Three hundred eighty thousand year long stable isotope and

. Stable isotope and faunal records from the central Red Sea show high-amplitude oscillations for the past 380,000 years. Positive (cid:127) (cid:127)80 anomalies indicate periods of significant salt buildup during periods of lowered sea level when water mass exchange with the Arabian Sea was reduced due to a reduced geometry of the Bab el Mandeb Strait. Salinities as high as 53%o and 55%o are inferred from pteropod and benthie foraminifera (cid:127)80, respectively, for the last glacial maximum. During this period all planktonic foraminifera vanished from this part of the Red Sea. Environmental conditions improved rapidly after 13 ka as salinities decreased due to rising sea level. The foraminiferal fauna started to reappear and was fully reestablished between 9 ka and 8 ka. Spectral analysis of the planktonic (cid:127)80 record documents highest variance in the orbital eccentricity, obliquity, and precession bands, indicating a dominant influence of climatically - driven sea level change on environmental conditions in the Red Sea. Variance in the precession band is enhance(cid:127) compared to the global mean marine climate record (SPECMAP), suggesting an additional influence of the Indian monsoon system on Red Sea climates.

which span 380 kyr and significantly extend the documented record of environmental change in the Red Sea. The records show that Red Sea climates responded not only to glacialinterglacial conditions but also to higher-frequency variations in the orbital precession band. As only fragmentary continental records are available from the surrounding landmasses, the records may also extend our understanding of the long-term climatic history of this region.

Material and Methods
We have generated a planktonic stable isotope record and records of foraminiferal abundance and faunal variability along a 21 m long Meteor Core Sta. 174/KL11 from the central Red Sea ( Table D. At core depth of 0.9-1.7 m no pl•to•c forami•fera were available for isotope measurements due to •e "apla•tonJc zone" which was caused by extremely high •inities during •e last [laci• m•mum •6• at approximately • 8 ka. for core KL11 was obtained through correlation of the isotope records with the global mean SPECMAP 5180 curve ( ( Figure   1, Table 1) [lmbrie et al., 1984]. The base of core KL11 reaches SPECMAP interglacial climatic event 11.2, a shortlived cold spell at approximately 375 ka (Figures 1 and 2). The age model suggests that mean sedimentation rates at the core site are 5.6 cm kyr. The isotope measurements were carried out at the C 14-Laboratory at Kiel University using a Finnigan MAT 251 mass spectrometer, which is linked on-line to a fully automated carbonate preparation device. The carbonate samples are dissolved in separate glass vials thus minimizing potential memory effects. Reproducibility was 0.08%0, and all isotope values are referred to the Peedee belemnite (PDB) scale. The planktonic foraminiferal counts were done on the size fraction > 150 •m and normalized to 1 g of sediment.

Stable Isotopes and Hydrography
The planktonic 5•SO profile along core KL11 shows highamplitude fluctuations with glacial-interglacial shifts up to 3.5-4.0%0 (Figure 2a). The coeval global glacial-interglacial variation of the world ocean's water masses was 1.  (Figure 2b) are coherent with the precession index pointing to low-latitude climatic forcing of the isotope-salinity anomalies, perhaps monsoon• variability. The orbital precession record has been lagged by 3 kyr to account for the phase lag which was determined by cross-spectral analysis (see Figure 5 and   Table  2). The record of planktonic 8•80 change in excess of the mean-ocean 6 w change is shown in Figure 2b. This record shows that the offset (A6•80) between planktonic •80 and mean-ocean 6 w was systematically larger during glacial periods, whereas during full -interglacial periods the offset was similar to or even slightly less than today. Superimposed on the long-term trend are short-lived anomalies which correlate with minima in the orbital precession index, reflecting the influence of precessional variability on low-latitude climate (see below).
The oxygen isotope composition of sea water in the Red Sea changes by 0.29%o for each 1%o change in salinity [Craig, 1966] In addition to the planktonic 6•80 we have generated a benthic 6180 record using a combined isotopic record of epibenthic Hanzawaia sp., Cibicides mabahethi in order to aFrom Vogelsang [1990]. However, as we will show below, the presence of distinct 5180 and faunal anomalies throughout our records implies that local climatic influences affected salinity in addition to the dominant sea level forcing.

Spectral Analysis
In order to compare statistically our planktonic 5180 record with the record of climate change we performed cross-spectral analysis using the record of 65øN July insolation [Berger and Loutre, 1991] (Table 2). This high coherency is expected because we have used the SPECMAP record to calibrate the age model for core KL11. The high coherencies and low phase angles confirm that the age model of core KL11 reproduces the orbital SPECMAP age model to within 0.4-1.3 ka ( Table 2) Spectral analysis of paleoclimatic records from the Arabian Sea has shown significant coherence between biological, chemical, and lithogenic tracers and insolation variance in the precession band with a phase lag of 122 ø corresponding to a time lag of 8 kyr [Clemens and Prell, 1990;Clemens et al., 1991]. This has been used as evidence for a strong contribution of monsoonal circulation to the paleoclimatic signals in response to radiative forcing, cross-equatorial heat transport and heat release from the Tibetan plateau [Clemens et al., 1991].

At present we have no independent tracer from core KL11
to test the monsoonal influence on the Red Sea area. Variance in the precession band is slightly enhanced in the planktonic 5180 record compared to the SPECMAP stack ( Figure 6). Since the KLll 5•80 record is phase-locked to global ice volume by using the SPECMAP age model, we cannot use this enhanced variance as independent evidence for monsoonal signals. However, in view of the presence of strong monsoonal signals in palcoclimatic records from the Arabian Sea [e.g., Prell, 1984;Kutzbach and Street-Perrot, 1985; Prell and van Carnpo, 1986; Anderson and Prell, 1992; Prell and Kutzbach, 1992] just outside the Red Sea, we speculate that the enhanced precessional 5180 variance in core KL11, in conjunction with faunal anomalies (see below), is a preliminary indication of palcomonsoon signals in the Red Sea.

Faunal Records of Paleohydrography and Paleoclimate
The abundance pattern of planktonic foraminifera along core KLll may be used as an independent indicator of palcosalinity, since most foraminiferal species have a narrow range of tolerance with respect to salinity. The abundance of planktonic foraminifera closely follows the 5•80-climate signal in that abundances decrease from full -interglacial to fullglacial conditions when salinities increase (Figure 7a). The distribution pattern shows short-term abundance maxima during high sea level stands, which correlate with low, closeto-modern salinity values as deduced from the planktonic 5•80 record, implying that rich planktonic communities dominated during humid climates when salinity was low due to increased moisture transport by the monsoonal winds. During periods of enhanced aridity and increased salinity the planktonic community was decreased and dominated by Globigerinoides ET AL.' GLOBAL SEA LEVEL AND RED SEA HYDROGRAPHY   [Milliman et al., 1969] were deposited and seem to be almost barren of planktonic foraminifera. It is only 1 cm above the LAL, or about 180 years after the aplanktonic zone [cf./llmogi-Labin et al., 1991] that the first planktonic foraminifera reappeared at site KL11 and another 1-4 cm, or up to 720 years aRer the end of the aplanktonic zone, that the planktonic foraminiferal community was almost fully reestablished. Apparently, environmental conditions improved very rapidly after 13 ka due to rising sea level. Thus the foraminiferal fauna appeared again, but was present still in low abundances. Between 9 and 8 ka, when climatic conditions were at an optimum, the total planktonic and benthic fauna was fully reestablished.
During the last deglaciation east African climates changed from arid to fully humid conditions in response to an orbitally -driven increase of monsoonal rainfall [Kutzbach and Guetter, 1986;Pachur and KrOpelin, 1987;COHMAP Members, 1988;Pachur et al., 1990;Bonnefille et al., 1990;Gasse et al., 1990;Lezine and Casanova, 1991]. Even though the Red Sea today receives little or no fresh water from surrounding land areas, we speculate that increased humidity may have lowered evaporation rates over the Red Sea, thereby enhancing the effects of sea level rise on lowering salinities. This would have contributed to the speed at which the faunal community reestablished itself during the last deglaciation. Together with the presence of strong precessional variance in the/5•80 record from core KL11, this leads us to conclude that during the past 380 kyr not only was the hydrography of the Red Sea controlled by global sea level variation but that variations in monsoonal strength also contributed to the fresh water balance by altering humidity and evaporation.

Summary
Core KL11 from the central Red Sea provides a continuous sedimentary record for the last 380 kyr showing glacialinterglacial 5•80 amplitudes that are up to 3 times higher than those of the mean-ocean 5•80 record. Correcting the planktonic 5•sO record for mean-ocean•S5 O, we derive salinifies of up to 53%o or even slightly higher for the LGM. These measured and calculated values are approximately 3%o higher than previously assumed values. During the glacial maximum, isotope stage 6 and 10 salinities were also significantly increased but lower than the maximum values inferred for the LGM. The planktonic fauna varied along with the salinity changes and was drastically reduced during glacial maxima due to the high salinity levels; during the LGM an aplanktonic zone developed owing to the highest salinities inferred for the entire 380 kyr period which apparently resulted in extremely hostile conditions. Global sea level variations are the main factor in controling the Red Sea's salinity. However, sea level cannot explain the full range of salinitiy changes deduced from our isotope data. An additional climatically -driven component is needed to add to the sealevel-driven salinity changes. This component is conceivably linked to variations in monsoonal strength. Support for this contention is provided by enhanced variance of planktonic 5•sO in the orbital precession band.