossil water in large stalagmite voids as a tool for paleoprecipitation stable isotope composition reconstruction and paleotemperature calculation. Chemical Geology 184, 83-95. Ginat, H. (1997). The geology and landscape evolution in Hayyon and Zihor valleys. Ph.D. Thesis (in Hebrew, English Abstract) , 115 pp. Ginat, H., Zilberman, E., and Saragusti, I. (2003). Early Pleistocene lake deposits and Lower Paleolithic finds in Nahal (wadi) Zihor, Southern Negev desert, Israel. Quaternary Research 59, 445-458. Goede, A., McCulloch, M., McDermott, F., and Hawkesworth, C. (1998). Aeolian contribution to strontium and strontium isotope variations in a Tasmanian speleothem. Chemical Geology 149, 37-50. Goodfriend, G. A. (1990). Rainfall in the Negev Desert during the middle Holocene, based on (super 13) C of organic matter in land snail shells. Quaternary Research 34, 186-197. Goodfriend, G. A. (1991). Holocene trends in 18O in land snail shells from the Negev Desert and their implications for changes in rainfall source areas. Quaternary Research 35, 417-426. Goodfriend, G. A. (1999). Terrestrial stable isotope records of Late Quaternary paleoclimates in the eastern Mediterranean region. Quaternary Science Reviews 18, 501-513. Goodfriend, G. A., and Magaritz, M. (1988). Palaeosols and late Pleistocene rainfall fluctuations in the Negev Desert. Nature 332, 144-146. Greenbaum, N., Porat, N., Rhodes, E., and Enzel, Y. (2006). Large floods during late Oxygen Isotope Stage 3, southern Negev desert, Israel. Quaternary Science Reviews 25, 704-719. Harmon, R. S., Ford, D. C., and Schwarcz, H. P. (1977). Interglacial chronology of the Rocky and Mackenzie Mountains based on 230Th/234U dating of calcite speleothems. Canadian Journal of Earth Sciences 14, 1543-1552. Haviv, I., Stone, J., Enzel, Y., Zilberman, E., Whipple, K., Matmon, A., and Fifield, L. K. (2006). Climatic control on erosion rates of dolo-limestone hilltops. In: (add editors) Israel Geological Society Annual Meeting, Bet-Shean, add page. Henderson, G. M., and Slowey, N. C. (2000). Evidence from U-Th dating agains Northern Hemisphere forsing of the penulmate deglaciation. Nature , 61-66. Hendy, C. H. (1971). The isotopic geochemistry of speleothems-I. The calculation of the effects of different modes of formation on the isotopic composition of speleothems and their applicability as palaeoclimatic indicators. Geochimica et Cosmochimica Acta 35, 801-824.
Hendy, C. H., and Wilson, A. T. (1968). Palaeoclimatic data from speleothems. Nature 219, 48-51. Herut, B., Starinsky, A., and Katz, A. (1993). Strontium in rainwater from Israel: Sources, isotopes and chemistry. Earth and Planetary Science Letters 120, 77-84. Holmgren, K., Karlen, W., and Shaw, P. A. (1995). Paleoclimatic Significance of the Stable Isotopic Composition and Petrology of a Late Pleistocene Stalagmite from Botswana. Quaternary Research 43, 320-328. Hurrel, J. W. (1995). Decadal trends in the north atlantic oscillation: regional temperatures and precipitation. Science 269, 676-679. Kahana, R., Ziv, B., Enzel, Y., and Dayan, U. (2002). Synoptic climatology of major floods in the Negev Desert, Israel. International Journal of Climatology 22, 867882. Kallel, N., Paterne, M., Duplessy, J. C., Vergnaud-Grazzini, C., Pujol, C., Labeyrie, L., Arnold, M., Fontugne, M., and Pierre, C. (1997). Enhanced rainfall on Mediterranean region during the last sapropel event. Oceanologica Acta 20, 697712. Kallel, N., Duplessy, J. C., Labeyrie, L., Fontugne, M., Paterne, M., and Montacer, M. (2000). Mediterranean pluvial periods and sapropel formation over the last 200 000 years. Palaeogeography, Palaeoclimatology, Palaeoecology 157, 45-58. Kaufman, A., Wasserburg, G. J., Porcelli, D., Bar-Matthews, M., Ayalon, A., and Halicz, L. (1998). U-Th isotope systematics from the Soreq cave, Israel and climatic correlations. Earth and Planetary Science Letters 156, 141-155. Koepnick, R. B., Burke, W. H., Denison, R. E., Hetherington, E. A., Nelson, H. F., Otto, J. B., and Waite, L. E. (1985). Construction of the seawater 87Sr/86Sr curve for the Cenozoic and Cretaceous supporting data. Chemical Geology (Isotope Geoscience Section) 58, 55-81. Kolodny, Y., Stein, M., and Machlus, M. (2005). Sea-rain-lake relation in the Last Glacial East Mediterranean revealed by O C in Lake Lisan aragonites. Geochimica et Cosmochimica Acta 69, 4045-4060. Kolosovsky, A. (2003). "Reconstruction of paleotemperatures and determination of of origin of organic matter in sediments from the eastern Mediterranean Sea (MSc Thesis, in Hebrew, English abstract)." Hebrew University, Jerusalem. Krichak, S. O., and Alpert, P. (2005). Signatures of the NAO in the atmospheric circulation during wet winter months over the Mediterranean region. Theoretical and Applied Climatology 82, 27-39. Krom, M. D., Cliff, R. A., Eijsink, L. M., Herut, B., and Chester, R. (1999). The characterisation of Saharan dusts and Nile particulate matter in surface sediments from the Levantine basin using Sr isotopes. Marine Geology 155, 319-330. Kronfeld, J., and Livnat, A. (1987). Uranium-series dating of Rift Valley travertines and lacustrine deposits: paleoclimatic implications and stratigraphic application. Israel Ministry of Energy and Infrastructure. Krumgalz, B. S., Hecht, A., Starinsky, A., and Katz, A. (2000). Termodynamic constraints of the Dead Sea evaporation: Can the Dead Sea dry up? Chemical Geology 165, 1-11. Kuperman, G. (2005). "Reconstruction of flood periods by dating speleothems in Nahal Hazera, northern Negev." Hebrew University of Jerusalem, M.Sc. Thesis (in Hebrew, English Abstract). Lauritzen, S. E. (1995). High resolution paleotemperature proxy record for the last interglaciation based on Norwegian speleothems. Quaternary Research 43, 133146.
Li, H.-C., Ku, T.-L., You, C.-F., Cheng, H., Edwards, R. L., Ma, Z.-B., Tsai, W.-s., and Li, M.-D. (2005). 87Sr/86Sr and Sr/Ca in speleothems for paleoclimate reconstruction in Central China between 70 and 280 kyr ago. Geochimica et Cosmochimica Acta 69, 3933-3947. Livnat, A., and Kronfeld, J. (1985). Paleoclimatic implications of U-series dates for lake sediments and travertines in the Arava Rift Valley, Israel. Quaternary Research 24, 164-172. Livnat, A., and Kronfeld, J. (1990). Pleistocene lakes in the Arava Rift of Israel (Sayif Formation): facies and paleoenvironmental setting. Journal of African Earth Sciences 10, 409. Ludwig, K. R. (2003). Isoplot 3.00, Special Publication No. 4. In "A Geochronological Toolkit for Microsoft Excel." (K. R. Ludwig, Ed.). Berkeley Geochronology Center, 2455 Ridge Road, Berkeley CA 94709, USA. Magaritz, M., and Goodfriend, G. A. (1987). "Abrupt Climatic Change." Reidel D., Publishing Company. Martinson, D. G., Pisias, N., Hays, J. D., Imbrie, J., Moore, T. C. J., and Schackleton, N. J. (1987). Age dating and the orbital theory of the ice ages: development of the high-resolution 0-300,000 year chronostratigraphy. Quaternary Research 27, 129. Martrat, B., Grimalt, J. O., Lopez-Martinez, C., Cacho, I., Sierro, F. J., Flores, J. A., Zahn, R., Canals, M., Curtis, J. H., and A., H. D. (2004). Abrupt temperature changes in the Westerm Mediterranean over the past 250,000 years. Science 306, 1762-1765. Matthews, A., Ayalon, A., and Bar-Matthews, M. (2000). D/H ratios of fluid inclusions of Soreq cave (Israel) speleothems as a guide to the Eastern Mediterranean Meteoric Line relationships in the last 120 ky. Chemical Geology 166, 183-191. McDermott, F. (2004). Palaeo-climate reconstruction from stable isotope variations in speleothems: a review. Quaternary Science Reviews 23, 901-918. McDermott, F., Frisia, S., Huang, Y., Longinelli, A., Spiro, B., Heaton, T. H. E., Hawkesworth, C., Borsato, A., Keppens, E., Fairchild, I. J., Borg, K., Verheyden, S., and Selmo, E. M. (1999). Holocene climate variability in Europe: Evidence from 18O, textural and extension rate variations in three speleothems. Quaternary Science Reviews 18, 1021-1038. McDermott, F., Mattey, D. P; Hawkesworth, C. (2001). Centennial-scale Holocene climate variability revealed by a high-resolution speleothem delta (super 18) O record from SW Ireland. Science 294, 1328-1331. McGarry, S., Bar-Matthews, M., Matthews, A., Vaks, A., Schilman, B., and Ayalon, A. (2004). Constraints on hydrological and paleotemperature variations in the Eastern Mediterranean region in the last 140 ka given by the [delta]D values of speleothem fluid inclusions. Quaternary Science Reviews 23, 919-934. Miller, G. H., Wendorf, F., Ernst, R., Schild, R., Close, A. E., Friedman, I., and Schwarcz, H. P. (1991). Dating lacustrine episodes in the eastern Sahara by the epimerization of isoleucine in ostrich eggshells. Palaeogeography, Palaeoclimatology, Palaeoecology 84, 175-189. Neev, D., and Emery, K. O. (1967). The Dead Sea. Bulletin of the Geological Survey of Israel 41, 147pp. Njitchoua, R., Sigha-Nkamdjou, L., Dever, L., Marlin, C., Sighomnou, D., and Nia, P. (1999). Variations of the stable isotopic compositions of rainfall events from the Cameroon rain forest, Central Africa. Journal of Hydrology 223, 17.
O'Neil, J. R., Clayton, R. N., and Mayeda, T. K. (1969). Oxygen isotope fractionation of divalent metal carbonates. Journal of Chemical Physics 30, 5547-5558. Olausson, E. (1961). Studies of deep sea cores. Sediment cores from the Mediterranean Sea and the Red Sea. Rep. Sweden Deep Sea Expedition 1947-48 VIII, 337-391. Osmond, J. K., and Dabous, A. A. (2004). Timing and intensity of groundwater movement during Egyptian Sahara pluvial periods by U-series analysis of secondary U in ores and carbonates. Quaternary Research 61, 85-94. Petit-Maire, N., Sanlaville, P., Abed, A., Yasin, S., Bourrouilh, R., Carbonel, P., Fontugne, M., and Reyss, J. L. (2002). New data for an Eemian lacustrine phase in southern Jordan. Episodes 25, 279-280. Picard, L. (1943). Structure and evolution of Palestine (with comparative notes on neighbouring countries). Hebrew University., Jerusalem, Department Geological Bulletin 4, 187 pp. Plakht, J. (1995). Climatic conditions during the last stages of development of Makhtesh Ramon. Israel Journal of Earth Sciences 44, 149-157. Plakht, J. (2000). Quaternary units in makhteshim, Negev Desert: Similarities and pecularities. Israel Journal of Earth Sciences 49, 179-187. Polyak, V., and Asmerom, Y. (2005). Orbital control of long term moisture in southwestern USA. Geophysical Reseach Letters 32, 1-4. Poulson, T. L., and White, W. B. (1969). The cave environment. Science 165, 971-981. Rabikovitz, S. (1992). The Soils of Israel: Formation, Nature, and Properties (In Hebrew, with 1:250,000 soil map of Israel in English and Hebrew). , pp. 489. Hakibbuts Hamehuchad Publishing House, Israel. . Rink, W. J., Richter, D., Schwarcz, H. P., Marks, A. E., Monigal, K., and Kaufman, D. (2003). Age of the Middle Palaeolithic site of Rosh Mor, central Negev, Israel: Implications for the age range of the Early Levantine Mousterian of the Levantine corridor. Journal of Archaeological Science 30, 195-204. Robinson, L. F., Henderson, G. M., and Slowey, N. C. (2002). U-Th dating of marine isotope stage 7 in Bahamas slope sediments. Earth and Planetary Science Letters 196, 175-187. Roche, D. M., Dokken, T. M., Goosse, H., Renssen, H., and Weber, S. L. (2007). Climate of the Last Glacial Maximum: sensitivity studies and model-data comparison with the LOVECLIM coupled model. Climate of the past 3, 205-224. Rohling, E. J., Cane, T. R., Cooke, S., Sprovieri, M., Bouloubassi, I., Emeis, K. C., Schiebel, R., Kroon, D., Jorissen, F. J., Lorre, A., and Kemp, A. E. S. (2002). African monsoon variability during the previous interglacial maximum. Earth and Planetary Science Letters 202, 61-75. Rossignol-Strick, M. (1983). African monsoons, an immediate climate response to orbital insolation. Nature 304, 46-49. Rossignol-Strick, M., and Paterne, M. (1999). A synthetic pollen record of the eastern Mediterranean sapropels of the last 1 Ma: implications for the time-scale and formation of sapropels. Marine Geology 153, 221-237. Ryan, W. B. F. (1972). Stratigraphy of the late Quaternary sediments in the Eastern Mediterranean. In "The Mediterra n ean Sea: A Natural Sedimentation Laboratory " (D. J. Stanley, Ed.), pp. 149-169. Dowden, Hutchinson and Ross, Stroudsberg, Pennsylvania, US. Said, R. (1993). "The river Nile: Geology, Hydrology and Utilization." Pergamon Press Ltd, Oxford.
Schwarcz, H. P. (1986). Geochronology and isotopic geochemistry of speleothems. In "Handbook of Environmental Isotope Geochemistry." (P. Fritz, and J. C. Fontes, Eds.), pp. 271-300. Elsevier, Amsterdam. Schwarcz, H. P., Blackwell, B., Goldberg, P., and Marks, A. E. (1979). Uranium series dating of travertine from archaeological sites, Nahal Zin, Israel. Nature 277, 558560. Shackleton, N. J., and Opdyke, N. D. (1973). Oxygen isotopes and paleomagnetic stratigraphy of equatorial Pacific core V28-238: Oxygen isotope temperatures and ice volumes on a 105 year and 106 year scale. Quaternary Research 3, 39-55. Shay-El, Y., and Alpert, P. (1991). A diagnostic study of winter diabatic heating in the Mediterranean in relation with cyclones Quarterly Journal of Royal Meteorological Society 117, 715-747. Shuman, B., and Donnelly, J. P. (2006). The influence of seasonal precipitation and temperature regimes on lake levels in the northeastern United States during the Holocene. Quaternary Research 65, 44-56. Solomon, A., and Garfunkel, Z. (1989). The evolution of the northern Negev monoclines. Israel Geological Society Annual Meeting , Ramot, p.131. Soudry, D., Ehrlich, S., Yoffe, O., and Nathan, Y. (2002). Uranium oxidation state and related variations in geochemistry of phosphorites from the Negev (southern Israel). Chemical Geology 189, 213-230. Soudry, D., Nathan, Y., and Roded, R. (1985). The Ashosh-Haroz facies and their significance for the Mishash paleogeography phosphorite accumulation in the northern and central Negev, southern Israel Israel Journal of Earth Sciences 34, 211-220. Spotl, C., Mangini, A., Frank, N., Eichstadter, R., and Burns, S. J. (2002). Start of the last interglacial period at 135 ka: evidence from a high Alpine speleothem. Geology 30, 815-818. Steinitz, G., and Yosef, B. (1991). The Miocene-Pleistocene history of the Dead Sea segment of the Rift in light of K-Ar ages of basalts. Israel Journal of Earth Sciences 40, 199-208. Street, A. F., and Grove, A. T. (1979). Global Maps of Lake-Level Fluctuations since 30,000 yr B. P. Quaternary Research 12, 83-118. Szabo, B. J., Haynes, J. C. V., and Maxwell, T. A. (1995). Ages of Quaternary pluvial episodes determined by uranium-series and radiocarbon dating of lacustrine deposits of Eastern Sahara. Palaeogeography, Palaeoclimatology, Palaeoecology 113, 227-242. Tsoar, H., and Pye, K. (1987). Dust transport and the question of desert loess formation. Sedimentology 34, 139-153. Usami, M., Suga, S. I., Lototskaya, A., Ziveri, P., Ganssen, G. M., and van Hinte, J. E. (1998). Calcareous nannofloral response to Termination II at 45oN, 25oW (northeast Atlantic). Marine Micropaleontology 34, 47-70. Vaks , A. (2001). "Climatic changes during the last 160 kyr from the speleothem record of karstic cave located on desert boundary, Ma'ale-Efrayim, Eastern Shomeron, Israel." M.Sc. Thesis (in Hebrew, English Abstract), Hebrew University. Vaks, A., Bar-Matthews, M., Ayalon, A., Schilman, B., Gilmour, M., Hawkesworth, C. J., Frumkin, A., Kaufman, A., and Matthews, A. (2003). Paleoclimate reconstruction based on the timing of speleothem growth and oxygen and carbon isotope composition in a cave located in the rain shadow in Israel. Quaternary Research 59, 182-193.
Vaks, A., Bar-Matthews, M., Ayalon, A., Matthews, A., Frumkin, A., Dayan, U., Halicz, L., Almogi-Labin, A., and Schilman, B. (2006). Paleoclimate and location of the border between Mediterranean climate region and the Saharo-Arabian Desert as revealed by speleothems from the northern Negev Desert, Israel Earth and Planetary Science Letters 249, 384-399. Vaks, A., Bar-Matthews, M., Ayalon, A., Matthews, A., Halicz, L., and Frumkin, A. (2007). Desert speleothems reveal climatic window for African exodus of early modern humans. Geology 35, 831-834. Vogel, J. C., Fuls, A., Danin, A. (1986). Geographical and environmental distribution of C3 and C4 grasses in Sinai, Negev, and Judean deserts. Oecologia 70, 258-265. Waldman, N., Starinsky, A., and Stein, M. (2007). Primary carbonates and Ca-cloride brines as monitors of a paleo-hydrological regime in the Dead Sea basin. Quaternary Science Reviews in-press . Walker, J., Cliff, R. A., and Latham, A., G. (2007). U-Pb isotopic age of the StW 573 Hominid from Sterkfontein, South Africa Science 314, 1592-1594. Woodhead, J., Hellstrom, J., Maas, R., Drysdale, R., Zanchetta, G., Devine, P., and Taylor, E. (2006). U-Pb geochronology of speleothems by MC-ICPMS. Quaternary Geochronology 1, 208-221. Yaalon, D. H. (1971). Soil forming processes in time and space. In: Yaalon, D. H. (Ed.) Paleopedology: Origin, nature and dating of paleosols. International Society of soil Science and Israel Universities Press , pp. 29-39. Yaalon, D. H., and Dan, J. (1974). Accumulation and distribution of loess-derived deposits in the semi-desert fringe areas of Israel. Zeitschrift fur Geomorphologie Supplementband 20, 91-105. Yaalon, D. H., and Ganor, E. (1973). The influence of dust on soils during the Quaternary. Soil Science 116, 146-155. Yair, A. (1990). The role of topography and surface cover upon soil formation along hillslope in arid climate. Geomorphology 3, 287-299. Yonge, C. J. (1982). Stable isotope studies from the water extracted from speleothems. Ph.D. Thesis, McMaster University , 298 pp. Zangvil, A., and Druian, P. (1990). Upper Air Trough Axis Orientation and the Spatial Distribution of Rainfall over Israel. International Journal of Climatology IJCLEU 10, 57-62. Zilberman, E. (1991). "Landscape evolution in the central, northern, and northwestern Negev during the Neogene and the Quaternary (Ph.D Thesis, in Hebrew, English abstract)." Hebrew University, Jerusalem.
Appendix 1: Figures
(A) Map indicating the extent of the SaharanArabian Desert (shaded gray area). The rectangle marks the research area in this study; (B) Rainfall map of Israel and adjacent lands: Palestinian Authority in Gaza (PA), north-eastern Egypt western Jordan, south-western Syria and southern Lebanon. Isohyets are indicated by black lines. Peqiâ€™in Cave is shown in northern Israel by a gray circle labeled PQ. The dotted rectangle indicates the "rain shadow" desert and Negev Desert research area; (C) Map showing the relief (Hall, 1997), precipitation and location of the caves in the research area. The isohyets are marked by white lines and caves by filled black circles numbered as following: â€œ Rain shadowâ€desert: 1) Ma'ale-Efrayim Cave; 2) Kanaim Cave; Northern Negev: 3) Ma'ale-Dragot Caves; 4) Tzavoa Cave; Central and Southern Negev: 5) IzzimCave, 6) Hol-Zach Cave, 7) Makhtesh-ha-QatanCave, 8) Ashalim Cave, 9) EvenSid-Ramon mini-caves, 10) Mitzpe-Ramon Cave, 11) WadiLotzCave, 12) Ma'ale-ha-Meyshar Cave, 13) ShizafonQuarry mini-caves and 14 ) Ktora Cracks (collapsed cave). Other caves where speleothem studies wher e previously made performed are shown by gray circles and nam ed as following: SQ â€“Soreq Cave, J â€“Jerusalem Cave. Major population centers are s hown by white circles. The rain sampling stations at Ma'ale-Ef rayim, Beâ€™erSheva, Arad, Makhtech-Ha-Qatan(near cave 7), Mitzpe-Ramon (near caves 9 and 10) and NeotSmadar(near the caves 13 and 14), are i ndicated by white cylinders.Beâ€™er-Sheva-AradValley shown as BA Valley, fragmented lines show major wadis (W.). Figure 1: Geographical location of the study area, annual precipitation and sample sites. A Northern Negev Central NegevB A V a l l e y2 6 5 41005 02 0 03 0 04 0 05 0 0 C60 0 Beâ€™erSheva Jerusalem Tel-AvivGazaAradS o u t h e r n Ne g e vElat AqabaDe a d S e aEa st e rn M e d i t err a n ea n S e a5 0 32 N 31 N 30 N BJordan J o r d a n V a l l e y 30 km 800 8 0 0 1 0 0 0 6 0 0 6 0 0 40 0 200 Ea s t e r n M ed i t e r r a n e a n S e a Syria LebanonPQ 4 0 0 400 6 0 0 6 0 0 600 50 km JordanIsrael Egypt31 N 32 N 30 N 33 N 2 0 04 0 0 2 0 0 Sinai Peninsula PAPA Negev Desert8 9 10 11 12 13 1SQ14 7 A r a v a V a l l e yC M RC M RR ai n s h ad ow d e s e r t J Gulf of Elat (Aqaba) W . P a r a nW . Zi n 3J u d e a D e s e r tE l a t M o u n t a i n s
Loess / Alluvial terraces ZhihaFormation Arava Formation Dead Sea Group HazevaGroup AvdatGroup Mount Scopus GroupPleistocene Miocene Pliocene Eocene Paleocene Maastricht SenonianJudea GroupTuronian CenomanianCurnubGroup Arad GroupAlbian Aptian Callovian Quaternary Tertiary Cretaceous Jurassic Figure 2:Geology of the research area. Most caves (#1-10, and #12-14, Fig. 1C) are located in Cenomanian and Turonian limestone and dolomite of Judea Group, but one cave (Wadi-LotzCave, # 11, Fig. 1C) is located in Eocene limestone of AvdatGroup. 2A: Geological stratigraphic section of the research area.
Figure 2B: Geologic map of the research area. Caves marked by numbers and black circles as in Fig. 1C. 2 6 5 4 8 9 10 11 12 1SQ13 7J3 14 31 N 30 N 32 N Late Quaternary Pliocene Pliocene Miocene Early-Middle Oligocene Miocene EoceneVolcanic rocks PalaeoceneTertiary QuaternarySenonian Coniacian CretaceousAlbianCenomanian Turonian Mesozoic Lower Cretaceous Precambrian crystalline basement Cambrian Jurassic Triassic Ordovician Legend:
Figure 3: Typical cross sections of the Ashalim Cave speleothemsa s observed in the field, hand specimens and under SEM : 1 â€“Flowstone ASH-15 in Ashalim Cave (#8, Fig. 1): the Basal massive member is yellow (top centre â€“ laminae D-I) and composed of large (> 2 cm) calcite cr ystals. The laminae C, D and E (marked by rectangles), are dated by U-Pbmethod to 1.269 0.013 Ma, 3.045 0.021 Ma and 3.048 0.014 Ma respectively. The white and brownish thin lamina between the two stratigraphic members (enlarged and marked by arrow) represents a growth break (hiatus). (2) Enlargements of boundarybetween the Basal member (lamina D) and Intermediate member (laminae A-C): 2.1 â€“in plain light, 2.2 â€“under SEM. (3, 4) SEM analysis shows that it contains silica, alumina, microcrystalline calcite, Fe-oxides and halite. 5 â€“Stalactite ASH-11 in Ashalim Cave: the Basal member is yellow (top centre), theIntermediate member is brown, and the Young member composed of 4 thin laminae is shown enlarged in (6). Numerous fine white laminae (white arrows in top) representing hiatuses divide between the wider calcite laminae. 6 â€“Enlargement of the Young member in stalactite ASH-11 showing the 5 calcite laminae with overall thickness of less than 2 cm. Four white thin laminae dividing between the calcite laminae represent hiatuses (marked by black arrows in top). The U-Th ages of the laminae are marked by arrows in bottom. 5 >550ka 203ka 117ka 132 ka 10 cm 1 6 hiatuses Host rock DEA B CFGHI 2 3 4 D C 2.2 2.1
EHost rock C DF GHI IK J A1.1 137 ka A1.2 319 ka >550 ka B AhiatusesB -Gypsum lamina hiatuses B -The cross section of flowstone KTO(1)-1 from the Ktora Cracks (#14 in Fig. 1). Black arrows on the right mark the hiatuses in calcite deposition between the Basal and the Intermediate members, and inside the Intermediate member. The oldest major hiatus is lamina H, composed of two gypsum laminae (beige) and thin calcite lamina in between (dark). The two gypsum laminae represent periods of hyper-arid conditions that occurred between the deposition of the Basal and the Intermediate members. The laminated Intermediate member (laminae E, F and G) is of lesser width, and shows frequent A -The Basal member (laminae I, J, and K) makes up the most of the speleothem volume and contains only 2 hiatuses at the bottom. This is indicative of generally continuous growth from slowly dripping water. The oldest laminae J and K that are marked by rectangles were dated by U-Pbmethod to 2.71 0.06 Ma and 3.32 0.06 Ma, respectively. alternations between calcite laminae and hiatuses (arrows) and therefore represents generally drier conditions than the Basal member. The Young member is thinly laminated (laminae A1, A2, B, C and D at the top, <1 cm width). The laminae A1 and A2 are dateable by the U-Th method and their ages are 137 ka and 319 ka respectively, whereas laminae C and D are older than 550 ka. The gypsum lamina B in the Young member formed during the long hyper-arid period during Middle-Pleistocene. Figure 4: Cross section of the flowstone KT O(1)-1 from Ktora Cracks (#14, Fig. 1C)
Figure 5:The mineralogy and petrography of the calcite and fine laminae representing hiatuses within the Young member of the stalactite ASH-11. A -Abrasion of the calcite crystals on the present day surface of the stalactite; B â€“etched calcite crystals of the ancient surface (hiatus plane) between laminae A2 (203 ka) and A3 (>550 ka); C â€“petrographic microscope picture (crossed polar light) showing the gypsum on the stalactite surface, calcite crystals of the lamina A1, detrital matter in hiatus between the laminae A1 and A2, and columnar fabric of calcite crystals in lamina A2; D â€“SEM pictur e of the present day surface of the stalactite with gypsum plates over etched calcite crystals; E â€“SEM picture of a halite sphere and the flakes of the clay minerals within the fine lamina representing the hiatus between the lami nae A1 and A2; F -SEM picture of the gypsum plates over calcite crystals formed during the growth break between thelaminae A2 and A3. C D E 1 mm gypsum detritusA 1 A 2 F B A
Figure 6: Flowstone MMR-7(2) from Maâ€™ale-ha-Meyshar Cave. A â€“Photograph of the present-day surface of the flowstone covered by a â€œpatinaâ€of gr ey-whitish microcrystalline calcite and detrital materials. The two youngest calcite laminae were dated at 204-198 ka and 127-120 ka. These laminae are separated one from other by fine laminae rich with detrital materials, microcrystalline calcite and evaporite minerals, representing hiatuses. It is easy to separate between the calcite laminae along these hiatus planes exposing the older speleothem surfaces. Surface older than 204 ka (hiatus plane) A hiatuses Present day surface120-127 ka 198-204 ka B â€“SEM analysis of the material accumulated on present-day speleothem surface shows microcrystalline calcite, clay minerals (silica, alumina), gypsum, halite, quartz grains, phosphate and minerals of Fe, Ti, and Zn. C â€“SEM analysis of the material accumulated on speleothem surface older than 204 ka (reddish colored hiatus plane in A) shows relatively similar mineralogy as in B (microcrystalline calcite, clay minerals (silica, alumina), quartz grains, and minerals of Fe, Mg and Zn), with larger amounts of Fe, Cr and Zn, smaller amounts of gypsum and halite, and an absence of Ti and phosphate. B C
Figure 7: SEM analysis of the minerals. (A) Celestine (SrSO4), (B) Minerals of Ti, Fe, Mn, Mg, with clay minerals from present day surface of flowstone MMR-8; (C) Barite (BaSO4) and celestine from the hiatus between laminae E and F in flowstone ASH-33. Ti Ti A B C
Figure 8: Petrography of stalactite KN-8 from the Kanaim Cave showing laminae of beige columnar calcite alternating with white laminae of microcrystalline calcite, probably formed by condensation corrosion and representing hiatuses in speleothem deposition. The corrosion of calcite probably induced radionuclide remobilization causing the age reversal between laminae A and B. AB 147 ka 183 ka A B
A B C Host rockFigure 9:Phreatic speleothems in Maâ€™ale-ha-Meyshar Cave . A : Phreatic calcite overgrowth layer (thin arrows) covering the cave walls and ceiling. The reddish paleosols(thick arrow) postdate the phreatic speleothems and possibly were washed into the cave after it was uplifted above the groundwater level. C: the cave â€œraftsâ€covering the floor of a small cave near Maâ€™aleha-Meyshar major cave. B: Laminated calcite overgrowth on the cave walls
Figure 10:The relative frequencies of the speleothem ages in â€œrain shadowâ€desert . A -Kanaim Cave, Judea Desert; B â€“Maâ€™ale-Efrayim Cave, including the modified data of Vaks et al., (2003) and the new 7 ages. Figure 11 (next page): Dating of central and southern Negev speleothems. A â€“Age distribution and periods of speleothem depos ition of central and southern Negev Desert as determined by U-Th dating. The horizontal axis marks the age (ka). The ages are arranged vertically according to their speleothem number (bold font) and in dividual sample number (smaller font). Black circles with error bars define individual age measurements. Ve rtical rectangles mark speleothem depositional periods: green rectangles indicate periods of intensive speleothem deposition (two or more contemporaneous ages), and yellow rectangles indicate periods of limited sp eleothem deposition (one age only). The grey rectangles on the right define the periods between 350 ka to 600 ka in which the age uncertainties are large, and precise timing of the speleothem deposition periods is impossible. Dark grey rectangles show periods of higher relative frequency of the speleothem deposition, light grey rectangles show periods when the frequency is low. B â€“The relative age frequencies indicating the fraction of speleothems deposited at a certain age with 95% confidence. C â€“Timing of humid periods in southern Saharan-Arabian Desert (rectangles mark the humid periods) compared with periods of speleothem deposition in the Negev: Ref. 1 â€“(Fleitmannet al., 2003); Ref. 2 â€“ (Osmond and Dabous, 2004) (dark blue â€“most humid periods, light blue â€“less humid periods, white â€“dry periods); Ref. 3 â€“(Crombieet al., 1997); Ref. 4 â€“(Szaboet al., 1995)).Relative frequency Relative frequencyA B 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 050100150200250300350400 Age (ka) 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 050100150200250300350400Age (ka)
Age (Ka) 100 200 300 500 600 400 KTO(1)-1A1 A2Speleothem samples ASHACH-4SHACH-3A BSHACH-1A BMMR-8A BMMR-7(2)A1 A2 BMMR-7A1.1 A1.2 A3.1 A3.2ESID-7(3)A1 BESID-7(1)A1 A2 A3 BESID-2A1.1 A1.2 A2 B A1 A2 B1 B2ASH-34C1 C2 C30 BRelative frequency of ages 0.01 0.02ASH-33A1 A2 B1 B2 B3 B4 C1 C2 D1 D2 E1.1 E1.2 E2.1 E2.2 A1.1 A1.2 A2.3 A2.2 A2.1 A2.4ASH-11 MKTC-5MKTC-4 A A C H1HZ-1 IZ-1 HZ-2 HZ-3(2) HZ-3(5)A1 A2 B A2 A3 A B(I) B(II) A B1 B2 C1 C2 D E A1(I) A1(II) B A2.2 A2.1 A3 A5MMR-13A B1 B2 C E F G H I 0.03 0.04 Saharan Arabian humid periods C Ref.1 Ref.2 Ref.3 Ref.4 H2 H9 Figure 11:
A B CFigure 12:Plots of the relative age frequencies versus age for 150-100 ka (A); 230-180 ka (B) and 550-250 ka (C). 0 0.01 0.02 0.03 0.04 0.05 100 110 120 130 140 150 Age (ka) 0 0.005 0.01 0.015 0.02 0.025 180 190 200 210 220 230 Age (ka) 0 0.001 0.002 0.003 0.004 0.005 250300350400450500550 Age (ka)Relative age frequency Relative age frequency Relative age frequency
Figure 13: Sample ASH-33 from the Ashalim Cave. 1 â€“2 pieces of flowstone ASH-33 with a drill holes for U-Th dating in the left piece. The dotted lines correlate between the top of the lamina C and the base of lamina E in the two pieces; 2 â€“the detailed stratigraphy of the sample. The hiatus E-F appears as orange clay lamina, two rock fragments (RF) overlay the hiatus. The areas photographed under microscope (3, 4, and 5) are marked by rectangles in 2; 3 -columnar fabric typical for speleothems grown in a closed system; 4 -the lamina E2.2 grown in a open system when the secondary calcite has grown in the pores between the rock fragments and the clay lamina (plane polar light); 5 -the same in cross polar light. BC1C 2C 4C 3D 1D 2E 1 . 1 E 1 . 2 E 2 . 2 E 2 . 1FGH hiatus (clay layer) hiatus 1 2 3 4 5RF RFRFRFR FR Fclay layer pores pores 1 cm 1 mm RF
050010001500200025003000Age (ka) 0 -2 -4 -6 -8 -10 -1218O ( PDB) Figure 14: 18O values of central and southern Negev speleothems during the last~3Ma. Samples younger than 350 ka are enclosed by the large rectangle. The small rectangles enclose samples that grew at 516 51 ka, 1.269 0.013 Ma (preliminary U-Pbage) and betwee n 3.02 Ma and 3.07 Ma (preliminary U-Pbages) . Samples older than 350 ka are marked by green circles. The samples in the light blue oval fiel ds are between 550 ka to 1.3 Ma, and older than 3.07 Ma, they were not dated and are arranged according to their stratigraphicorder. Young member Intermediate member Basal member Last 350 ka (Enlarged in Figure 15) 516 51 ka 1.269 0.013 Ma 3.023.07 Ma
86 87 88 89Age (ka) -7.5 -8 -8.5O ( PDB) 280285290295300305Age (ka) -1 -2 -3 -4 -5O ( PDB) y Figure 15: Oxygen isotope profiles of central and southern Negevspeleothems during the last 350 ka . (A) speleothems deposited during the last 350 ka; (B) between 89 and 86 ka; (C) between 161 and 154 ka; (D) between 305 ka and 285 ka; (E) between 350 ka and 31 5ka. U-Th ages with s error bars for samples deposited during short episodes are shown in top of plots B-E. 80120160200240280320Age (ka) -2 -4 -6 -8 -10 -1218O ( PDB) Legend:ESID-2 SHACH-1 ASH-33 ASH-34 MMR-8 HZ-3(2) ESID-7(1) HZ-1 MMR-7(2) ASH-11 MKTC-5 A B D E C 310320330340350Age (ka) -10 -10.5 -11 -11.5 -12O ( PDB) 154156158160162Age (ka) -2 -3 -4 -5 -6O ( PDB)
115120125130135140Age (ka) -7 -8 -9 -10 -1118O ( PDB) A B C D 115120125130135Age (ka) -7 -8 -9 -10 -1118O ( PDB) Legend:ESID-2 SHACH-1 ASH-33 ASH-34 MMR-8 HZ-3(2) ESID-7(1) HZ-1 MMR-7(2) ASH-11 MKTC-5 195200205210215220225Age (ka) -6 -7 -8 -9 -10 -11O ( PDB) 190195200205210215220225Age (ka) -6 -7 -8 -9 -10 -11 -1 2 18O ( PDB) Figure 16: Oxygen isotope profiles in central Negev during the two major periods of speleothem deposition. (A) Oxygen isotope profiles of central and southern Negev speleothems sampled between 133 ka and 115 ka. U-Th ages with 2 errors are given in the top of the diagrams; (B) data as in A with running average of 18O between 133 ka and 126 ka (marked by black line); (C) Oxygen isotopic profiles between 221 ka and 197 ka. UTh ages with 2 errors are given in the tops of the diagram; (D) data as in C with running average of 18O (marked by black line).
050010001500200025003000Age (ka) 2 0 -2 -4 -6 -813C ( PDB) Young member Intermediate member Basal member Last 350 ka (Enlarged in Figure 18) 516 51 kaFigure 17: 13C values of central and southern Negev speleothems during the last~3 Ma . Samples younger than 350 ka are enclosed by a large rectangle. The small rectangles enclose samples that grew at 51651 ka, 1.269 0.013 Ma (preliminary U-Pbage) and between 3.02 Ma and 3.07 Ma (preliminary U-Pbages). Samples older than 350 ka are marked by green circles. The samples in the light blue fields are between 550 ka to 1.3 Ma, and older than 3.07 Ma, were not dated and are arranged according to their stratigraphicorder. 1.269 0.013 Ma 3.023.07 Ma
86 87 88 89Age (ka) -4 -4.5 -5 -5.5 -613C ( PDB) 153154155156157158159160161162Age (ka) -6 -7 -8 -913C ( PDB) 280285290295300305Age (ka) -3 -4 -5 -6 -7 -8 9 C ( PDB) 310320330340350Age (ka) 1 0 -1 -2 -3C ( PDB) Figure 18: Carbon isotope profiles of central and southern Negevspeleothems during the last 350 ka. (A) all speleothems formed during the last 350 ka; (B) between 89 and 86 ka; (C) between 161 and 154 ka; (D) between 305 ka and 285 ka; (E) between 350 ka and 315 ka. U-Th ages with s error bars for samples deposited during short episodes are shown in the top of plots B-E. 80120160200240280320Age (ka) 2 0 -2 -4 -6 -8 -1013C ( PDB) LEGEND:SHACH-1 ASH-33 ASH-34 MMR-8 ESID-2 ASH-11 ESID-7 HZ-1 HZ-3(2) MMR-7(2) MKTC-5 B CD E A
Figure 19:Carbon isotope profiles in central Negev duri ng the two major periods of speleothem deposition (A) Carbon isotope profiles of central and southern Negev speleothems between 133 ka and 115 ka. U-Th ages with 2 errors are given at the top of th e diagram; (B) Data as in A, with running average of 13C marked by the black line; (C) Carbon isotopic profiles of central and southern Negev speleothems between 221 ka and 197 ka. U-Th ages with 2 errors are given on the top of the diagram; (D) Data as in C, with running average of 13C marked by the black line . A B C D 115120125130135Age (ka) 0 -2 -4 -6 8 13C ( PDB) 190195200205210215220225Age (ka) 2 1 0 -1 -2 -3 -4 -5 -6 -7 -8 -9 -1013C ( PDB) 115120125130135140Age (ka) 0 -1 -2 -3 -4 -5 -6 -7 -8 -9 -1013C ( PDB) 195200205210215220225Age (ka) 2 0 -2 -4 -6 -813C ( PDB)
Figure 20: Northern Negev rainfall isotopic composition. A -Oxygen and hydrogen isotopic composition of northern Negev rainfall during the 2004-2006 rainy seasons (September-May) including results of Vaks et al. (2006). (B) enlarged central part of A. Events that have undergone most significant evaporation marked by ov al in A and form the Beâ€™er-Sheva evaporation line (black dashed line in A and B). -8-6-4-20218O ( SMOW) -40 -30 -20 -10 0 10D ( SMOW) -12-10-8-6-4-2024618O ( SMOW) -70 -60 -50 -40 -30 -20 -10 0 10 20 30 4 0 D ( SMOW) A B Legend:Global Meteoric Water Line (MWL) Mediterranean Meteoric Water Liine (MMWL) Be'er-Sheva rain events <10 mm Arad rain events Be'er-Sheva rain events >10 mm tropical moisture Be'er-Sheva rain events >10 mm EM moisture Be'er-Sheva, 2004-2006 average Arad, 2004-2006 average Be'er-Sheva evaporation line
Figure 21: Central and southern Nege v rainfall isotopic composition. A -Oxygen and hydrogen isotopic composition of the rainfall in central and southern Negev during 2004-2006 (September to May) in Mitz pe-Ramon and Neot-Smadar, and during 2002-2006 in Makhtesh-ha-Qat an (including the data of Kuperman(2005)); B -enlarged central part of A; C â€“the average isotopic compositions of the five rainfal l collecting stations in the Negev plotted relative to the evaporation line of Beâ€™er-Sheva rainfall shown in Fig. 20 (A, B). -8-6-4-2024618O ( SMOW) -40 -30 -20 -10 0 10 20 30 40D ( SMOW) -8-6-4-2018O ( SMOW) -30 -20 -10 0 10D ( SMOW) -5-4-3-218O ( SMOW) -20 -15 -10 -5D ( SMOW) Neot Smadar rain events Makhtech-ha-Qatan 2004-2006 average Arad, 2004-2006 average Be'er Sheva, 2004-2006 average Mitzpe-Ramon 2004-2006 average Neot-Smadar 2004-2006 average Legend:MWL MMWL Makhtesh-ha-Qatan rain events Mitzpe-Ramon rain events Be'er-Sheva evaporation line A B C
050010001500200025003000Age (ka) -60 -50 -40 -30 -20D FI ( SMOW) Figure 22: D values of Negev Desert speleothem fluid inclusions (FI) during the last 220 ka with the trend line in grey (A); during the last 3.1 Ma, samples with pr eliminary U-Pbages are enclosed in two small rectangles (B). FI of Tzavoa, Ashalim, Hol-Zakh and Maâ€™ale-ha-Meyshar Caves are shown by blue, green, purple and orange triangles, respectively. The samples in the light blue oval filed (B) are between 550 ka to 1.3 Ma, they were not dated and are arranged according to their stratigraphicorder. A BYoung member Intermediate member Basal member 050100150200Age (ka) -60 -50 -40 -30 -20D FI ( SMOW) 1.269 0.013 Ma 3.02-3.07Ma Legend:Tzavoa Cave Hol-Zakh Cave Ma'ale-ha-Meyshar Cave Ashalim Cave
0 1500 3000Age (ka) 0 100 200 300 400 500 600Sr (ppm) Figure 23:Srconcentrations in the Negev and Jordan Valley speleothems ; during the last 3.3 Ma, samples with preliminary U-Pbages enclosed in rectangles (A); during the last 300 ka in Jordan Valley (B); during the last 200 ka in the northern Negev: Tzavoa cave speleothems are shown by olive green data points linked by lines, and Maâ€™ale-Dragot by green data points (C); during the last 350 ka in the central and southern Negev (D). The samples in the light blue oval fields are between 550 ka to 1.3 Ma, and older than 3.07 Ma, they were not dated and are arranged according to their stratigraphic order (A). 050100150200Age (ka) 0 100 200 300 400 500 600Sr (ppm) Legend:Ma'ale-Dragot Ma'ale-Efrayim Tzavoa Hol-Zakh Ashalim Even-Sid-Ramon Ma'ale-ha-Meyshar Shizafon Ktora Cracks 050100150200250300350Age (ka) 100 150 200 250 300 350Sr (ppm) B C ABasal member Intermediate member Young member and other young speleothems D 100150200250300350Age (ka) 0 100 200 300 400 500 600Sr (ppm) 1.269 0.013 Ma 3.02-3.07Ma 2.71.06Ma 3.32.06Ma
050100150200250300350Age (ka) 0.707 0.708 0.709 0.71 0.711 0.71287Sr / 86Sr Figure 24: 87Sr/86Sr isotopic ratios of Maâ€™ale-Efrayim Cave speleothems ; during the last 300 ka (A); 87Sr/86Sr isotopic ratios of Maâ€™ale-Efrayim Cave speleothems compared with 87Sr/86Sr isotopic composition of dolomitichost rock, bulk soil above the cave and silicate fraction of the soil (B). B ATerra-Rosa soil â€“ silicate fraction Terra-Rosa soil â€“ bulk Dolomite host rock 050100150200250300350Age (ka) 0.7081 0.7082 0.7083 0.7084 0.7085 0.708 6 87Sr / 86Sr
05010015020025030035 0 Age (ky BP) 0.7079 0.708 0.7081 0.7082 0.7083 0.7084 0.7085 0.708687Sr/86Sr 0 1500 3000Age (ka) 0.7076 0.7078 0.708 0.7082 0.7084 0.708 6 87Sr/86Sr Legend:Ma'ale-Dragot Cave Tzavoa Cave Hol-Zakh Cave Ashalim Cave Even-ve-Sid mini-caves Ma'ale-ha-Meyshar Cave Shizafon mini-caves Ktora Cracks A BFigure 25:87Sr/86Sr ratios of the Negev Desert vadose speleothems ; formed during the last 3.3 Ma, preliminary U-Pbages shown by small rectangles (A); and during the last 350 ka (B). The samples in the light blue oval fields (A) are between 550 ka to 1.3 Ma, and older than 3.07 Ma, they were not dated and are arranged according to their stratigraphic order. Young member Intermediate member Basal member 1.269 0.013 Ma 2.71.06Ma 3.02-3.07Ma 3.32.06Ma Age (ka)
050100150200250300350Age (ky BP) 0.708 0.709 0.71 0.71187Sr / 86Sr 0 1500 3000Age (ka) 0.708 0.709 0.71 0.71187Sr/86Sr Young member Intermediate member Basal memberFigure 26: 87Sr/86Sr ratios in Negev Desert vad ose speleothems compared with 87Sr/86Sr ratios of dolomitic host rock, of bulk soil above the caves and of its silicate fraction ; during the last 3.3 Ma (A); 87Sr/86Sr ratios in Negev Desert vadose speleothems during the last 350 ka comp ared with the same parameters as in A. Preliminary U-Pbages are enclosed in small rectangles. The samples in the lightblue oval fields are between 550 ka to 1.3 Ma, and older than 3.07 Ma, they were not dated and are arranged according to their stratigraphicorder. 350 B A Legend:Ma'ale-Dragot Cave Tzavoa Cave Hol-Zakh Cave Ashalim Cave Even-ve-Sid mini-caves Ma'ale-ha-Meyshar Cave Shizafon mini-caves Ktora Cracks Negev host rocks Bulk Negev soils Silicate fractions of Negev soils 1.269 0.013 Ma 2.71.06Ma 3.02-3.07Ma 3.32.06Ma Silicate fractions of soils Bulk soils Silicate fractions of soils Bulk soils Host rocks Host rocks
Figure 27:Hydrogen and oxygen isotopic composition of paleo-cave waters based of speleothem fluid inclusion analysis : (A) Last 220 ka in all caves; (B) 2 last glacial periods (Tzavoa Cave); (C) NHP-1 in central and southern Negev; (D) NHP-2 in central and southern Negev. The MMWL (dashed red line) and MWL (blue line) are shown for reference. For each sample the range in the calculated 18O represents 2 C variations in cave temperatures (see text for explanation). -13-12-11-10-9-8-7-6-5-418O ( SMOW) -70 -60 -50 -40 -30 -20 -10D (SMOW) -13-12-11-10-9-8-7-6-5-418O ( SMOW) -70 -60 -50 -40 -30 -20 -10D ( SMOW) A B C D -13-12-11-10-9-8-7-6-5-418O ( SMOW) -70 -60 -50 -40 -30 -20 -10D (SMOW) -13-12-11-10-9-8-7-6-5-418O ( SMOW) -70 -60 -50 -40 -30 -20 -10D ( SMOW) Ashalim 117-124 ka Ashalim 117-126 ka Ashalim 126-130 k a Ashalim 127-132 ka Ashalim 131-132 ka Ashalim 126-130 ka Ashalim 131-132 ka Ashalim 199-202 ka Ashalim 201-203 ka MWL MMWL Tzavoa 32-35 ka Tzavoa 61-64 ka Tzavoa 65-67 ka Tzavoa 124-126 ka Tzavoa 128-131 ka Tzavoa 157-161 ka Tzavoa 172-175 ka Tzavoa 175-177 ka Tzavoa 195-198 ka Tzavoa 198-201 ka Hol-Zakh 126-131 ka Hol-Zakh 131-132 ka Ashalim 202-208 ka Ashalim 210-213 ka Ashalim 218-220 ka Ma'ale-ha-Meyshar 120-128 ka Ma'ale-ha-Meyshar 195-206 ka Ashalim 127-132 ka Legend:
04080120160200Age (ka) 0 10 20 30 40 50 60 MWL MMWLHighest d-excess in the present day rainfallFigure 28: d-excess values from speleothem fluid inclusions in central and southern Negev Desert during the last 220 ka plotted relative to the MWL (blue horizontal line), MMWL (red horizontal line) and maximu m d-excess in present-day rainfall from EM origin (red dotted horizontal line) (Bar-Matthews et al., 2003; Kuperman, 2005). Tzavoa Cave samples ar e marked by blue, Ashalim Cave by green, Hol-Zakh by red and Maâ€™ale-haMeyshar by orange two-end arrows. For each sample the range d-excess represents 2oC variations in cave temperatures (see text for explanation).d-excess (SMOW )
A B C D F E 124 ka 128 ka 137 ka 200 ka 197 ka 117 ka 130 ka 200 ka 218 ka 218 ka 200 ka 130 ka 197 ka 205 ka >550 ka 124 ka 131 ka >550 ka 137 ka319 ka >550 ka GFigure 29: Thinning of the speleothem sequences from the north to the south. Figure 29 (A-G):Width of the NHP-1 and NHP-2 sequences in the Negev speleothems . In Tzavoa Cave, northern Negev: width of the NHP-1 and NHP-2 sequences is ~ 6 cm (A, B). Ashalim Cave central Negev (C, D), and Maâ€™ale-ha-Meyshar Cave, the boundary between central and sou thern Negev (E, F): width of NHP-1 and NHP-2 sequences is up to 2.5 cm. Ktora Cracks, southern Negev: the width of NHP-1 sequence is reduced to a few mil limeters, while the NHP-2 sequence disappears (G). Stalactite TZ-4 Stalactite TZ-22(1) Flowstone ASH-33 Flowstone KTO(1) Flowstone MMR-8 Stalagmite MMR-13 Flowstone ASH-33, lower part
31.53130.53029.5Latitude ( o N) 0 1 2 3 4Speleothem laminae thic k ness (cm) 31.53130.53029.5Latitude ( o N) 0 1 2 3 4Speleothem laminae thic k ness (cm) Legend:Stalactite laminae thickness (cm) Flowstone laminae thickness (cm) Stalagmite laminae thickness (cm) Legend:Stalactite laminae thickness (cm) Flowstone laminae thickness (cm) Stalagmite thickness (cm) NHP-2 Figure 29 (H, I): Thinning of the speleothem laminae from th e north to the south in the Negev Desert during the NHP-1 (H) and during the NHP-2 (I).The maximal thickness of the laminae decrease from the nor th to the south.Laminae with thickness of 0 cm represent depositional hiatuses.HICaves CavesTZ TZ HZ HZ ASH ASH ESID ESID MMR MMR KTO KTO
0 50 100 300150 250 200 300 350 150Isohyet (mm) N S100 50 30 Age (ka)?? ? ? ? ? NHP-1 NHP-2 NHP-3 500 ?NHP-4? ? 1 0 05 02003 0 04 0 050 06 0 0 Beer-Sheva Tel-Aviv GazaS o u t h e r n N e g e vElat AqabaDe a d S e aEa s t e r n M e d i t e r r a n e a n Se a5 0 Arad Jerusalem Fig. 29J: Shortening of the speleothem deposition periods from the north to the south. Speleothem deposition during the glacial periods Speleothem deposition during the interglacial periods (incomplete data shown by striped pattern with question marks) Speleothem deposition in timeSpeleothem deposition in space 1 J 2 SQ 3 4 7 65 8 9 10 11 13 14 12 Southern boundary of continuous speleothem deposition Southern limit of intensive glacial speleothem deposition Approximate southern limit of the speleothem deposition during the NHP-2 ? ?
0 0.01 0.02 0.03 110115120125130135140 0 0.005 0.01 0.015 0.02 190195200205210215220225 py 115120125130135140 -3 -4 -5 -6 -7 -8 -9 -10 -1118O ( PDB) 80 120 160 200 -2 -4 -6 -8 -1018O ( PDB) BFigure 30: Comparison of the 18O profiles and age frequenciesof central Negev speleothems with 18O values of Tzavoa, Soreq and Peqiâ€™in speleothems. (A) 18O profiles of the central Negev speleothems (color) compared with those of Tzavoa Cave, northern Negev (grey) and Soreq and Peqiâ€™in Caves in central and northern Israel (black). The rectangles in (A) i ndicate NHP-1 (left) and NHP-2 (right). (B-C) 18O profiles of central Negev speleothems (top) compared with 18O profiles of the Tzavoa, Soreq and Peqiâ€™in caves (middle), and with relative frequencies of speleothem deposition in central and southern Negev (bottom) during NHP-1 (B) and NHP-2 (C). The green rectangles show the highest relative frequencies of speleothem ages. Relative frequency of speleothem agesC Age (ka) ARelative frequency of speleothem agesAge (ka)Legend: Legend:ASH-33 ASH-34 MMR-8 HZ-3(2) ESID-2 Age (ka)Soreq speleothems Tzavoa speleothems ASH-33 ASH-34 ASH-11 ESID-2 MMR-8 ESID-7 HZ-1 HZ-3(2) MMR-7(2) Legend: Soreq speleothems Tzavoa speleothems g ASH-33 ASH-34 ASH-11 ESID-7 HZ-1 HZ-3(2) MMR-7(2) 115120125130135 -7 -8 -9 -10 -1118O ( PDB) MKTC-5 195200205210215220225 -7 -8 -9 -10 -1 1 18O ( PDB) Peqi'in speleothems 190195200205210215220225 -3 -4 -5 -6 -7 -8 -9 -10 -1118O ( PDB) Peqi'in speleothems Tzavoa speleothems
0 0.005 0.01 0.015 0.02 190195200205210215220225 py 0 0.01 0.02 0.03 110115120125130135140 80 120 160 200 2 0 -2 -4 -6 -8 -10 -12 -1413C ( PDB) A B Relative frequency of speleothem agesCRelative frequency of speleothem agesAge (ka) Age (ka)Figure 31: Comparison of the 13C profiles and age frequenciesof central Negev speleothems with 13C values of Tzavoa, Soreq and Peqiâ€™in speleothems. (A) 13C profiles of central and southern Negev caves (color) compared with those of Tzavoa Cave, northern Negev (grey) and Soreq and Peqiâ€™in Caves in central and northern Israel (black). The rectangles in (A) show NHP-1 (left) and NHP-2 (right). (B-C) 13C profiles of central and southern Negev speleothems (top) compared with 13C profiles of the Tzavoa, Soreq and Peqiâ€™in caves (middle), and with relative frequencies of speleothem deposition in central and southern Negev (bottom) during NHP-1 (B) and NHP-2 (C). The green rectangles show the highest relative frequencies of speleothem ages.Age (ka) 115120125130135 0 -2 -4 -6 -813C ( PDB) 115120125130135140 0 -2 -4 -6 -8 -10 -1213C ( PDB) 190195200205210215220225 0 -2 -4 -6 -8 -10 -1213C ( PDB) Legend: MKTC-5 Soreq speleothems Peqi'in speleothems Tzavoa speleothems MMR-8 ASH-11 ESID-7 HZ-1 HZ-3(2) MMR-7(2) ASH-33 ASH-34 ESID-2 Legend:MKTC-5 Soreq speleothems Peqi'in speleothems Tzavoa speleothems MMR-8 ASH-11 ESID-7 HZ-1 HZ-3(2) MMR-7(2) ASH-33 ASH-34 ESID-2 195200205210215220225 2 1 0 -1 -2 -3 -4 -5 -6 7 13C ( PDB)
Figure 32: Correlation between the peaks of relative freque ncies of thespeleothem agesduring the last 550 ka (B) and solar radiation energy in July at 65 N (Berger, 1978) (A), and in January at 65 S (Berger, 1978) (C). NHP-s are indicated by green rectangles and peri ods of lower relative frequency of speleothem ages are marked by the yellow rectangles. Periods of higher relative frequency of speleothem ages between 550 ka and 350 ka are indicated by grey rectangles. Relative frequency of speleothem ages Solar energy at 65 N(W/m2) Solar energy at 65oS in January (W/m2)Age (ka) NHP-1 NHP-2 NHP-3 NHP-4 0.03 0.02 0.01 0 A B C0 100 200 300 400 500 0 100 200 300 400 500 0 100 200 300 400 500 400 440 480 520 0 100 200 300 400 500 380 400 420 440 460 480 50 0
0100200300400500Age (ka) -40 -20 0 20 40 60Monsoon Index ((cal*cm -2 )*day -1 ) Figure 33: Relative frequencies of the speleothem age s compared with African monsoon index and timing of Mediterranean sapropels (Rossignol-Strickand Paterne, 1999): NHP-s (green rectangles), periods of limited speleothem deposition (yellow rectangles) during the last 350 ka, and periods of higher relative frequency of speleothem ages between 550 ka and 350 ka (grey rectangles). African monsoon index (M) in time (t) defined by Rossignol-Strick(1983) as: Mt = IT t+ (IT t-IE t), where ITand IEis insolation in units (cal*cm2)*day-1at Tropic of Cancer and Equator, respectively. S1S4S3S5S6S7S8S9Relative frequency of speleothem agesAge (ka)0.03 0.02 0.01 0 S10S11 S12SA510100200300400500
180190200210220230 -3 -4 -5 -6 -7 -8 -918O ( PDB) 100110120130140150 -3 -4 -5 -6 -7 -8 -918O ( PDB) 0 0.005 0.01 0.015 0.02 0.025 -40 -20 0 20 40 60 380 400 420 440 460 480 500 Solar radiation at 65 N in July (W/m2) Relative frequency of speleothem ages Solar radiation at 65 S in January (W/m2) 380 400 420 440 460 480 500 Figure 34: Correlation between the relative frequencies of the speleothem ages during the NHP-1 (A) and NHP-2 (B) and the: 18O profiles of Soreq and Peqiâ€™in Caves (central and northern Israel), monsoon index, and parameters of solar radiation in July at 65 N, in December at 30 N and 60 S (Berger, 1978). Peaks of positive monsoon index are marked by red lines; green lines show peaks of negative monsoon index. Timing of the sapropels S5, S7 and S8 is shown by the black bars (Fontugneand Calvert, 1992) and grey bars (Emeiset al., 2003) at the top. ABSolar radiation at 65 N in July (W/m2) Relative frequency of speleothem ages Solar radiation at 65 S in January (W/m2)Age (ka) Age (ka) Monsoon Index (cal*cm-2/day) -40 -20 0 20 40 60 Monsoon Index (cal*cm-2/day) Oxygen profile of Peqiâ€™in Cave speleothems Oxygen profile of Soreq Cave speleothems 100110120130140150 100110120130140150180190200210220230 180190200210220230 180190200210220230 180190200210220230 100110120130140150 100110120130140150S5 S7 S8 51 51 0 0.01 0.02 0.03 0.04 0.05 380 400 420 440 460 480 500 380 420 460 500
0 0.001 0.002 0.003 0.004 0.005 Solar radiation at 65 N in July Relative frequency of speleothem ages Solar radiation at 65 S in JanuaryFigure 35: Correlation between the relative frequencies of the speleothem ages and sapropel events (black vertical bars at the top) during the period between 550 ka and 250 ka (A) with: African monsoon index values (B), solar radiation energy in July at 65 N (C), and in December at 60 S (D) (Berger, 1978). NHP-3 and 4 are indicated by green rectangles, periods of less intensive speleothem deposition by yellow rectangles; light yellow rectangles at the right mark higher frequencies of speleothem deposition during the period between 550 ka and 350 ka, where the age uncertainties are large. NHP-3NHP-4A B C D 400 420 440 460 480 500Monsoon Index ((cal*cm-2)*day-1)250300350400450500550 250300350400450500550 250300350400450500550 250300350400450500550 -40 -20 0 20 40 60 51 S10 S11S12 SA 380 400 420 440 460 480 500A g e ( ka )
SQ J 1 3 4 6 8 9 12 13 14 7J o rd an Va lleyC e n t r a l I sr a e lNegev DesertLake L i s a nSinai Peninsula Nile DeltaFigure 36: Location of the caves where Srisotopic composition of speleothems was studied (black filled circles with labels given in Appe ndix 2 and Fig.1C) Soreq (SQ) and Jerusalem (J) caves, central Israel are indicated with open circle s. Glacial (G) and interglacial/present-day (I) locations of the EM coastline are indicated in th e figure, together with the glacial Lake Lisan, Sinai Peninsula and Nile Delta. The area exposed during the glacial sea low stands shown in light gray; Wind direction during the passage of the Cyprus cyclones is marked by arrows.EM SeaG I
A BFigure 37:87Sr/86Sr ratios of Maâ€™ale-Efrayim Cave speleothems, Jordan Valley (A) and Negev Desert speleothems (B) compared with those of the speleothems ofSoreq (Ayalon et al., 1999) and Jerusalem (Frumkin and Stein, 2004) caves. 050100150200250300350Age (ka) 0.7077 0.7078 0.7079 0.708 0.7081 0.7082 0.7083 0.7084 0.7085 0.7086 0.708787Sr/86Sr Legend:Soreq Cave Jerusalem Cave Ma'ale-Efrayim Cave 050100150200250300350Age (ky BP) 0.7077 0.7078 0.7079 0.708 0.7081 0.7082 0.7083 0.7084 0.7085 0.7086 0.708 7 87Sr / 86Sr Legend:Ma'ale-Dragot Cave Tzavoa Cave Hol-Zakh Cave Ashalim Cave Even-ve-Sid mini-caves Ma'ale-ha-Meyshar Cave Shizafon mini-caves Ktora Cracks Soreq Cave Jerusalem Cave Negev Desert speleothems Age (ka)
050100150200250300350Age (ka) 0 100 200 300 400 500 60 0 Sr (ppm) 050100150200250300Age (ka) 0 100 200 300 400 500 60 0 Sr (ppm) Figure 38: Comparison between Srconcentrations in central Israel caves (Soreq and Jerusalem) (Ayalon et al, 1999; Bar-Matthews et al., 2001; Frumkin and Stein, 2004) and caves of northern Negev (Tzavoa and Maâ€™ale-Dragot), Jordan Valley (Maâ€™ale-Efrayim) (A), and central and southern Negev (B). Legend:Hol-Zakh Ashalim Even-Sid-Ramon Ma'ale-ha-Meyshar Shizafon Ktora Cracks Jerusalem Soreq Legend:Ma'ale-Dragot Tzavoa Jerusalem Soreq Ma'ale-Efrayim A B
00.0050.010.0150.020.0251/Sr (1/ppm) 0.7079 0.708 0.7081 0.7082 0.7083 0.7084 0.708587Sr/86Sr 00.0050.010.0150.020.0251 / S r (1 / ppm) 0.7075 0.708 0.7085 0.709 0.7095 0.71 0.7105 0.71187Sr/86Sr 00.0050.010.0150.020.0251 / S r (1 / ppm) 0.7076 0.7078 0.708 0.7082 0.7084 0.708687S r / 86S r Figure 39: (A) 87Sr/86Sr vs1/Sr plot in speleothems, host rocks and bulk soils. The bulk compositions of soils probably define a mixing line (orange) between the proximal source (0.7083-0.7085?) and Saharan dust (0 .7165-0.72) (Kromet al., 1999), whereas speleothems define an a pparent mixing line, possibly a result of dilution w ith rainwater (green); (B) 87Sr/86Sr vs 1/Sr plot for vadose and phreatic speleothems including the Basal member; (C) 87Sr/86Sr vs 1/Sr plot for vadose speleothems younger than 550 ka, as wellas in Intermediate member. Enclosed data fields are for Tzavoa and Maâ€˜ale-Efrayim Cave speleothems. Their Sr isotopic values define the composition of the proximal source during the glacial and interglacial periods,with a slight increase in host rock contribution and dilution evident during the interglacials.Ktora Cracks Intermediate Member Basal Member Makhtesh-ha-Qatan phreatic speleothems Ma'ale-ha-Meyshar phreatic speleothems Host rocks Bulk soils Ma'ale-Efrayim Cav e Ma'ale-Dragot Cave Tzavoa Cave Hol-Zakh Cave Ashalim Cave Even-Sid-Ramon mini-caves Ma'ale-ha-Meyshar Cave Shizafon mini-caves Legend:A B C Terra-Rosa soils Negev Desert soilsGlacial Interglacial Dilution Host rock Dust-born soil Glacial and Interglacial Maâ€™ale-Efrayim data field Tzavoa Cave data field Maâ€™ale-Efrayim (Jordan Valley) Maâ€™ale-Dragot (northern Negev)
SQ J 1 3 4 6 8 9 12 13 14 7J o rd an Va lle yC e n t r a l I sr a e lNegev DesertD e a d S e aSinai Peninsula 180 210 1 8 0 1501 2 0 90 60 Nile Delta Figure 40: Annual dust fall (gr/m2) map of Ganorand Foner(1996, 2001) and the location of the studied caves and the caves of central Israel. The marks and labels of the caves are is as in Fig. 36. The wind direction during the passage of the Cyprus cyclones is marked bythe arrow.
Appendix 2 â€“ The studied caves Cave positions were determined by global position (GP) system. Ma'ale-Efrayim Cave (#1 in Fig. 1C) â€“ GP: 32 03'05.81"N, 35 23'03.34"E. At an elevation of 250 m asl, located on the eastern slope of the Central Mountain Ridge (CMR) of Israel near the Jordan Valley, 60 km of the east from the Mediterranean Sea coast. The crest of the CMR is 900-1000 m asl and 10 km to the west of the cave, and receives 650-700 mm annual rainfall. The topogr aphically depressed position of the cave relatively to westerly winds causes a deep "rain-shadow", and present day annual precipitation above the cave is 250-300 mm, with average yearly temperature of 21-22 C. The cave depth is 2-20 m below the surface. The cave was probably formed along NESW tectonic fissure in Cenomanian dolomite hos t rock (Fig. 2). The cave was discovered during road cutting and before it had no natural opening. At present, the lowest part of the cave traps cold air during the winter, whereas the temperature in the upper part is around the annual mean value. The speleothems inside the cave are vadose (formed in unsaturated zone) and include stalagmites (m ore than 2 m in height), stalactites and flowstone. Kanaim Cave (#2 in Fig. 1C) â€“ GP: 31 18'07.43"N, 35 18'23.42"E. At an elevation of 280 m asl, located on the eastern slope of the CMR in Judea Desert, 9 km NE of town of Arad and 8 km to the west of Dead Sea. Th e southern edge of CMR is 600-900 m asl and located 22 km to the west, receives 200-350 mm annual rainfall. The topographically depressed position of the cave relatively to westerly winds causes a "rain shadow" and the present day precipitation above the cave is ~100 mm, with average yearly temperature of ~21-22C. The cave depth is 5-30 m below the surface. The cave system is hundreds of meters in length and contains gypsum speleothems, which locally reach thicknesses of a few cm on the cave walls. Vadose calc ite speleothems are found in a few deep chambers >50 meters from the entrance and mainly include stalagmites and stalactites (up to tens of cm in length). The cave has one natural opening. Maâ€™ale-Deragot Cave system (#3 in Fig. 1C) GP: 31 18'3.09"N, 3504'4.12"E. At an elevation of 630-720 m asl, located in north ern Ma'ale-Dragot (Tel-Arad) Quarry. The caves are located on the southern edge of CM R of Israel a few km to the north from Be'er-Sheva Arad Valley and 65 km to the E-SE from the Mediterranean Sea. The cave
system is within Cenomanian and Turonian dolomite and limestone host rocks (Fig. 2) and the depth of the caves is 10-90 m belo w the surface. The mean annual present-day precipitation above the caves is 280-300 mm; and the mean annual temperature is ~1718C. Until the caves were discovered during quarrying, they had no natural openings. The stratigraphically oldest speleothems in lowest part of the caves are phreatic, composed of alternating calcite and hematite laminae. Most of the speleothems are vadose and includes stalagmites (more than 2 m in height), stalactites and flowstone. Tzavoa cave (#4 in Fig. 1C) â€“ GP: 31 12'3.24"N, 35 13'4.38"E. At an elevation of 550 m asl on the eastern flank of m ountain ridge (~70 m below the ridge crest), about 4 km of the south from town of Arad, and 80 km ESE from the Mediterranean Sea. The cave system is located in Turonian limestone host rock (Fig. 2), 20-50 m below the surface, within present-day 150-160 mm isohyets, w ith mean annual temperature of 18-19C. The cave system is hundreds of meters in length a nd contains vadose speleothems: stalagmites (more than 1 m in height), stalactites and flow stone. Calcite veins tens of cm in width are found in the host rock. Field relations between the cave and the veins show that these veins are older than the cave space. The cave has several natural openings. Izzim Cave (#5 in Fig. 1C) â€“ GP: 31 08'2.24"N, 35 03'5.11"E (limestone quarry). At an elevation of ~500 m above the sea level, 72 km SE from the Mediterranean Sea coast. The cave is a few tens of meters in length, a nd located in Turonian limestone host rock (Fig. 2), 10-15 m below the surface. Before it was discovered by quarrying, the cave had no natural openings. The mean annual rainfa ll above the cave is of 120-140 mm and the mean annual temperature is ~18-19 C. The cave contains vadose calcite speleothems: stalagmites (of up to 2 m in height), st alactites and flowst one. Numerous gypsum speleothems are also found in the cave, some of them formed on the surface of the calcite speleothems. Hol-Zakh Cave (#6 in Fig. 1C) â€“ GP: 3109'2.24"N, 3512'0.37"E (limestone quarry). At an elevation 150 m asl. This is a small cave located within Tu ronian limestone host rock (Fig. 2), at depth of 15-20 m below the surface, 84 km south-east from the coast of eastern Mediterranean Sea, between present-day 80-100 mm isohyets, with mean annual temperature of 21-22 C. Although the cave located in th e same latitude of the Izzim Cave, the rainfall amount is smaller because of its location in the "rain shadow" of the
mountain ridge to the west. The cave was exposed during quarrying, has no natural openings, and contains vadose speleothems: st alactites and flowstone crusts of maximum length/thickness of ~3 cm. Ashalim Cave (#7 in Fig. 1C) â€“ GP: 30 56'3.62"N, 3444'2.25"E. At an elevation of ~400 m asl, located ~2 km SE to Ashalim villa ge and 67 km SE from the coast of eastern Mediterranean Sea. The cave system is hundr eds of meters long, located in Turonian limestone host rock (Fig. 2) at depths of 0-25 m below the surface; between present-day 120 mm and 100 mm isohyets, with mean annual temperature of 19-20 C. The cave has one natural opening, a few meters in diameter with a funnel-like shape. Most of the speleothems are vadose: stalagmites, stalacti tes and flowstone with maximal thickness of 30-40 cm. In lowest parts of the cave some phreatic speleothems cover the cave walls, and "stalactites" of bat guano were found. Makhtesh-ha-Qatan Cave (#8 in Fig. 1C) â€“ GP: 30 57'0.53"N, 35 13'1.16"E. Elevation of ~20 m bsl. This is small cave located ~4 m above the Wadi Hazera (ephemeral stream) channel, in the outlet of Makhtesh-ha-Qatan (erosi onal crater), on the eastern flanks of mountain ridge rising ~550 m asl, and 99 km south-east from the coast of eastern Mediterranean Sea. The precipitat ion on the ridge crest we st of the cave is ~100 mm, but the depressed posit ion of the cave relatively to westerly winds creates a "rain shadow" effect, and the annual precipit ation above the cave is 50-70 mm with mean annual temperature of ~23 C. The cave located within Cenomanian dolomite host rock (Fig. 2), at depth of 0-4 m below the surface; it has one natural opening and its length is several meters. Today this cave is intermittently flooded by Wadi Hazera, and sandy flood sediments cover the cave floor. In the past the cave was probably below the channel of Wadi-Hazera and was flooded during the high groundwater levels. The cave contains phreatic calcite speleothems ~20 cm thick, al ternating with thin (<1 cm) laminae of vadose speleothems. Speleothems are composed of calcite and thin iron oxide laminae. Gypsum crystals are also found on the su rface of the calcite speleothems. Even-Sid-Ramon mini-caves (#9 in Fig. 1C) â€“ GP: 30 38'2.28"N, 34 48'3.57"E. At an elevation of ~800 m asl, located in a quarry 2 km to the nor th of the town of MitzpeRamon, 96 km south-east from the coast of the eastern Mediterranean Sea. The precipitation is 90-100 mm, with m ean annual temperature of ~17 C. Small caverns and
fissures in Cenomanian limestone host rock (Fig. 2), are located 2-6 m below the surface and were exposed during the quarrying. Before it they had no natural openings. They contain mainly vadose speleothems: stal actites, flowstone and pool deposits with maximum thickness of ~8 cm. Some of the cav es contain red clay sediments, probably paleosols washed into the caves from the surface. Mitzpe-Ramon Cave â€“ (#10 in Fig. 1C) â€“ GP: 30 36'3.54"N, 34 47'3.99"E. At an elevation of ~850 m asl. The cave is located in the western limit of the town of MitzpeRamon, 95 km from the coast of the eastern Me diterranean Sea. The ca ve length is a few tens of meters, and it is located at a de pth of 0-20 m below the surface in Cenomanian dolomite host rock (Fig. 2). Present day annua l rainfall is 90-100 mm with mean annual temperature of 16-17 C. The cave has one natural opening. It contains vadose speleothems, mainly flowstone a nd stalactites up to 20 cm thick. Wadi-Lotz Cave â€“ (#11 in Fig. 1C) â€“ GP: 30 28'1.22"N, 34 34'3.85"E. At an elevation of ~900 m asl. The cave is located ~2 km from Israeli-Egyptian border, ~100 km from the coast of the EM Sea. Th e rainfall above the cave is ~100 mm, and the mean annual temperature is 16-17 C. The cave is located within Eo cene limestone host rock (Fig. 2), inside a steep cliff. The cave has the shape of an overturned funnel 15 m in depth, starting with natural opening in its top (3-4 m in diameter), and widening towards its bottom, forming the large chamber with the major natu ral entrance of ~10 m in diameter. Most of the speleothems are vadose flowstone up to 15 cm thick, and either covers the walls in the upper part of the shaft, or are exposed on the surface above the upper opening and partly eroded. The cave floor is cove red by thick layer of bird guano. Ma'ale-ha-Meyshar Cave (#12 in Fig. 1C) â€“ GP: 30 29'3.56"N, 34 55'5.47"E (near road #40), at an elevation of 450 m asl. The cave is located 115 km S-SE from Mediterranean Sea coast, on 50 mm isohyet, with mean annual temperature ~19 C. The major cave was partly destroyed by road cutting. The remaining chamber is 10-12 m in length, 5-12 m below the surface, and it is a part of cave system composed of isolated small caves containing phreatic speleothem s 1-25 cm thick, covered by thin vadose flowstone, stalactites and stalagmites up to 3 cm thick. Reddish and beige clay sediments are found inside the caves (probably paleosol s washed to the caves). The host rock is
Cenomanian dolomite (Fig. 2). The cave had no natural opening before it was exposed by road cutting. Shizafon Quarry mini-caves (#13 in Fig. 1C) â€“ GP: 30 02'1.83"N, 3500'1.13"E (limestone quarry), at an elevation of ~400 m asl. The site is located 165 km S-SE off the coast of eastern Med iterranean Sea, between present-da y 40 mm to 20 mm isohyets, with the mean annual temperature is 20-21 C. Small vugs, caverns and cracks containing vadose flowstone 0.5-5 cm thick occur in U pper-Turonian limestone host rock (Fig. 2). The mini-caves are located at a depth of ~10 m below the surface and had no natural openings before exposed. Ktora Cracks (#14 in Fig. 1C) â€“ GP: 30 01'5.68"N, 3503'4.93"E; at an elevation of ~400 m asl. The site is located ~55 km north of town of Elat, 168 km S-SE from the coast of eastern Mediterranean Sea, between present-day 40 mm to 20 mm isohyets, with the mean annual temperature is 20-21C. This is an exposed wall of a collapsed cave covered by vadose flowstone 5-20 cm thick. Some of the speleothems are covered by rock fall and talus. This collapsed cave, and other caves in this site were formed within large fissures in Upper Turonian limestone host rock (Fig. 2) which developed as a result of sliding of large blocks of rock down slope (Yacoby, 2004). The walls collapsed in many of the cracks clogging the upper parts of the cracks with breccias fo rmed the cave ceiling. The direction of the crack forming the cave is 230 (SW-NE). Reference: Yacoby, Y. (2004). The open fractures in Ma'ale-Grofit area. MSc Thesis (in Hebrew, English abstract), Ben Gurion University of the Negev, Be'er-Sheva, Israel .
Appendix 3: Analytical procedures of U-Th and U-Pb dating methods A.3.1. U-Th dating: chemical separati on between U and Th (Kuperman, 2005) 1) Depending on the uranium concentrati on (Table 1), 10-800 mg calcite powder was dissolved in 7N HNO3. 2) The sample solutions were centrifuge d 10 minutes in 4000 rpm to deposit the insoluble residue (most of it was silicates). 3) The aliquots were separate d from the insoluble residues. 4) The insoluble residues were dissolved using a 5 ml mixture of concentrated HF and HNO3 and evaporated to dryness by heati ng in Teflon beaker. Silica was removed by reaction with HF yielding gaseous SiF4. The residues were dissolved in 7M HNO3 and added to the sample aliquots. 5) 236U-229Th spike was added to the solution and mixed. 6) The solution was evaporated to dryness in a Teflon beaker and then dissolved in 5 ml 7M HNO3. 7) 2ml Bio-Rad AG 1X8 200-400 mesh resi n was added into mini-columns and cleaned by passing through 2 aliquots of 5 ml 7M HNO3 and three aliquots of doubledistilled H2O. 8) The resin on columns was pre-conditi oned to the sample solution by passing through 5 ml of 7M HNO3. 9) The sample solution was loaded onto mini-columns. 10) The matrix was removed by passing through 2 aliquots of 2 ml 7M HNO3. 11) The resin on columns was pre-conditioned for Th collection by the addition of 1.5 ml HCL 6M. 12) Th was eluted into Teflon beakers by passing 3.5 M HCl 6M through the column. 13) The resin on columns was pre-conditio ned for U collection by adding of 1.5 ml of 1M HBr.
14) U was eluted to Teflon beakers by a ddition of 3.5 ml 1N HBr to the columns. 15) Evaporation of the U and Th solutions to dryness. 16) Dissolution of the residues in 2 ml (U) and 5 ml (Th) of 0.1M HNO3. 17) 50 l of U solution were dissolved in 2.5 ml 0.1M HNO3 (diluted by factor of 50) and the original U concentration in solution was measured using Perkin Elmer Sciex ICP-MS Elan 6000. 18) Samples with U concentration >50 ppb were diluted to 50 ppb to adjust the concentration to MC-ICP-MS Faraday Cups working conditions. 19) Addition of 25 l of UNBL 112A standard containing 50 ppb U to Th solutions and mixing. A.3.2. U-Th dating: measurement of U and Th isotopic ratios by MC-ICP-MS (Kuperman, 2005) High precision U-Th dating was performed using a Nu Instruments MC-ICP-MS, equipped with 12 Faraday cups and 3 ion c ounters. The sample was introduced to the MC-ICP-MS through an Aridus micro-concentric desolvating nebuliser sample introduction system. The solution injected in to an Ar plasma at ~6000K, the various elements are ionized and their beam is accelerated in electric field. The beam shape is smoothed by a system of high voltage lenses. The isotopic species of different masses are separated by magnet and are counted by th e collector system of 12 Faraday cups (H6, H5, H4, H3, H2, H1, Ax, L1, L2, L3, L4 and L5) and 3 ion counters (IC) (IC0, IC1 and IC2). The difference between the Fara day cups and the ion counters is in their sensitivity. Ion counters are sensitive to ion beams with low concentration, giving weak electric signal (234U, 230Th, 236U, 229Th) whereas the Faraday cups are built to measure ion beams with high concentration giving strong electric signal (i.e. 238U, 235U). U measurement were made in 3 cycles in which ion counters were calibrated relatively to Faraday cups and three atomic ratios 235U/238U, 234U/236U, and 235U/236U were determined. During the first cycle th e instrumental mass bias was corrected
(using the exponential equa tion) by measuring the 235U/238U atomic ratio and correcting with the natural 235U/238U ratio (0.0072), and the Faraday cups were calibrated to ion counter IC0. During the second cycle, 235U and 238U were measured by the Faraday cups and then were calibra ted to ion counter IC1. During the third cycle 235U and 238U were measured by the Faraday cups, and 236U and 234U were measured by ion counters, IC0 and IC1, respectively. Determination of 230Th/229Th and 232Th/229Th was performed by calibration to the isotopic ratios of uranium, because the collector configuration of MC-ICP-MS does not enable the direct measurement of 230Th. The measurement was performed in three cycles. During the first cycle 235U and 238U were measured by the Faraday cups and then they were calibrated to ion counter IC1, by which 229Th was measured. In second cycle 235U was measured by IC1 and 238U was measured by Faraday cups. During the third cycle, 235U and 238U were measured by the Faraday cups, and then they were calibrated to ion counter IC1 by which 230Th was measured. The age determination was possible due to the accurate determination of 234U and 230Th concentrations by isotope dilution analysis using the 236U-229Th spike: 234U/236U, 235U/236U, 235U/238U, 230Th/229Th and 232Th/229Th ratios were determined and the known ratio of 236U/229Th in spike (2000) enabled calculation of 230Th/234U ratio. 234U/238U ratio was calculated with reproducibility of ). A3.3 U-Pb dating method (Woodhead et al., 2006). A3.3.1 Chemical purification and extraction of U and Pb Five speleothem laminae ASH-15-C, ASH-15-D, ASH-15-E, KTO(1)-1-J and KTO(1)-1-K were sawn into slabs ~5-10 mm thick and 10-30 mm in length. From each lamina 4-5 sub-samples of 70-280 mg were cut, providing between 100 pg and 1.2 ng of Pb, depending on Pb content. After extraction from the speleothem, sub surfaces were cleaned by repeated ultras onification in 0.5 M HCl, followed by washing in ultra-pure water. Samples were then dried and weighted. All material handling subsequent to cutting was conducted in a multiple-HEPA-filtered
environment, and all acids utilized in th e dissolution and chemical extraction of Pb were (as a minimum) twice distilled in either quartz or Teflon stills. Following Woodhead and Hergt (1997), samples were placed in new, cleaned 30 ml TPX polymethyl-pentene beakers with ~1 ml of ultrapure water, and dissolved by dropwise addition of 6M HCl. After complete dissolu tion (no inorganic residues were observed in any of the samples), sample soluti ons were aliquoted for spiking, using 233U-205Pb tracer. Pb and U were purified using conve ntional anion exchange chemistry with HBr/HCl media (employing a single column pass only to reduce blank contributions), and EICHROM TRU-resin with HNO3/HCl/HF media respectively (Luo et al., 1997; Manhes et al., 1978). Total processing blanks were 5-15 pg Pb and < 5 pg U; a value of 10 pg has been used for the former in the blank subtraction calculation. A3.3.2 Measurement using MC-ICP-MS Isotopic analyses were carri ed on a Nu Plasma MC-ICP-MS coupled to a Cetac Aridus desolvation system with a Glass E xpansion OpalMist nebu lizer (uptake rate ~40 l/min). At a typical sensitivity of 100-12 0V/ppm, Pb samples, dissolved in 0.5 ml of 0.3 M HNO3, yielded total Pb signals upwards of 4 10-13A (40mV) for a 200 pg sample size, allowing 205, 206, 207, 208Pb to be measured on Faraday cups. The correction for mass bias was performed by reference to external standards. For radiogenic samples (samples without non radiogenic Pb) the error attributable to the mass bias correction was only a minor component of the overall analytical uncertainty. U isotope dilution analyses were performed using the natural 235U/238U ratio in the sample for internal bias correc tion, after correction for a minor 235U contribution from the 233U spike. The 234U/238U of the samples was measured in Geological Survey of Israel using methods described in Vaks et al. (2006). Initial 234U/238U activity ratios were calculated using the formula [234U/238U]0 = 1 + (([234U/238U]s â€“ 1) exp( 234t)) (Kaufman et al., 1998), where square bracket s indicate the activity ratio, subscripts 0 = initial, s = measured in sample, t = time in years, and 234 is the decay constant for 234U, for which value of 2.835 10-6 was used. Blank corre ction and preliminary
isochron construction (prior to disequilibrium correction) were performed using the data reduction software PBDAT (Ludwig, 1993) and Isoplot Ex. 3.04 (Ludwig, 2001). A3.3.3 Isochrone formalization Use of the isochron co nstructions including 204Pb and 208Pb data was avoided because of the interference of the 204Pb mass by 204Hg, and because the concentrations of 208Pb in highly radiogenic speleothem samples (206Pb/204Pb>500) are vanishingly small. Isochron construction instead used the Terra-Wasserburg (T-W) diagram (Terra and Wasserburg, 1972), renamed the â€œsemi-total Pb-U isochronâ€ by Ludwig (1998). The T-W diagram uses parameters 238U, 207Pb and 206Pb, which are well determined on the Faraday cups. In the case of a suite of coeval, cogenetic speleothem subsamples with differing 206Pb/204Pb, plots of 238U/206Pb and 207Pb/206Pb data uncorrected for both blank and common Pb should define a mixing line between common Pb (a combination of blank and intrinsic common Pb) and pure radiogenic Pb located on the T-W Concordia (although in the case of speleothems the latter curve is influenced by initial isotopic disequilibr ium effects, see below). Plots of isotope data determined on different sub-samples from the same lamina show little systematic difference in the resulting ages when cal culated either with or without blank correction, although usually age errors are reduced for bla nk corrected data. Therefore the data were blank corrected before plotting and calculati ng T-W disequilibrium Concordia intercepts with Isoplot Ex (Ludwig, 2001). A3.3.4 Correction for U series disequilibrium Speleothems precipitate out of U-series equilibrium and 234U/238U excess in cave seepage waters frequently leading to wide range of 234U/238U values in speleothems. In addition, the time required for 230Th and 231Pa ingrowth (clean speleothems precipitate with essentially no Th and Pa) in the 238U-206Pb and 235U-207Pb decay chains must also be taken into account. The correction for initial 234U and 230Th was achieved using the Isoplot spreadsheet functi ons of Ludwig (2001) incorporated into
an â€˜in houseâ€™ iterative r outine. The present day 234U/238U activity ratios of all five speleothems are very close to equilibrium. For estimation of 234U/238U activity ratios in flowstone samples KTO(1 )-1-J and K, the initial 234U/238U=1.06 activity ratio of the youngest lamina A1 of this flowstone (d ated by U-Th to 137 .7 1.5 ka, Table 1) have been used. In ASH-15-C, D and E laminae the average initial 234U/238U=1.05 activity ratio for the young speleothems from this cave have been used. References of Appendix 3: Kaufman, A., Wasserburg, G. J., Porcelli, D., Bar-Matthews, M., Ayalon, A., and Halicz, L. (1998). U-Th isotope system atics from the Soreq cave, Israel and climatic correlations. Earth and Planetary Science Letters 156, 141-155. Kuperman, G. (2005). "Reconstruction of fl ood periods by dating speleothems in Nahal Hazera, northern Negev." Hebrew University of Jerusalem, M.Sc. Thesis (in Hebrew, English Abstract). Ludwig, K. R. (1993). PBDAT: a computer program for processing Pb-U-Th isotope data. In "U.S.G.S. Open File Report 88-542." pp. 33. Ludwig, K. R. (1998). On the treatmen t of concordant uranium-lead ages. Geochimica et Cosmochimica Acta 62, 665-676. Ludwig, K. R. (2001). Isoplot/Ex rev. 2.49. Special Publication, 1a. In "A geochronological tool kit for Microsoft Excel." (K. R. Ludwig, Ed.). Berkeley Geochronology Center, Berkeley, USA. Luo, X., Renkamper, M., Lee, D.-C., and Halliday, A. N. (1997). High precision 230Th/232Th and 234U/238U measurements using energy-filtered ICP magnetic sector multiple collector mass spectrometry International Journal of Mass Spectrometry Ion Proceeding 171, 105-117. Manhes, G., Minster, J. F., and Allegre, C. J. (1978). Comparative uranium-thoriumlead and rubidium-strontium study of the Saint Severin amphoterite: consequences for early solar system chronology. Earth and Planetary Science Letters 39, 14-24.
Terra, F., and Wasserburg, G. J. (1972). UTh-Pb systematics in three Apollo 14 basalts and the problem of in itial Pb in lunar rocks. Earth and Planetary Science Letters 14, 281-304. Vaks, A., Bar-Matthews, M., Ayalon, A., Matthews, A., Frumkin, A., Dayan, U., Halicz, L., Almogi-Labin, A., and Sc hilman, B. (2006). Paleoclimate and location of the border betw een Mediterranean climate region and the SaharoArabian Desert as revealed by speleo thems from the northern Negev Desert, Israel Earth and Planetary Science Letters 249, 384-399. Woodhead, J. D., and Hergt, J. M. (1997). A pplication to the 'double spike' technique to Pb-isotope geochronology. Chemical Geology (Isotope Geoscience) 138, 311-321. Woodhead, J., Hellstrom, J., Maas, R., Drys dale, R., Zanchetta, G., Devine, P., and Taylor, E. (2006). U-Pb geochronol ogy of speleothems by MC-ICPMS. Quaternary Geochronology 1, 208-221.
Appendix 4: Tables of results
Table 1 page 1: U-Pb ages of speleothemsCave number at Fig. 1C Cave/Cite Sample ID Sample mass (gr) Pb (ppb) U (ppb)206Pb/238U (%)206U/207U (%)238U/206Pb (%)207Pb/206Pb (%) Isochron slope Intersept with 207Pb/206Pb axis Age(Ma)for equilibrium initial 234U/238U Age(Ma) for assumed initial 234U/238U Assumed initial 234U/238U 8 Ashalim ASH-15-C 0.110 1.53 4298 0.0002420 0.656 4.3470 1.900 4132.7 0.656 0.230044 1.900 -0.000134 0.000001 0.783 0.002 1.286 0.008 1.269 0.013 1.05 0.03 Cave Flowstone 0.130 6.41 3288 0.0007317 0.265 1.6675 0.156 1366.7 0.265 0.599700 0.156 lamina 0.070 1.64 4102 0.0002550 0.966 3.8683 2.420 3921.6 0.966 0.258511 2.420 0.090 1.73 3870 0.0002692 0.779 3.4878 1.650 3714.2 0.779 0.286714 1.650 0.070 2.35 3313 0.0003517 0.869 2.5271 1.120 2843.0 0.869 0.395710 1.120 ASH-15-D 0.254 2.42 1946 0.0007046 0.241 3.1371 0.388 1419.2 0.241 0.318766 0.388 -0.000358 0.000007 0.827 0.011 3.063 0.018 3.045 0.021 1.05 0.03 Flowstone 0.129 1.65 1612 0.0006390 0.542 3.7323 1.270 1564.9 0.542 0.267931 1.270 lamina 0.267 1.92 1859 0.0006398 0.261 3.7043 0.535 1563.1 0.261 0.269957 0.535 0.267 1.56 1848 0.0005857 0.275 4.6425 0.791 1707.4 0.275 0.215401 0.791 0.278 1.19 1650 0.0005499 0.307 5.6635 1.160 1818.4 0.307 0.176569 1.160 ASH-15-E 0.187 1.36 1960 0.0005406 0.376 5.9810 1.580 1850.0 0.376 0.167196 1.580 -0.000360 0.000008 0.831 0.015 3.066 0.009 3.048 0.014 1.05 0.03 Flowstone 0.225 1.29 2426 0.0004920 0.301 10.0470 2.130 2032.4 0.301 0.099532 2.130 lamina 0.146 1.51 1846 0.0005781 0.465 4.8113 1.520 1729.8 0.465 0.207844 1.520 0.174 1.17 1931 0.0005151 0.425 7.4687 2.360 1941.4 0.425 0.133892 2.360 0.204 1.13 1945 0.0005084 0.37 8.0294 2.190 1967.0 0.370 0.124542 2.190 14 Ktora KTO(1)-1-J 0.194 3.04 616 0.0019118 0.264 1.7436 0.155 523.1 0.264 0.573526 0.155 -0.000272 0.000008 0.715 0.005 2.734 0.059 2.713 0.060 1.06 0.02 Cracks Flowstone 0.153 3.33 626 0.0020394 0.38 1.7215 0.229 490.3 0.380 0.580889 0.229 lamina 0.270 2.37 532 0.0017548 0.308 1.7882 0.202 569.9 0.308 0.559222 0.202 0.280 2.02 586 0.0014228 0.324 1.9085 0.243 702.8 0.324 0.523972 0.243 KTO(1)-1-K 0.146 1.54 338 0.0018544 0.776 1.8809 0.556 539.3 0.776 0.531660 0.556 -0.000334 0.000008 0.714 0.002 3.338 0.062 3.317 0.063 1.06 0.02 Flowstone 0.186 6.50 552 0.0042229 0.204 1.5755 0.124 236.8 0.204 0.634719 0.124 lamina 0.224 3.15 376 0.0031084 0.295 1.6498 0.169 321.7 0.295 0.606134 0.169 0.230 2.29 347 0.0025233 0.374 1.7171 0.221 396.3 0.374 0.582377 0.221
Table 1 page 2: U-Th ages of speleothemsCave number at Fi g . 1C Region Cave/Cite Sample description Sub-sample numbe r LaminaAge (ka) 238U ppm234U/238U2 230Th/234U 230Th ppb 232Th ppb230Th/232Th 1Jordan Valle y Ma'ale-Efrayim Cav e Stalagmite ME-12 1ME-12-ou t 12.061.271.04 5 1.0435 3 0.0026 3 0.1050 8 0.0052 4 0.0018 7 10.0 5 35.72 2ME-12-A116.400.280.7171.0886 5 0.0009 2 0.1402 7 0.0011 3 0.0017 8 1.833185. 8 3ME-12-A 2 24.7 8 0.460.73 8 1.1086 9 0.001910.2044 6 0.0017 0 0.0027 3 3.326155. 8 4ME-12-K 7 71.8 8 2.530.1721.0900 9 0.0019 6 0.4874 9 0.0060 4 0.0014 9 5.10655.9 8 Conical stalactite ME-55ME-5-A1 49.21 4 1.3540.3171.0760 2 0.0027 7 0.3657 2 0.0039 4 0.0020 3 4.7182.77 6ME-5-A2.1 70.62 9 1.6210.2731.096710.0026 5 0.4816 8 0.0038 7 0.0023 5 10.6 5 42.11 7ME-5-A2. 2 71.92 2 1.0410.29 5 1.0788 7 0.0009 0 0.4872 6 0.0024 8 0.0025 3 2.76173.7 8ME-5-B 2 158.47 5 3.54 9 0.3001.0597 9 0.0016 7 0.7756 3 0.0036 9 0.0040 3 9.6179.2 9 9ME-5-B 3 157.12 2 3.6730.3441.073 3 0.001510.7743 9 0.0039 4 0.0046 6 15.0358.43 10ME-5-C1 211.5116.6440.2841.0416 2 0.0018 7 0.8654 7 0.0041 5 0.0041 8 18.0 9 43.66 11ME-5-C 2 221.18 6 9.8170.3271.0456 4 0.0020 2 0.8788 7 0.005710.0049 0 2.52367.6 12ME-5-C 3 219.43 7 5.4420.32 9 1.0481 3 0.0012 6 0.8771 8 0.0032 0 0.0049 4 12.0277.37 13ME-5-D1 288.034.30.3941.038810.0035 7 0.9398 7 0.0106 4 0.0062 8 13.8285.63 14ME-5-D 3 300. 8 34.80.4131.0224 0 0.0028 4 0.9435 3 0.0094 6 0.0064 9 1.93 5 631.6 15ME-5-D 5 2188.8.131.52 5 1.0332 0 0.002610.9296 6 0.0094 0 0.0063 5 4.316277.6 2Judea Desert Kana'im Cav e Small conical stalactite KN-116KN-1-A1 31.460.224.29 5 1.0313 0 0.0011 0 0.2515 3 0.0007 3 0.0181 6 44.1277.3 9 near Dead Sea 17KN-1-A 2 131.42.16.0540.9573 6 0.0008 6 0.6960 7 0.0027 0 0.0657 6 6.142201 5 18KN-1-B1 256. 9 6.03.9070.9976 8 0.0011 2 0.9052 2 0.0023 0 0.0575 3 1.97 9 545 6 Small conical stalactite KN-619KN-6-A1 184.108.40.20660.9787 0 0.0012 5 0.7571 5 0.0022 0 0.4488 9 0.6811238 7 Small conical stalactite KN-820KN-8-A 182.6 @ 2.33.5650.958810.001130.805740.001680.044901.5885314 21KN-8-B 146. 5 2.84.05 5 1.0008 2 0.0010 4 0.7400 3 0.003210.0489 6 0.8441087 6 Stalactite KN-9 22KN-9-A1 37.90 * 0.253.5771.0312 0 0.0011 0 0.2946 8 0.0007 6 0.0181 6 145.1 2 23.52 23KN-9-A 2 141.31.46.0070.9552 9 0.001210.7211 2 0.0015 9 0.0674 6 1.2561007 8 24KN-9-B 260.06.52.6660.9894 3 0.000910.9056 9 0.002410.0389 5 0.51114311 Stalactite KN-10(1) 25KN-10(1)-A (I ) 207. 5 3.83.57 8 1.0187 2 0.0014 7 0.8554 7 0.0022 8 0.0508 4 12.8 9 740.1 26KN-10(1)-A (II ) 247. 9 6.33.1961.006710.0009 6 0.8993 6 0.0027 3 0.0471 7 0.3222746 6 27KN-10(1)-B1259. 8 7.12.36 8 1.0055 2 0.0020 4 0.9096 9 0.0024 8 0.035310.32 8 2020 6 28KN-10(1)-B 2 3220.127.116.11 9 0.9975 4 0.0011 6 0.9482 4 0.0018 7 0.0794 0 0.3104811 7 29KN-10(1)-B 3 353. 9 18.94.4860.9975 2 0.0012 6 0.9605 6 0.0029 2 0.0700 7 0.3673586 0 5Central NegevIzzim Cave Stalagmite IZ-1 30IZ-1-A2 372.622.30.8371.015120.000950.972600.003100.0134715.90160 31IZ-1-A3 500.185.30.6731.007340.001420.992800.003290.010972.996696 6Central NegevHol-Zakh Cave Flowstone HZ-1 32HZ-1-A 18.104.22.16860.952610.001680.689500.001970.082731.39811101 33HZ-1-A1 22.214.171.12400.960410.001430.683130.001370.074548.7841596 34HZ-1-A2 126.96.36.19920.950250.00090.691880.001980.074845.7582442 35HZ-1-B 131.71.712.9270.930240.000990.693140.002080.135871.89913432 Stalactite HZ-2 36HZ-2-A 131.33.09.1820.936520.001390.692930.003800.0971412.981405 37HZ-2-B(I) 188.8.131.520.953560.001260.710660.003720.1513110.812626 38HZ-2-B(II) 184.108.40.206.966930.001410.710510.005120.1358311.332249 Stalactite HZ-3(2) 39HZ-3(2)-A 133.35.011.400.949940.001910.700170.006370.123593.4186820 40HZ-3(2)-B1 133.42.717.460.922780.001340.696630.003360.1829230.021149 41HZ-3(2)-B2 220.127.116.1180.997030.001380.683400.003020.084392.3136897 42HZ-3(2)-C1 200.05.613.000.97050.001530.834960.003710.171763.3659626 43HZ-3(2)-C2 222.25.019.080.964960.001320.862200.002570.258842.86017008 44HZ-3(2)-D 247.410.84.8591.009060.002280.899420.006020.071882.8164825 45HZ-3(2)-E 318.104.22.16890.984930.001850.954620.003950.134412.42910420 Stalactite HZ-3(5) 46HZ-3(5)-A1(I)22.214.171.12430.93660.001070.635600.002480.082815.6261781 47HZ-3(5)-A1(II)126.96.36.19900.939910.001160.698840.002580.077844.5133283 48HZ-3(5)-A2.1188.8.131.5200.935930.001310.698640.002270.085175.8422763 49HZ-3(5)-A2.2184.108.40.2060.953270.000890.692790.001410.078075.2812791 50HZ-3(5)-A3 126.31.318.360.931510.000760.678810.001740.1892928.341260 51HZ-3(5)-A5 220.127.116.111.012060.00110.687750.003050.081394.9643110 52HZ-3(5)-B 187.52.415.960.966310.000900.815180.001780.204964.4848603
Table 1 page 3: U-Th ages of speleothemsCave number at Fi g . 1C Region Cave/Cite Sample description Sub-sample numbe r LaminaAge (ka) 238U ppm234U/238U2 230Th/234U 230Th ppb 232Th ppb230Th/232Th 7Central NegevMakhtesh-ha-Qatan CaveStalactite MKTC-4 53MKTC-4-A 139.7*2.00.15201.189900.002200.743550.002380.0022019.6921 (intermittently grown 54MKTC-4-B >550 0.23281.151580.002731.075390.004970.004701.132812 submerged under water) Stalactite MKTC-5 55MKTC-5-A 131.7*6.60.76421.110200.002200.713850.008690.01000137.4014 (intermittently grown 56MKTC-5-C 18.104.22.168291.231650.002010.792500.004320.025031.8402590 submerged under water)57MKTC-5-H1296.19.00.98291.171630.001310.976970.002950.018340.8194205 58MKTC-5-H2303.512.51.26631.178710.001810.983670.003860.023941.9702279 59MKTC-5-H9322.214.171.12491.230560.001691.018840.003830.045943.1102271 8Central NegevAshalim Cave Stalactite ASH-2 60ASH-2-A >550 8.4160.999630.001510.996300.005220.136641.71514944 Stalactite (Column) ASH-1161ASH-11-A1.1126.96.36.19911.039810.000950.662370.002750.04371.4315761 62ASH-11-A1.2123.61.44.0551.027810.000780.681920.002080.04630.79111039 63ASH-11-A2.1127.45.92.7511.04340.001570.694740.008370.03250.9536541 64ASH-11-A2.2203.48.712.561.029470.000850.851770.008370.17951.11330363 65ASH-11-A2.3194.83.412.561.030110.000830.838860.002540.17691.29325755 66ASH-11-A2.4188.8.131.521.02710.000600.866540.003290.18331.32825979 67ASH-11-A3 >550 11.921.010490.001221.035570.002070.20331.74821815 68ASH-11-A4 >550 16.170.999170.002031.034000.002940.27232.38421428 Flowstone ASH-15 69ASH-15-A+B>550 1.8610.999390.001081.035320.001130.03142.3942460 70ASH-15-C1 >550 3.4120.999250.001501.025070.002190.05700.76613956 Frowstone ASH-33 71ASH-33-A1 184.108.40.20681.038920.002090.661660.004710.036962.3713013 72ASH-33-A2 124.03.63.3711.029480.001910.683270.005130.038662.1723508 73ASH-33-B1 220.127.116.1131.025850.001030.703730.003220.042171.3965805 74ASH-33-B2 18.104.22.1681.027450.000880.698070.003200.048591.0358969 75ASH-33-B3 22.214.171.1241.024530.000870.698030.004060.048521.1887780 76ASH-33-B4 127.53.63.6341.024560.001380.693120.005060.042071.8244447 77ASH-33-C1 201.96.09.8541.01670.000800.847050.004190.138342.57610180 78ASH-33-C2 199.77.011.731.018260.000780.844130.005020.164342.30613498 79ASH-33-D1 203.04.711.531.018340.000880.849060.003230.162484.4446934 80ASH-33-D2 208.34.111.001.018390.001170.856540.002560.156446.4064629 81ASH-33-E1.1207.93.612.371.018190.000610.855900.002380.175751.25426447 82ASH-33-E1.2213.64.714.271.019960.000600.863940.002950.204952.62714713 83ASH-33-E2.1218.02.813.291.019970.000730.869610.001610.192173.04111929 84ASH-33-E2.2(I)199.52.515.591.027370.001100.846280.002260.221022.25818499 85ASH-33-E22(II)126.96.36.1991.022470.000810.844950.002950.2060021.011850 86ASH-33-E22(III)198.84.314.051.023680.002470.843680.003110.197855.7006574 87ASH-33-F-up>550 13.181.019300.001211.014960.003510.222352.58916219 Flowstone ASH-34 88ASH-34-A1 188.8.131.5251.042000.000780.667370.002640.043145.3331526 89ASH-34-A2 184.108.40.20661.057370.001080.682930.002720.042690.46317381 90ASH-34-B1 127.21.63.0931.046000.000690.694340.002220.036621.5104581 91ASH-34-B2 220.127.116.1131.057820.001060.711390.002290.057321.5427003 92ASH-34-C1 199.33.312.841.029630.000930.845750.002290.1822910.753196 93ASH-34-C2 200.14.415.251.025010.000660.846050.003170.2156014.472820 94ASH-34-C3(I)192.32.919.681.029450.00070.834840.002220.2757215.403370 95ASH-34-C3 (II)188.46.913.991.022990.000950.827270.005570.1930413.872636 96ASH-34-C3 (III)194.64.315.191.026890.002310.837970.002960.2130111.003667 97ASH-34-D >550 16.451.012150.00130.999390.004310.271291.74029461
Table 1 page 4: U-Th ages of speleothemsCave number at Fi g . 1C Region Cave/Cite Sample description Sub-sample numbe r LaminaAge (ka) 238U ppm234U/238U2 230Th/234U 230Th ppb 232Th ppb230Th/232Th 9Central NegevEven-Sid-Ramon mini-cavesFlowstone ESID-198ESID-1>500*0.058941.013280.001391.113360.001537.27600.00108429 (single lamina) Stalactite (Column) ESID-299ESID-2-A18.104.22.168.428 2 1.0934 6 0.0013 2 0.5591 7 0.0028 3 0.004 3 1.46 5 566 100ESID-2-A1.222.214.171.12491.068440.001260.861830.005780.00312.690239 101ESID-2-A22126.96.36.199461.065050.000610.949570.002700.01110.9892134 102ESID-2-B516.450.62.1931.056010.000681.012560.002090.03821.3625282 103ESID-2-C>5500.11821.006330.001251.427950.019550.00280.3361553 Stalactite SHAR-1104SHAR-1-A>5500.15061.017880.001051.109660.004018.25290.00277364 Stalactite ESID-7(1)105ESID-7(1)-A1+A2188.8.131.5261.050060.001040.712970.004170.00182.065177 106ESID-7(1)-A34184.108.40.206541.029940.00210.997010.003040.00834.245375 107ESID-7(1)-B462.798.80.88101.020070.004020.993340.004870.01469.399298 108ESID-7(1)-C1>5500.31531.011460.000781.024040.002070.00531.476680 Stalactite ESID-7(3)109ESID-7(3)-A1212.3*3.30.28111.041590.000980.873880.001920.0041726.6630 110ESID-7(3)-B304.99.01.5381.013690.000080.943410.002530.023972.8641582 Stalactite ESID-11111ESID-11-A555.7201.30.18911.055950.001961.015660.047000.003310.889701 Stalactite ESID-13112ESID-13-A>5502.2551.029110.00131.009760.002080.038212.2993119 113ESID-13-B1>5500.93651.017630.001491.035550.003370.016091.3932172 114ESID-13-C>5500.76051.053240.001091.077130.002200.014071.0902427 Stalagmite ESID-18115ESID-18-A1>5500.73661.029390.000891.032340.002580.012763.023793 116ESID-18-A2>5501.1401.028730.000891.040060.003520.019892.8501312 117ESID-18-E>5501.8871.013710.001061.047030.002430.032651.2125055 Stalagmite ESID-22118ESID-22-A1>5500.60741.029490.001211.021730.003040.010424.430442 119ESID-22-A2>5501.1361.025490.001071.037530.002950.019711.6532240 120ESID-22-D>5501.5581.015950.001241.035450.003080.026721.6133111 11Central NegevWadi Lotz CaveStalagmite fragment LOTZ-3121LOTZ-3-A>5500.37941.04840.000851.048940.001850.006802.530509
Table 1 page 5: U-Th ages of speleothemsCave number at Fi g . 1C Region Cave/Cite Sample description Sub-sample numbe r LaminaAge (ka) 238U ppm234U/238U2 230Th/234U 230Th ppb 232Th ppb230Th/232Th 12Southern NegevMa'ale-ha-Meishar CaveFlowstone MMR-7122MMR-7-A1.1220.127.116.11431.071820.001160.654830.001910.0054311.5889 123MMR-7-A1.218.104.22.168581.070820.001330.644410.001940.005463.685281 124MMR-7-A3.122.214.171.124241.064280.002030.662950.002970.003780.881820 125MMR-7-A3.2194.63.00.38241.076510.001060.846650.002230.0056816.1167 126MMR-7-B>5500.71641.029940.000661.012820.002860.012189.154280 Flowstone MMR-7(2)127MMR-7(2)-A1126.96.36.199871.072950.001990.677350.002730.005443.358305 128MMR-7(2)-A2127.0*4.30.61271.07830.002040.697800.002900.0075749.7629 129MMR-7(2)-B203.09.40.29861.063260.001360.857470.006750.004443.283256 Flowstone MMR-8 130MMR-8-A 188.8.131.52711.068650.001190.851370.002790.004116.268125 131MMR-8-B 184.108.40.20651.064270.000920.862190.001840.003705.193135 Stalagmite MMR-13132MMR-13-A 138.04.10.37571.061670.003030.726100.005120.0047210.2387 133MMR-13-B1220.127.116.11851.060270.001410.687860.004640.0043811.6971 134MMR-13-B218.104.22.168041.058250.001410.711310.007030.004552.532339 135MMR-13-C 22.214.171.124401.058250.001530.730750.004360.004707.731115 136MMR-13-E 132.02.60.36541.060140.001260.709990.003340.004481.665507 137MMR-13-F 126.96.36.199711.065930.000750.712590.002980.004554.782179 138MMR-13-G 188.8.131.52351.070340.001520.709630.005770.005003.591263 139MMR-13-H 184.108.40.206311.072060.001390.696750.003320.005882.862387 140MMR-13-I 220.127.116.1161.077530.000930.708300.002390.021357.516533 141MMR-13-K >550 0.38111.014350.00271.019460.005650.006423.767321 13Southern NegevShizafon mini-cavesFlowstone SHACH-1142SHACH-1-A341.712.00.65781.024980.001170.964700.002120.0106024.9980 143SHACH-1-B333.419.00.30811.01150.001140.957000.003790.004863.376274 Flowstone SHACH-3144SHACH-3-A351.712.60.21351.038830.001050.972960.002110.003524.426150 145SHACH-3-B431.664.70.22321.024170.002170.989880.005000.003691.732401 146SHACH-3-C>550 0.20511.023910.000661.018320.003490.003491.108600 147SHACH-3-E>550 0.18451.045160.001151.030290.002900.003241.389449 Pool crust SHACH-4148SHACH-4 488.9*34.90.47211.028720.001410.999780.001400.0079272.7621 Flowstone SHACH-6149SHACH-6-A>550 0.35521.032310.000811.011450.000840.006054.021284 14Southern NegevKtora Cracks Flowstone KTO-1(1)150KTO-1(1)-A1137.7*1.50.98231.04330.001470.723180.001600.01232224.5210 151KTO-1(1)-A2318.104.22.168641.025160.000820.954650.001900.0159373.2241 152KTO-1(1)-C >550 0.80181.007370.001191.001080.003680.013184.085619 *Ages corrected for 230Th from detrital origin, other ages don't need correction because their 230Th/232Th ratios are above 30 (Kaufman et al., 1998). @ Age overturn in lamina KN-8-A, relative to KN-8-B. KN-8-A is a surface lamina, affected by condensation corrosion which possibly induced radionuclide remobilisation (Auler and Smart, 2004). Thus the age is not reliable. References : Auler, A. S., and Smart, P. L. (2004). Rates of condensation corrosion in speleothems. Speleogenesis and Evolution of Karst Aquifers (www.speleogenesis.info) 2, p.2 (4 pages). Kaufman, A., Wasserburg, G. J., Porcelli, D., Bar-Matthews, M., Ayalon, A., and Halicz, L. 1998, U-Th isotope systematics from the Soreq Cave, Israel, and climatic correlations., Earth and Planetary Science Letters, v. 156, p. 141-155.
Table 2: D and 18O values of present-day rainfall in the Negev Desert page 1Station with GP and elevation sampledate of rain eventamount (mm) 18 O D d-excessBeer-ShevaBSH-129.10.04 52.3-2.1-8.57.9 GP: 31 16' 39" N; 034 49' 23" E BSH-217.-18.11.04 18.4-4.3-9.824.3 Elevation: 380 m aslBSH-318.11.04 9.2 -4.2-11.122.3BSH-421-22.11.04 68.1-5.4-8.234.7BSH-526-28.11.04 19.1-3.03.027.1BSH-607-08.12.04 10.4-4.3-15.618.5BSH-712.12.04 drops BSH-815-16.12.043.6-5.0-13.726.5BSH-924-25.12.04 26.6-11.8-68.325.7BSH-1025-26.12.04 3.7 -4.9-18.420.5BSH-1102-04.01.05 22.8-5.6-28.616.0BSH-1205-06.01.05 36.1-9.1-54.019.0BSH-1306-07.01.05 1.2 -4.2-10.823.1BSH-1419-20.01.05 3.7 -4.7-19.417.9BSH-1524.01.05 14.8-4.4-17.218.2BSH-1606-09.02.05 39.6-2.9-3.319.5BSH-1711-12.02.05 6.6 -3.7-8.021.3BSH-1812-13.02.05 2.5 -6.4-31.719.6BSH-1921.02.05 1.4 -1.37.417.8BSH-2004.03.05 1.5 -1.2-8.11.6BSH-2108-10.03.05 40.9-3.2-13.712.2BSH-2212.03.05 4.3 -22.214.171.124BSH-2302.04.05 0.6 -3.2-9.415.9BSH-2403.04.05 1.4 -1.47.318.7SummaryWinter 2004-2005388.7 mm -4.8-18.220.6(Vaks et al., 2006) BSH-10118-21.10.05 4.1 -0.49.613.1BSH-10207.11.05 1.9 -2.7-5.316.1BSH-10315-16.11.05 4.6 -2.5-2.117.7BSH-10421.11.05 13.3-4.3-12.521.5BSH-10522.11.05 5.9 -1.1-2.16.7BSH-10620.12.05 3.3 -2.94.027.0BSH-10723.12.05 5.5 -3.6-1.626.9BSH-10824-26.12.05 23.1-4.1-7.825.2BSH-10904.01.06 0.6 6.337.4-13.3BSH-11007-09.01.06 11.5-4.4-8.126.9BSH-11115.01.06 0.3 -2.4-5.713.1BSH-11216-18.01.06 1.7 -1.77.321.0BSH-11326-28.01.6 5.0 -4.2-7.726.0BSH-11428.01.06 3.6 -126.96.36.199BSH-11503.02.06 4.2 -2.8-6.715.9BSH-11604.02.06 0.7 -0.74.59.9BSH-11713-17.02.06 61.2-6.8-29.224.8BSH-11826.02.06 0.2 4.526.0-10.1BSH-11909-10.03.06 3.6 -1.40.511.3BSH-12001-02.04.06 12.6-6.0-40.67.1BSH-12104-06.04.06 23.7-4.2-19.513.9BSH-12216.04.05 10.2-3.2-17.38.4BSH-12324.04.06 5.3 -1.3-7.43.1BSH-12425.04.06 0.5 2.612.7-8.1SummaryWinter 2005-2006206.5 mm-4.5-16.519.4Summar y 2004-2006595.2 mm-4.7-17.620.2
Table 2 page 2Station with GP and elevation sampledate of rain eventamount (mm) 18 O D d-excessArad GP: 31 16' 01" N; 35 13' 11" E ARAD-129.10.04 31.3-3.1-10.814.2 Elevation: 605 m asl ARAD-201.11-14.12.04 51.8-3.2-11.514.0 ARAD-314.12-02.02.05 35.9-7.2-28.528.8 ARAD-403.02-13.03.05 41.4-3.7-2.327.0 ARAD-513.03-22.05.05 0.71.35.7-4.9 SummaryWinter 2004-2005161.1 mm-4.2-12.720.6 ARAD-10102.10-15.12.055.1-2.4-0.319.1 ARAD-10215.12.05-02.02.0611.1-3.6-11.816.6 ARAD-10302.02-07.03.06 19.7-5.6-28.416.5 ARAD-10407.03-10.04.06 6.9-3.6-13.3515.5 ARAD-10510-30.04.06 10.1-2.3-12.75.4 SummaryWinter 2005-200652.9 mm-4.0-17.214.5Summar y 2004-2006214 mm-4.1-13.819.1Makhtesh-ha-Qatan GP: 30 57' 14" N; 35 12' 34" E Elevation: Sea level MQ-129.10.04 65.0-2.5-12.57.7 MQ-201.11-14.12.04 14.62.525.35.4 MQ-314.12.04-02.02.05 15.6-3.9-7.424.0 MQ-403.02-13.03.05 26.3-1.90.515.9 SummaryWinter 2004-200 5 121.5 mm-1.97-4.4911.3 ***02.10-15.12.05 no rain ***15.12.05-02.02.06 no rain MQ-10102.02-07.03.06 10.40.23.01.8 MQ-10207.03-10.04.06 7.3-7.7-27.034.52 MQ-10310-30.04.06 7.0-2.6-18.82.05 SummaryWinter 2005-200624.7 mm-2.9-11.911.4Summary $ 2002-2006 $ 216.9 mm $ -2.4-6.113.4Mitzpe-Ramon GP: 30 36' 59" N; 034 48' 06" E MR-129.10.04 5.62.530.310.5 Elevation: 820 m asl MR-201.11-14.12.04 11.130.816.59.9 MR-314.12.04-02.02.05 18.2-5.6-30.814.2 MR-403.02-13.03.05 13.3-3.5-4.423.8 SummaryWinter 2004-200548.2 mm-2.62-5.5115.4 MR-10102.10-15.12.05 1 -0.92229.28 MR-10215.12.05-02.02.06 4.1-3.97.838.6 MR-10302.02-07.03.06 25.1-4.2-21.112.5 MR-10407.03-10.04.06 7.9-3.9-15.315.5 MR-10510-30.04.06 11.8-3.6-19.39.5 SummaryWinter 2005-200649.9 mm-3.9-16.514.7Summar y 2004-200698.1 mm-3.3-11.115.1
Table 2 page 3Station with GP and elevation sampledate of rain eventamount (mm) 18 O D d-excessNeot-Smadar GP: 30 02' 55"N; 35 01' 27" Enot sampled#29.10.04 5.0lostlost Elevation: 405 m asl NSM-122.11.04 0.2-1.921.737.1 NSM-2A16.12.04 0.7-4.3-10.823.5 NSM-2B24-25.12.04 0.2lostlost NSM-2C05.01.05 0.8lostlost NSM-2D06.01.05 0.1lostlost NSM-2E09.01.05 1.5-2.9-23.5-0.1 NSM-2F16.01.05 0.80.82.9-3.7 NSM-3A11.02.05 0.50.923.816.9 NSM-3B12.02.05 drops NSM-3C20.02.05 0.10.82.9-3.7 NSM-3D08.03.05 2.2-1.3-9.31.3 NSM-3E09.03.05 5.0-2.9-8.714.4 NSM-3F10.03.05 1.3-6.1-12.336.3 NSM-423.04.05 0.55.128.8-12.1 Summary*Winter 2004-200518.9 mm*-2.3-7.011.3 no rain02.10-15.12.05 NSM-10115.12.05-02.02.06 1.4-1.2-4.54.8 NSM-10202.02-07.03.06 4.04-2.9-15.17.9 NSM-10307.03-10.04.06 1.7-0.4-5.1-2.2 NSM-10410-30.04.06 1.8-0.51.855.9 SummaryWinter 2004-20058.9 mm-1.7-8.25.1Summary * 2004-200627.8 mm*-2.0-7.58.8Ein-Netafim Spring GP: 29 35' 48.69"; 34 52' 51.52" NET-1 23.06.2004spring water-4.2-18.015.8 550 m asl NET-2 23.06.2004spring water-3.7-18.810.6Averag e -3.9-18.413.2$The summary of the rainfall data for Makhtesh-ha-Qatan includes the data from this study (winters 2004-2006) and from Kuperman (2005) (winters 2002-2004).#Neot-Smadar station was constructed on 02.11.2004. *In Neot-Smadar during the winter 2004-2005 the 18O and D were measured only on 12.8 mm from the 18.9 mm of total rainfall.
Table 3: D values and calculated 18O values of the Negev Desert speleothem fluid inclusions page 1CaveSample NumberAge range (ka) 18Ocalcite ( PDB) Standard deviation Dextractedwater ( SMOW) Standard deviation DFI ( SMOW) Numberof replicates Minimum temperature ( C) Maximum temperature ( C) Calculated minimum 18O ( SMOW) Calculated maximum 18O ( SMOW) Average d-excess ( 3.5) Tzavoa TZ-4-A2 124-126 -10.00.7 -80.0 7.1 -41.2 3 18 22 -10.35 -9.44 38 TZ-4-B1-2 128-131 -9.6 0.6 -71.0 6.9 -31.7 4 15 19 -10.56 -9.62 49 TZ-15-B-1 32-35 -5.420.09 -67.5 2.3 -31.1 6 12 16 -7.03 -6.06 21 TZ-15-F 61-64 -6.0 0.3 -64.014.6 -32.2 4 13 17 -7.35 -6.39 23 TZ-15-G 65-67 -4.7 0.4 -57.4 3.5 -23.0 4 12 16 -6.31 -5.34 24 TZ-22(1)-C2 157-161 -6.6 0.3 -69.4 1.5 -34.0 4 12 17 -8.21 -7.00 27 TZ-22(1)-F4+G1172-175 -9.8 0.5 -84.9 3.7 -48.7 3 12.516.5-11.46 -10.5039 TZ-22(1)-H 175-177 -9.9 0.7 -78.8 1.1 -43.6 4 13.517.5-11.32 -10.3743 TZ-22(1)-K2+K3195-198 -10.10.4 -92.1 1.5 -57.0 3 16.520.5-10.77 -9.84 25 TZ-22(1)-L 198-201 -9.6 0.4 -83.0 5.5 -43.8 3 16.520.5-10.27 -9.35 35 Hol-Zakh HZ-1-A 126-131 -9.9 0.3 -87.9 2.8 -53.4 4 19.523.5 -9.86 -8.96 22 HZ-1-B 131-132 -9.5 0.4 -92.0 6.1 -57.6 2 18 22 -9.77 -8.85 17 AshalimASH-33-A 117-124 -8.3 0.3 -80 6.3 -41.3 3 19 23 -8.29 -7.38 21 ASH-33-B123 126-130 -8.5 0.3 -87.2 5.2 -48.9 3 18 22 -8.77 -7.85 18 ASH-33-B4 130-132 -8.4 0.2 -73.7 5.6 -36.1 2 16 20 -9.14 -8.21 33 ASH-33-C 199-202 -7.7 0.1 -94.6 4.0 -55.4 5 17.521.5 -8.10 -7.18 6 ASH-33-D 203-209 -8.1 0.4 -90.3 3.7 -54.1 8 19 23 -8.09 -7.19 7 ASH-33-E1 210-213 -7.3 0.4 -84.5 2.2 -51.3 4 15 19 -8.26 -7.32 11 ASH-33-E2.1 218-220 -7.6 0.3 -90 4.7 -56.9 3 14 18 -8.78 -7.83 10
Table 3 page 2CaveSample NumberAge range (ka) 18Ocalcite ( PDB) Standard deviation Dextractedwater ( SMOW) Standard deviation DFI ( SMOW) Numberof replicates Minimum temperature ( C) Maximum temperature ( C) Calculated minimum 18O ( SMOW) Calculated maximum 18O ( SMOW) Average d-excess ( 3.5) AshalimASH-34-A 117-124 -8.0 0.5 -72.1 2.2 -34.4 3 19 23 -8.02 -7.12 26 ASH-34-B 127-133 -8.3 0.4 -80.5 4.1 -42.2 4 16 20 -9.07 -8.14 27 ASH-34-C 201-203 -7.4 0.1 -87.6 5.3 -53.0 4 17.521.5 -7.71 -6.79 5 ASH-15-A1Early-Middle Pleistocene-9.580.04 -73.1 4.9 -34.2 3 ? ? ? ? ? ASH-15-A2Early-Middle Pleistocene-9.9 0.1 -82.6 2.5 -43.4 3 ? ? ? ? ? ASH-15-B1Early-Middle Pleistocene-9.710.04 -85.4 4.7 -46.9 4 ? ? ? ? ? ASH-15-B2Early-Middle Pleistocene-9.9 0.1 -89.2 3.7 -51.1 5 ? ? ? ? ? ASH-15-C1Early-Middle Pleistocene-9.440.09 -90.6 4.3 -52.9 2 ? ? ? ? ? ASH-15-C2Early-Middle Pleistocene-9.400.09 -65.1 1.9 -26.0 4 ? ? ? ? ? ASH-15-C3Early-Middle Pleistocene-9.4 0.1 -77.8 8.9 -39.1 3 ? ? ? ? ? ASH-15-YEE3000-3200 -11.20.5 -87.3 3.0 -47.3 4 ? ? ? ? ? Ma'ale-ha-MeysharMMR-7(2)-A 120-128 -10.50.2 -92.0 7.1 -55.9 2 17.521.5-10.93 -10.0228 MMR-8 195-207 -10.20.3 -86.4 1.3 -48.3 3 18 22 -10.52 -9.61 32
Table 4: Sr concentrations and 87Sr/86Sr ratios: speleothems, host rocks, bulk soils and their silicat e fractions in the Jordan Valley and the Negev Desert. Page 1: speleothems younger than 550 ka.Cave/Site/Group Sample Age (ka) Sr concentration (ppm)87Sr/86Sr U-Th dateable speleothems: Ma'ale-Efra y im Cav e ME-3-E 134.702.80267 0.708400.00001ME-5-A1 49.211.35220 0.708380.00001ME-5-A2-1 70.631.62187 0.708460.00001ME-5-A2-2 71.921.04126 0.708460.00001ME-5-B3 157.123.67262 0.708400.00001ME-5-B2 158.483.55193 0.708370.00001ME-5-C1 211.516.64106 0.708280.00001ME-5-C3 219.435.44116 0.708180.00001ME-5-C2 221.199.82154 0.708330.00001ME-5-D1 288.034.3113 0.708420.00001ME-5-D3 300.834.8135 0.708150.00001ME-12-A1 16.400.28342 0.708350.00001ME-12-A2 24.790.46296 0.708460.00001ME-12-C 30.800.60208 0.708420.00001ME-12-F 31.600.80176 0.708410.00001ME-12-I2 35.801.00157 0.708440.00001ME-12-I3 38.101.00310 0.708430.00001ME-12-I4 38.700.40187 0.708520.00001ME-12-J1 48.001.40125 0.708470.00003ME-12-J4 53.901.00150 0.708480.00001ME-12-J5 59.102.00117 0.708310.00001ME-12-K3 61.602.20143 0.708470.00001ME-12-K5 63.001.80177 0.708530.00001ME-12-K7 71.882.53160 0.708510.00001Ma'ale-Dra g ot Cave s MD ( 1 ) -3-A 65.540.4512700.708310.00001MD ( 1 ) -8-A1 3.83 0.05152 0.708070.00002MD ( 1 ) -9-A 27.750.28243 0.708380.00001MD ( 1 ) -9-B 45.140.78175 0.708380.00002MD ( 1 ) -9-C 112.261.44184 0.708230.00003Tzavoa Cav e TZ-4-A2 124.971.6257 0.708340.00001TZ-4-B1,2 128.882.4399 0.708310.00002TZ-4-B2 136.832.95151 0.708310.00002TZ-12-A1 20.730.14527 0.708410.00001TZ-12-A2 25.950.10385 0.708460.00001TZ-14-A 14.020.21473 0.708450.00001TZ-14-B+C 16.040.21312 0.708510.00001TZ-14-E 17.210.16290 0.708500.00001TZ-14-F 17.430.21462 0.708490.00001TZ-14-G 18.110.19450 0.708460.00001TZ-14-H 18.920.23335 0.708480.00001TZ-14-J 18.990.11357 0.708460.00001TZ-14-K 19.950.28384 0.708460.00001
Table 4 Sr data, page 2: speleothems younger than 550 ka.Cave/Site/Group Sample Age (ka) Sr concentration (ppm)87Sr/86Sr TZ-15-B1 33.680.32310 0.708440.00001TZ-15-B2 37.990.29433 0.708410.00001TZ-15-C 48.040.27337 0.708390.00001TZ-15-D 53.640.25325 0.708400.00001TZ-15-E 61.090.64257 0.708440.00001TZ-15-F 62.680.82309 0.708460.00001TZ-15-G 66.260.64311 0.708440.00001TZ-15-I 70.490.32402 0.708380.00001TZ-15-K 69.840.67335 0.708400.00001TZ-15-L 72.630.40400 0.708390.00001TZ-15-M 74.320.48368 0.708430.00001TZ-15-N 77.940.79400 0.708410.00001TZ-21-A3 74.963.71280 0.708360.00001TZ-22 ( 1 ) -C2 157.351.48354 0.708510.00001TZ-22 ( 1 ) -F2+F3 174.462.03166 0.708290.00001TZ-22 ( 1 ) -H 175.112.64173 0.708270.00001TZ-22 ( 1 ) -K2+K3 197.003.0078 0.708170.00001TZ-22 ( 1 ) -L 199.003.0066 0.708230.00001Hol-Zakh Cav e HZ-1-A 128.391.32252 0.708220.00001HZ-1-B 131.661.67510 0.708300.00001HZ-3 ( 5 ) -A2 131.931.50300 0.708210.00001HZ-3 ( 5 ) -B 187.492.45557 0.708220.00001 A shalim Cav e A SH-33-C 200.786.47408 0.708280.00001 A SH-33-D 205.694.42357 0.708310.00001 A SH-33-E 213.193.71505 0.708240.00001 A SH-34A 120.661.79500 0.708410.00001 A SH-34-B 129.941.65395 0.708390.00001Even-Sid-Ramon mini-cave s ESID-2-A2 292.288.47180 0.708270.00001ESID-2-A1.1 87.731.38203 0.708270.00001ESID-7-A1 132.781.5658 0.708150.00001Ma'ale-ha-Me y shar Cav e MMR-7-A1 111.451.20113 0.708210.00001MMR-7 ( 2 ) -A+B 123.842.0088 0.708220.00001MMR-8 201.053.31111 0.708170.00001Shizafon Quarr y mini-cave s SHACH-1A 339.4512.0043 0.708040.00001Ma'ale-Ktora Crack s KTO-1 ( 1 ) -A2 319.298.51193 0.707940.00001
Table 4 Sr data, page 3: speleothems older than 550 ka and host rocks.Cave/Site/Group Sample Age (ka) Sr concentration (ppm)87Sr/86Sr Intermidiate Member ( IM ) A shalim Cav e A SH-15A 550-1269 42 0.708460.00001 A SH-15-B 550-1269 58 0.708390.00001 A SH-15-C 1269 13 110 0.708260.00001Even-Sid-Ramon mini-cave s ESID-2-C 550-2700? 45 0.708210.00001ESID-2-J 550-2700? 44 0.708100.00001Ma'ale-Ktora Crack s KTO ( 1 ) -1-E 550-2713 41 0.708290.00001Avera g e IM 0.708290.00012Basal Member ( BM ) A shalim Cav e A SH-15-D1 3045 21 260 0.707890.00001 A SH-15-D2 3045 21 252 0.707900.00001 A SH-15-E 3048 14 290 0.707840.00001 A SH-15-F 3048-3300? 280 0.707830.00001 A SH-15-G1 3048-3300? 318 0.707850.00001 A SH-15-G2 3048-3300? 320 0.707860.00001 A SH-15-I 3048-3300? 320 0.707880.00001 A SH-15-J 3048-3300? 326 0.707880.00001Ma'ale-Ktora Crack s KTO ( 1 ) -1-J 2713 60 180 0.707890.00001KTO ( 1 ) -1-K ( I ) 3317 63 100 0.707670.00001KTO ( 1 ) -1-K ( II ) 3317 63 104 0.707680.00001Avera g e BM 0.707830.00008Phreatic S p eleothems Ma'ale-ha-Me y shar Cav e MMR-12-B ? 160 0.707710.00001MMR-12-C ? 286 0.707730.00001MMR-12-E u p ? 300 0.707710.00001MMR-12-E u p ( 1-8 ) ? 299 0.707770.00001MMR-12-E low ( 9,10 ) ? 405 0.707770.00001Avera g e MMR-12 0.707740.00003Makhtesh-ha-Qatan MKTC-5-C 157.213.7821400.707860.00001Carbonate Host Rocks: Ma'ale-Efra y i m A MMINADAV-MECenomania n 80 0.707560.00001Hol-Zakh SHIVTA-H Z Turonian 99 0.707630.00001ovel y the cave bearin g formation s MENUCHA-ROAD224Senonia n 950 0.707590.00001 A shalim SHIVTA-CHALUKKIMTuronian 107 0.707500.00001 A shalim NEZER-CHALUKKIMTuronian 118 0.707710.00001ovel y the cave bearin g formation s MOR-NAFCHA Eocenia n 596 0.707780.00001Even-Sid-Ramo n TAMAR MR-1 Cenomania n 78 0.707530.00001Ma'ale-ha-Me y sha r TAMAR MMR Cenomania n 148 0.707640.00001Ma'ale-Ktora and Shizafo n GROFIT-KTO Turonian 200 0.707410.00001Avera g e carbonate rock 0.707590.00011
Table 4 Sr data, page 4: bulk soils and their silicate fractions.Soils above the caves (bulk): Sample Age (ka) Sr concentration (ppm)87Sr/86Sr Mildl y -Arid Zone: Terra-Rossa soil sam p les : Ma'ale-Dra g o t A MASA-TR-1 Presen t 161 0.709990.00001Ma'ale-Dra g o t A MASA-TR-2 Presen t 186 0.709810.00001Ma'ale-Efra y i m ME-TR Presen t 121 0.710990.00001Avera g e Terra-Rossa soil ( bulk ) 0.710260.00052 A rid Zone: Loess soil sam p les : Tzavoa, Mishash formatio n TZ-MISH-LOESS Presen t 374 0.708890.00001Tzavoa, Shivta formatio n TZ-SHIV-LOESS Presen t 314 0.709090.00002 A shalim A SH-LOESS Presen t 333 0.708810.00001Even-Sid-Ramon ( above ) ESID-LOESS Presen t 397 0.708790.00001Even-Sid-Ramon ( 1.5 km from ) RUCHOT-LOESS Presen t 408 0.708410.00001Avera g e Loess soil ( bulk ) 0.708800.00022H yp er-Arid Zone: Hamada soil sam p les : Ma'ale-ha-Me y sha r MMR-SOIL-1 Presen t 339 0.708780.00001Ma'ale-ha-Me y sha r MMR-SOIL-2 Presen t 399 0.708590.00001Shizafon SHACH-HAMAD A Presen t 382 0.708960.00001Avera g e Hamada soil ( bulk ) 0.708780.00015Paleosoils washed to the caves ( silicates ) : Even-Sid-Ramon mini-caves ESIDp aleoso l ? 0.709770.00002Ma'ale-ha-Me y sha r MMRp aleoso l ? 0.713970.00002Cave in road cut near Even-Sid-Ramo n STRAM-loes s ? 0.710270.00002Soils above the caves ( silicates ) : Mildl y -Arid Zone: Terra-Rossa soil sam p les : Jordan Valle y Ma'ale-Efra y i m ME-T R Presen t 0.712280.00001Northern Ne g ev Ma'ale-Dra g o t A MASA-TR-1 Presen t 0.710630.00001Ma'ale-Dra g o t A MASA-TR-2 Presen t 0.710640.00001 A rid Zone: Loess soil sam p les : Tzavoa, Mishash formatio n TZ-MISH-LOESS Presen t 0.710430.00002Tzavoa, Shivta formatio n TZ-SHIV-LOESS Presen t 0.710850.00001 A shalim A SH-LOESS Presen t 0.710090.00002Even-Sid-Ramon ( 1.5 km from ) RUCHOT-LOESS Presen t 0.710720.00001Even-Sid-Ramon ( above ) ESID-LOESS Presen t 0.711050.00001Avera g e soils northern and 0.710630.00028central Ne g ev ( silicates ) H yp er-Arid Zone: Hamada soil sam p les : Ma'ale-ha-Me y sha r MMR-SOIL-1 Presen t 0.711100.00001Ma'ale-ha-Me y sha r MMR-SOIL-2 Presen t 0.710800.00001Shizafon SHACH-HAMAD A Presen t 0.711500.00001Avera g e Hamada soil ( silicates ) 0.711130.00029Average Negev Soils (silicates): 0.710880.00025
NHP . â€“ , . 87Sr/86Sr " " 0.7082 0.7085 , . . , , 87Sr/86Sr 0.7084~. UTh . U-Pb 2.7-3.3 . . 1.3 ~ . 87Sr/86Sr -2.7 , . , . , , . " " . NHP-1 142-109 .
13C , , . NHP-3 , NHP-2 NHP-1 ) ( , D 220 . 18O , , . , , . , -200 " . . , , 300-350 " -50 " . , . 18O 132 126-127 , 200 , NHP-1 NHP-2 . . NHP , ) 51 cal/cm2 day ( , . . MIS-11 -
, 150-160 " , , . , " ." 117 96 , 92 85 , . , 280-300 " . 18O , . D . 13C C3+C4 , C3 , , . , . 25 " . , , , . ) Negev Humi d Periods NHP ( 350 310-350 ) NHP-4 ( , 290-310 ) NHP-3 ( , 190-220 ) NHP-2 ( , 109-142 ) NHP-1 .( NHP-4 . , . NHP-1 . ) NHP-3 ( ) MIS-8 .( NHP ,
, ) ( . : ) 350 " , ( ; 350 " 150 " -150 " 30 " . : , " " , , . ) , / .( U-Th . , 18O 13C D . Sr 87Sr/86Sr . , 200275 " 300-350 " . , MIS-6 MIS-5 -135 ) Termination-II ( 225-205 . MIS-5 , . 18O 13C 67 25 ) , ( , . . , " " , .
" " , . : ' , , ' , , , ' , , , ' GSI/14/08 , , "