(This is a sample cover image for this issue. The actual cover is not yet available at this time.)

(This is a sample cover image for this issue. The actual cover is not yet available at this time.) This article appeared in a journal published by Elsevier. The attached copy is furnished to the author for internal non-commercial research and education use, including for instruction at the authors institution and sharing with colleagues. Other uses, including reproduction and distribution, or selling or licensing copies, or posting to personal, institutional or third party websites are prohibited. In most cases authors are permitted to post their version of the article (e.g. in Word or Tex form) to their personal website or institutional repository. Authors requiring further information regarding Elsevier s archiving and manuscript policies are encouraged to visit: http://www.elsevier.com/copyright

Quaternary Science Reviews 40 (2012) 21e29 Contents lists available at SciVerse ScienceDirect Quaternary Science Reviews journal homepage: www.elsevier.com/locate/quascirev Bipolar modulation of millennial-scale West African monsoon variability during the last glacial (75,000e25,000 years ago) Syee Weldeab Department of Earth Science, University of California, Santa Barbara, CA 93106-9630, USA article info abstract Article history: Received 8 September 2011 Received in revised form 15 February 2012 Accepted 23 February 2012 Available online xxx Keywords: Africa s paleoclimate West African monsoon Heinrich events Last glacial Bipolar modulation Time series of planktonic foraminiferal d 18 O and Ba/Ca-based sea surface salinity (SSS) estimates from the eastern Gulf of Guinea (eastern equatorial Atlantic) indicate changes in runoff that reflect variability of spatially integrated precipitation over the equatorial West African monsoon area. Millennial-scale and recurring runoff-induced SSS rises and declines in the range of 1.5 and 2 psu (practical salinity unit) reveal rapid oscillation between dry and wet phases. The timing of decreased runoff coincides with oscillation of DansgaardeOeschger stadials and Heinrich events, the most severe monsoon weakening correlating with the latter. d 18 O residual time series, derived by removing temperature, ice volume, and salinity components from the foraminiferal d 18 O record, suggest that weak monsoon precipitation during stadials and Heinrich events was accompanied by significant shifts in d 18 O precipitation toward higher values. Furthermore, d 18 O analysis of individual tests of Globigerinoides ruber pink (d 18 O indiv ) during dry episodes show a total range and variance of 2.3& and 0.25 (n ¼ 121), indicating that seasonal contrast of sea surface freshening was significantly reduced during Heinrich events relative to that of interstadials which show a total range and variance of 3.35& and 0.42 (n ¼ 140). On the basis of the timing and magnitude of changes in the monsoon record, it is evident that northern high latitude climate was the most dominant control on the West African monsoon variability. However, a southern high latitude imprint is also apparent during some episodes. This centennially resolved climate record demonstrates that the equatorial West African monsoon experienced profound changes in the amount, seasonal contrast, and moisture source of summer monsoon precipitation during the last glacial. The most plausible mechanism is a large-scale southward displacement of the monsoon trough, most likely initiated by large-scale reorganization of atmospheric circulation in response to northern high cooling and southern high latitude warmth. Ó 2012 Elsevier Ltd. All rights reserved. 1. Introduction Growing lines of evidence indicate that Africa s climate was sensitive to rapid and recurring ice sheet instabilities during the last glacial (demenocal et al., 2000; Johnson et al., 2002, 2011; Stager et al., 2002, 2011; Adegbie et al., 2003; Weldeab et al., 2005, 2007a,b, 2011; Shanahan et al., 2006; Brown et al., 2007; Scholz et al., 2007, 2011; Mulitza et al., 2008; Tjallingii et al., 2008; Castañeda et al., 2009; Niedermeyer et al., 2010; Sangen et al., 2011). Episodes of cold air temperature over Greenland, known as DansgaardeOeschger stadials (Dansgaard et al., 1993; Blunier and Brook, 2001), and meltwater flux into the North Atlantic during Heinrich events (Heinrich, 1988; Vidal et al., 1997; McManus et al., 2004) correlate with rapid decline in precipitation over much of Africa. Climate amelioration over the African E-mail address: weldeab@geol.ucsb.edu. continent coincides with relatively warm interstadial in northern high latitudes and vigorous Atlantic meridional ocean circulation (AMOC) (McManus et al., 2004), suggesting a strong atmospheric linkage. Key aspects concerning the mechanisms of climate linkage, timing, and magnitude of precipitation changes are, however, insufficiently understood. For instance, the lack of quantitative proxies for precipitation presents a serious obstacle to quantifying wetedry oscillations and assessing the severity of proposed mega droughts (Scholz et al., 2007; Mulitza et al., 2008; Stager et al., 2011). Age model uncertainty prevents the investigation of lead and lag patterns needed to infer whether tropical climate was merely responding to high latitude climate instabilities or was influencing the latter (Brown et al., 2007). Large-scale shifts in the intertropical convergence zone (ITCZ) are often invoked to explain the tight link between the African monsoon and northern high latitude climate. While the ITCZ hypothesis provides a viable mechanism for the northern and central parts of the African 0277-3791/$ e see front matter Ó 2012 Elsevier Ltd. All rights reserved. doi:10.1016/j.quascirev.2012.02.014

22 S. Weldeab / Quaternary Science Reviews 40 (2012) 21e29 monsoon, the extent of southward ITCZ displacement during Heinrich events and stadials is not well constrained. Furthermore, an alternative model such as proposed by Nicholson (2009) may be required to explain the climate mechanisms responsible for dry conditions in areas south of the mean summer position of the ITCZ. Nicholson (2009) showed that summer West African monsoon precipitation consists of a relatively weak rainfall zone at northernmost boundary of the monsoon area, which is associated with the ITCZ position, and an intense rain belt that is linked to ascending air masses sandwiched between the African Easterly Jet (AEJ) and Tropical Easterly Jet (TEJ) (Nicholson, 2009). Furthermore, modern data also show that during Sahel droughts the main rain belt retreated several hundred kilometers southwards, while the weak rainfall zone linked to the ITCZ remained over the northern Sahel (Nicholson, 2009). Most recent spatio-temporal reconstruction of West African monsoon (Weldeab et al., 2011) appears to support this model. In this contribution, we present a centennially resolved record of West African tropical hydrology, revealing rapid climate oscillations that were not discovered in a lower-resolution data set of the same core material (Weldeab et al., 2007a). The record covers the time interval between 75 and 25 thousand years before present (kyr, BP), and presents two independent proxies for hydrological changes whose differences in magnitude, timing, and pace provide crucial insights into regional climate patterns and their possible link to bipolar thermal oscillation. 2. Background Riverine discharge of large West African river systems into the eastern Gulf of Guinea provides a spatially integrated record of rainfall over the tropical and equatorial West African monsoon area (Fig. 1). Though runoff from the Niger River in the north and the Ogooué River in the south also contributes to the low sea surface salinity (SSS) in the Gulf of Guinea, the annual riverine discharge of approximately 86 km 3 from Sanaga, Nyong, and Ntem rivers, which drain the southern margin of the West African monsoon area, has the most impact on SSS (Antonov et al., 2010), isotope composition (LeGrande and Schmidt, 2006), and trace element budget of the seawater in the eastern Gulf of Guinea (Weldeab et al., 2007a). Dissolved Ba in runoff is, on average, 3 times higher (60 mg/l) than the Ba concentration in seawater (20 mg/l) (Turner et al., 1980). The concentration of dissolved Ba in sea surface water in coastal regions adjacent to large rivers is strongly controlled by the amount of runoff, and thus tightly anti-correlated with runoff-induced SSS changes as demonstrated by Ba and SSS measurements off Amazon, Congo, Ganges-Brahmaputra, and Mississippi rivers (Edmond et al., 1978; Carroll et al., 1993; Moore, 1997). Laboratory experiments show that the amount of Ba uptake into foraminiferal calcite is linearly correlated to the dissolved Ba concentration in seawater in which the foraminifers calcify, and the Ba incorporation is not influenced by changes in calcification temperature, ph, carbonate ion concentration, and total alkalinity (Lea and Spero, 1994; Hönisch et al., 2011). Thus, the temporal variability of runoff, reflecting precipitation of river basins, can be reconstructed by analyzing Ba/Ca and d 18 O in the calcite tests of low salinitytolerating, surface-dwelling planktonic foraminifer Globigerinoides ruber pink (Weldeab et al., 2007a). 3. Methods MD03-2707 core sediment was recovered from the eastern Gulf of Guinea (02 30.11 0 N, 09 23.68 0 E, 1295 m water depth) (Fig. 1) and presents a high-resolution sequence covering the last 155 kyr BP (Weldeab et al., 2007a). In this study, we focus on the Fig. 1. Map of West Africa showing basins of Niger, Sanaga, Nyong, Ntem, and Ogooué rivers, and the location of MD2707 in the eastern Gulf of Guinea. The inlet shows annual sea surface salinit (Antonov et al., 2010). The salinity map presents an extrapolation of large area, including the location of MD2707 for which no salinity measurements exist in the WA09.

S. Weldeab / Quaternary Science Reviews 40 (2012) 21e29 23 interval spanning from 75 to 25 kyr BP. For trace element analysis, 30e40 individuals of G. ruber variety pink were selected from the 250e300 mm fraction for each analysis. Shell samples were gently crushed and cleaned using the UCSB standard foraminifera cleaning procedure (Martin and Lea, 2002) that includes mechanical (sonication), oxidative, and reductive steps. Foraminiferal trace elements in dissolved samples were analyzed by the isotope dilution/internal standard method (Martin and Lea, 2002) using a Thermo Finnigan Element2 sector field ICPeMS. Analytical reproducibility of Ba/Ca, assessed by analyzing consistency standards matched in concentration to dissolved foraminifera solutions and analyzed over the course of entire study (578 samples), is estimated at 1.8%. To assess the removal of silicate and coating phases, Al/Ca, Fe/Ca, Mn/Ca, and other diagnostic elements (REEs/ Ca and U/Ca) were also analyzed. Out of 578 trace element measurements, 12 (2%) samples were disregarded due to anomalously high Ba/Ca values, most of which were also accompanied by either high Al/Ca, Fe/Ca, or Mn/Ca ratios. We established a record of Ba/Ca-based SSS estimates using the foraminiferal Ba/ Ca time series, an estimate of practical partition coefficient of D Ba ¼ 0.149 0.05 [Ba/Ca foram shell ¼ D * Ba Ba/Ca seawater ] (Lea and Spero, 1994; Hönisch et al., 2011), and a linear relationship between modern Ba/Ca seawater and SSS off Congo River [SSS (psu) ¼ 1.1 * Ba/Ca seawater þ 37.45 (1.06); r 2 ¼ 0.98] (Edmond et al., 1978). Error propagation in the SSS estimate results to an uncertainty of 1.2 practical salinity unit (psu). Analyses of d 18 O in bulk (250e300 mm) and individual (300e400 mm) G. ruber pink tests from MD03-2707 down core samples were made with Thermo MAT 253 mass spectrometer at the author s stable isotope lab, UCSB. The mass spectrometer is coupled online to a Kiel Carbonate Device (Type IV) for automated CO 2 preparation. Samples were reacted by individual 100% phosphoric acid addition. Results were corrected using NBS19 standard and are reported on the Peedee Belemnite (PDB) scale. d 18 O estimate of local seawater (d 18 O seawater ) is obtained by removing the temperature and ice volume components from the d 18 O G.ruber record. The temperature component is reconstructed using Mg/Ca-based calcification temperature of G. ruber pink (Weldeab, under review) and the Ted 18 O calcite ed 18 O seawater equationt ( C) ¼ (16.5 0.2) (4.8 0.16) * (d 18 O calcite d 18 O seawater )(Bemis et al., 1998). The ice volume component is obtained using high-resolution sea-level record (Siddall et al., 2003) that is digitally resampled and converted to a global d 18 O seawater record assuming d 18 O seawater changes of 0.0085& per meter eustatic sea-level change (Waelbroeck, 2002). Uncertainty in the reconstructed d 18 O seawater record arises due to errors in the d 18 O G.ruber measurements (0.07&), sea-level reconstruction (12 m z 0.102&), conversion of d 18 O G.ruber into d 18 O seawater (0.16&), and errors in the Mg/Ca-based estimate of calcification temperature (1.2 C z 0.25&), resulting to an error estimate of 0.32&. In a further step, we applied the Ba/Ca-based SSS record and modern SSSed 18 O seawater relationship (Schmidt et al., 1999) [d 18 O seawater ¼ 0.1713 * SSS 5.32, r 2 ¼ 0.8, n ¼ 477] to separate d 18 O seawater variability related to runoff-induced SSS changes (d 18 O SSSchanges) from d 18 O seawater related to changes in the isotope composition of precipitation and other processes (d 18 O residual ¼ d 18 O seawater d 18 O SSS-changes ). The modern d 18 O seawater and SSS data, used to establish the SSSed 18 O seawater relationship, are from tropical Atlantic (8.5 Ee80 W and 28.8 Ne27 S, water depth: 0e50 m, SSS: 20e40 psu) (Schmidt et al., 1999). The estimate of error propagation in the reconstructed d 18 O residual time series accounts for 0.45&. The age model for the MD03-2707 record has been developed by aligning the planktonic foraminiferal d 18 O record in MD03-2707 to the d 18 O record of GISP2 ice core (Blunier and Brook, 2001) (Fig. 2). The rationale for this approach is based on (i) the observation that changes in the 14 C-dated d 18 O MD03-2707 record (Weldeab et al., 2007a,b) are synchronous, within age model uncertainty, with air temperature variation over Greenland, as reflected in d 18 O of ice core records (Blunier and Brook, 2001); and (ii) based on the assumption that this linkage holds throughout the millennial-scale last glacial climate instabilities. A constant deposition rate based on the tie points, lends support to the above assumption. Fig. 2. Time series of d 18 O (B) and Ba/Ca (C) analyzed in Globigerinoides ruber pink (250e300 mm) from MD2707 core sediment plotted versus calendar age model. The age model of MD2707 section shown here has been established by tuning the d 18 O MD2707 record to the GISP2 d 18 O record (A) (Blunier and Brook, 2001). Black triangles along the d 18 O record show tie points of the age model. Numbers above the GISP2 record denote to DansgaardeOeschger interstadials and gray bars indicate time interval of Heinrich events and stadials.

24 S. Weldeab / Quaternary Science Reviews 40 (2012) 21e29 4. Results We generated 590 new d 18 O and 320 new Ba/Ca measurements on bulk G. ruber pink (250e300 mm) at 2 cm interval. This sampling interval corresponds to a time resolution of 40 and 250 years with 270 Ba/Ca measurements based upon a previous study (Weldeab et al., 2007a). We also analyzed 489 new d 18 O measurements on individual G. ruber pink (300e400 mm) from nine samples within Heinrich event 5, interstadial 12, and stadial 11. The new highresolution d 18 O calcite and Ba/Ca data sets, and the derived d 18 O seawater and d 18 O residual time series provide a centennial- to submillennial-scale record of changes in West African monsoon (Fig. 2). 5. Discussion 5.1. Millennial-scale oscillation of monsoon precipitation The foraminiferal d 18 O and Ba/Ca records over the last glacial reveal millennial-scale shifts between relatively weak and strong monsoon episodes (Fig. 2). The d 18 O seawater record, obtained by removing the temperature and ice volume components from d 18 O in G. ruber, shows numerous millennial-scale oscillations (Fig. 3B). Time intervals with the highest d 18 O seawater values coincide with the timing and duration of Heinrich events H3, H4, H5, H5a, and H6. Furthermore, every stadial in the GISP2 ice core record has also its counterpart in the MD03-2707 sequence, with relatively high d 18 O seawater values. Conversely, episodes synchronous with northern high latitude interstadials are characterized by markedly light d 18 O seawater values. The amplitude of these negative and positive d 18 O seawater excursions declines significantly toward full glacial conditions; this is most evident during interstadials 3, 4, and H6 (Fig. 3). Due to the proximity of the core site to the river mouths and the strong impact of runoff on the salinity and isotope composition of seawater in the eastern Gulf of Guinea (Fig. 1) (LeGrande and Schmidt, 2006), temporal oscillations of reconstructed d 18 O seawater in MD03-2707 most likely reflect changes in the amount of runoff and/or variation of oxygen isotope composition of precipitation over the riverine basins. This view is fully corroborated by the Ba/ Ca-based monsoon record, which provides independent insights into variations in runoff and, therefore, changes in precipitation over the river basins (see Background and Methods section). Down core Ba/Ca values in G. ruber pink vary between 0.75 and 1.5 mmol/ mol. Foraminiferal Ba/Ca-based SSS estimates (see Methods section) suggest that runoff-induced SSS varied between 32 and 27 psu. For comparison, modern annual SSS over the core site accounts for 29.5 psu, seasonally varying between 28.5 and 31.5 psu (Antonov et al., 2010). Episodes of elevated Ba/Ca-based SSS estimates coincide and overlap with the duration of DOstadials and Heinrich events, indicating reduced amounts of runoff and weak monsoon precipitation at those times. In the Ba/Ca record some of the short-lived interstadials are present in muted amplitude (D/O 10) or completely absent (D/O 9, and 11). Because the mean oceanic residence time of Ba (w10,000 years) (Turner et al., 1980) is significantly longer than the duration of these brief D/O interstadials and foraminiferal Ba/Ca is sensitive to brief Fig. 3. 5-point running average of d 18 O seawater (B) and Ba/Ca-based sea surface salinity (SSS) estimates (C) compared to the d 18 O ice records from Greenland (GISP2) (A) and Antarctica (Byrd) (E) based on common methane synchronized age model (Blunier and Brook, 2001). (D) Variation of solar insolation over 5 N (July, 21) is also shown (Berger and Loutre, 1991; Paillard et al., 1996). Uncertainties in the d 18 O seawater and Ba/Ca-based SSS estimates account for 0.32& and 1.2 psu, respectively.

S. Weldeab / Quaternary Science Reviews 40 (2012) 21e29 25 climate excursions such as 8.2 kyr events (Weldeab et al., 2007b), it is not clear why the imprint of D/O 9 and 11 is not recorded in the Ba/Ca series. Regardless, the two independent runoff proxies show that monsoon precipitation in equatorial Africa was marked by numerous millennial-scale oscillations. The overall patterns of last glacial West African Monsoon variability that emerge from d 18 O and Ba/Ca records are consistent with other components of the global monsoon as manifested in the d 18 O records of Arabian (Burns et al., 2003; Fleitmann et al., 2011), Indian (Kudrass et al., 2001; Sinha et al., 2005; Cai et al., 2006), and east Asian (Wang et al., 2001) monsoons, suggesting that large-scale atmospheric reorganization occurred in the low latitudes in tandem with ice sheet instabilities in both northern and southern high latitudes (Dansgaard et al., 1993; Blunier and Brook, 2001). However, the lack of a proxy that enables the separation of the d 18 O calcite records into the amount effect of precipitation and d 18 O precipitation changes related to source, pathway, and recycling of moisture prevents nuanced insights into regional climate variation, as suggested by modeling studies (Lewis et al., 2010; Pausata et al., 2011). In the following chapters, we use Ba/Ca-based runoffinduced SSS variation and its temporal relationship to the reconstructed d 18 O residual to assess non-mount changes in d 18 O calcites and its implication for understanding the severity of past monsoon declines. 5.2. Phase relationship between the Ba/Ca and d 18 O records Leads and lags between reconstructed changes in the foraminiferal Ba/Ca and d 18 O are independent from of age model error because the records were generated from aliquots of the same G. ruber sample. Because the MD03-2707 d 18 O record is tied to the GISP2 d 18 O age model (see Methods section), deviations in the timing, excursion, and pace of changes in the Ba/Ca record from those of the d 18 O record are real climate signals and indicate modulation of Ba/Ca by processes and influences other than northern high climate variability. Temporal variability of Ba/Ca in MD2707 is assumed to reflect changes in runoff. However, exposure of large shelf areas to erosion during sea-level low stands (Giresse et al., 1995; Ngueutchoua and Giresse, 2010) and reduced vegetation cover in the catchments during relatively dry phases (Dupont et al., 2000) may contribute to high rate of dissolved Ba input (Fleitmann et al., 2007). Considering age model uncertainties, a comparison of the Ba/Ca record with the sea-level record (Siddall et al., 2003) does not reveal any coherent patterns. A large sea-level drop and reduced vegetation cover over the river basins during marine isotope stage 4 (Dupont et al., 2000) is paralleled by low Ba/ Ca, suggesting that the timing, pace, and magnitude of Ba/Ca are largely determined by variations in runoff. A comparison between the Ba/Ca and d 18 O seawater records reveals differences in the timing, pace, and magnitude of climate excursions (Fig. 3). The onset of West African monsoon weakening at the end of interstadials 21,13, and 8, as suggested by the Ba/Ca record, predates the timing of changes manifested in the MD2707 d 18 O record. The termination of stadials 20, 7, H5, and H4 in the Ba/Ca record also leads those of the d 18 O record. We suggest that the early and gradual onset of Ba/Ca decline reflects a gradual climate deterioration in equatorial Africa predating the abrupt and severe decline of precipitation documented in the d 18 O record. If our interpretation is correct, the timing and pace of changes in the Ba/Ca record suggests that millennial-scale West African monsoon variability was also modulated by tropical and southern high latitude climate processes. The degree of modulation by these climate processes throughout the last glacial must have varied, as its influence is not discerned in every monsoon shift. Examples that may indicate an influence of southern high latitude in West Africa monsoon, as recorded in the Ba/Ca, include D/O 14. The onset of an abrupt decrease in Ba/Ca-based SSS estimates during D/O 14 lags by w450 years air temperature rise over Greenland and coincides with a precipitous decline in air temperature over Antarctica as revealed in the Byrd d 18 O record (Fig. 3E). Gradual increase of SSS toward the end of D/O 14 is also paralleled by a gradual rise in air temperature over Antarctica and leads air temperature cooling over Greenland (Fig. 3). Similar patterns can be observed during D/O 16 and 8 (Fig. 3). One possible mechanism for the linkage between West African Monsoon and Southern high latitude climate is provided by model studies (Broccoli et al., 2006; Chang et al., 2008). These models suggest that relatively warm (cool) southern high latitudes instigate a southward (northward) displacement of the ITCZ due to changes in low and mid latitude heat exchange, thereby weakening (strengthening) monsoon precipitation. A further possibility, as suggested by declining Ba/Ca-based SSS, is that the onset of climate amelioration in equatorial West Africa occurred concomitant with warmest phases over Antarctica and leads the termination of H6, H5, H4, and stadial 7 (Fig. 3). This is (seemingly) in contradiction with the aforementioned mechanism. Though somewhat speculative, one way to reconcile this observation with the concept of ITCZ displacement is that the southernmost ITCZ displacement and its spatially limited seasonal migration is instigated by the peak warmth in the southern high latitude and in the eastern Gulf of Guinea (Weldeab, in press), leading to more annually average precipitation in the coastal area. Indeed, a modeling study simulating Heinrich event-like conditions predicts such rainfall focusing over coastal areas of the Gulf of Guinea (Chang et al., 2008). Overall, while not every feature in the Ba/CaeSSS estimates is explainable on the basis of interhemispheric modulation, invoking southern hemisphere influence provides an additional tool to better understand the timing, pace, and magnitude of African monsoon precipitation during the last glacial. Our finding lends strong support to the emerging lines of evidence that suggest that records of millennial-scale last glacial climate variability in equatorial Africa (Mulitza and Rühlemann, 2000; Brown et al., 2007), Indian Summer Monsoon (Cai et al., 2006), and East Asian monsoon (Cai et al., 2006; Rohling et al., 2009) harbor a significant climate imprint of southern high latitude origin. 5.3. Changes in the d 18 O of monsoon precipitation The reconstructed d 18 O seawater time series (Fig. 4) represents a composite of signatures related to the amount of riverine discharge and changes in isotope composition of the runoff (Fig. 4). Using Ba/Ca-based SSS estimates and modern SSSed 18 O seawater relationship (see Methods section) and assuming that this relationship does not significantly vary throughout the investigated time interval (Schmidt, 1999; Rohling, 2007), we removed the salinity component from the d 18 O seawater record and derived the residual oxygen isotope composition (d 18 O residual ) (see Methods section). d 18 O residual is assumed to reflect oxygen isotope changes of seawater in response to variability of d 18 O in West African precipitation. We note that this assumption presents an oversimplification of processes and components that shape the d 18 O residual (Fig. 4). Nonetheless, we argue that at this particular core site changes in the oxygen isotope composition of precipitation presents the dominant factor in determining large-scale excursions in d 18 O residual. If this interpretation is correct, the d 18 O residual record suggests that the severely weakened monsoon precipitation during Heinrich events and stadials was significantly enriched in heavy oxygen isotope. The enrichment is most likely related to multipleinfluences such as changes in moisture source, the degree of rainout, atmospheric mixing, evapo-precipitation, recycling of moisture, and/or changes in precipitation seasonality (LeGrande and

26 S. Weldeab / Quaternary Science Reviews 40 (2012) 21e29 Fig. 4. 5-point running average of d 18 O residual as derived by removing the temperature, ice volume, and salinity components from the d 18 O record analyzed in Globigerinoides ruber pink plotted versus calendar age model. Marked increase in d 18 O residual during Heinrich events and stadial are indicated by gray shaded areas. Uncertainty in the d 18 O residual accounts for 0.45&. Schmidt, 2009; Lewis et al., 2010). One modeling study simulating Heinrich event-like conditions suggested that the d 18 O of precipitation in all northern hemisphere monsoon systems experienced a significant shift toward higher values (Lewis et al., 2010). A common dominant factor for the d 18 O shift in all monsoon systems is most likely the weakening of monsoon coupled with large-scale ITCZ shift. In addition, regional characteristics of monsoon such as proximity to moisture source, orographic barriers, and the relative contribution of winter and summer monsoon most likely shaped the magnitude of the d 18 O precipitation shift. Because the main riverine catchments are located in coastal areas, changes in the amount effect and increase in the relative contribution of winter precipitation are the most plausible mechanisms explaining the inferred shift in the d 18 O precipitation. Invoking an increase in the relative contribution of winter rainfall is strongly supported by the data and discussion in 4.3. Overall the data are consistent with model results that predict a shift of d 18 O toward heavier values in monsoon precipitation during Heinrich events (Lewis et al., 2010) and suggest a fundamental atmospheric reorganization occurred over the monsoon area, affecting not only the amount (Fig. 3B and C) but also the isotope signature of precipitation (Fig. 4). This finding has an important implication for assessing the severity of precipitation decline based only on d 18 O of marine and terrestrial archives. 5.4. Contrast in seasonality and shift of rainfall zone over the monsoon area Large-scale displacement of the mean summer position of the ITCZ is often invoked to explaining past variation of proxies of monsoonal hydrology. This concept is broadly supported by modeling studies focusing on orbital and millennial time scales (Kutzbach and Liu, 1997; Chiang et al., 2003; Timmermann et al., 2005; Broccoli et al., 2006; Liu et al., 2006; Chang et al., 2008; Timm et al., 2010). To assess changes in sea surface conditions that would relate to southward displacement of the ITCZ over the Gulf of Guinea during Heinrich events, we analyzed d 18 O in individual tests of G. ruber pink (d 18 O indiv ) from narrowly spaced time intervals within Heinrich event 5 (HE 5), stadial 11 (preceding D/O 11) and interstadial 12 (Fig. 4). The life span of G. ruber pink is approximately 2 weeks (Hemleben et al., 1989; Spero, 1998) and the variance and range of d 18 O indiv measurements from samples with multi-decadal resolution reflects sub-monthly-to-seasonal contrasts in mixed layer depth (Spero, 1998; Koutavas et al., 2006). In the eastern Gulf of Guinea, the modern seasonality of SSS variation is largely determined by runoff related to monsoon precipitation and current mixing. The ITCZ remains throughout the year north of the eastewest trending coastal zone and SE trade winds prevail over the eastern Gulf of Guinea. Significant changes in the monsoon strength and southward shift of NE trade winds over the Gulf of Guinea are expected to leave a strong imprint in the variance and range of d 18 O indiv of G. ruber pink. Analysis of d 18 O indiv from four narrowly spaced samples within stadial 11 show mean d 18 O values of 1.36& 0.52(1s) (n ¼ 35), 1.04& 0.37(1s) (n ¼ 59), 1.3& 0.59(1s) (n ¼ 64), and 1.22& 0.39(1s) (n ¼ 71), with a total range and pooled variance of 2.53& and 0.23 (n ¼ 228), respectively (Fig. 5A). 121 d 18 O indiv measurements from two samples within HE 5 show a narrow total range of 2.33&, relatively heavy mean value of 1.06& 0.62(1s), and a pooled variance of 0.25. In contrast, d 18 O indiv measurements from 3 samples within interstadial 12 yield mean values of 1.88& 0.69 (1s) (n ¼ 39), 1.85& 0.63 (1s) (n ¼ 48), and 1.75& 0.64 (1s)(n ¼ 53), with a total range and pooled variance of 3.35& and 0.42 (n ¼ 140), respectively. The pooled variance and total range in HE 5 (stadial 11) samples is lower by 41% (45%) and 30% (25%) as compared to that of interstadial 12 samples, indicating that during stadials and Heinrich events sub-monthly, intra- and interseasonal contrast of sea surface conditions were significantly reduced relative to those of interstadials. Assuming that the d 18 C indiv record can be used as indicator of upwelling (Koutavas et al., 2006) and is not overprinted by other unknown processes, the poor correlation between the d 18 C indiv and d 18 O indiv values during interstadial 12 (r 2 ¼ 0.0005, n ¼ 140), stadial 11 (r 2 ¼ 0.04, n ¼ 228), and HE 5 (r 2 ¼ 0.001, n ¼ 121) indicates that no seasonal upwelling was established during these time intervals (Fig. 5B). This is supported by SST changes of w1 C(Weldeab, in press) and therefore, could explain at most, a d 18 O change of 0.21& (Bemis et al., 1998). The dominant source for a total range of 3.35&, 2.54&, and 2.3& is therefore most likely runoff-induced SSS changes, with 0.27& in d 18 O per salinity unit (Fairbank et al., 1982). During interstadial 12, the intra- and seasonal contrast was most pronounced, suggesting a strong summer monsoon precipitation and possibly large spatial extent of monsoon area. In contrast, during HE 5 and, to lesser degree, during stadial 11 monsoon strength was severely weakened (HE5: mean d 18 O indiv: 1.06& 0.62) and the seasonal contrast in runoff-induced SSS change was markedly reduced as well (total range: 2.33& and variance 0.25). It is also noteworthy that the difference between the seasonal contrast in sea surface conditions during HE 5 and interstadial 12 arises mainly due to changes in the wet season (Fig. 5A). This observation can be attributed to a weak summer monsoon precipitation (Fig. 3) and shift in the d 18 O signature of monsoon precipitation toward high values (Fig. 4). The amount of winter

S. Weldeab / Quaternary Science Reviews 40 (2012) 21e29 27 Fig. 5. (A) d 18 O analyzed in individual test (300e400 mm) of Globigerinoides ruber pink (d 18 O indiv) plotted versus calendar age model of MD2707. Blue open circles, red open circle, and vertical balk show individual measurements from a sample, mean value, and standard deviation (1s). Bold black line indicates 5-point running average of d 18 O measurement in bulk tests (250e300mm) of Globigerinoides ruber pink. (B) d 18 O indiv plotted versus d 13 C indiv analyzed in tests of Globigerinoides ruber pink in samples from HE 5, DO 12, and stadial 11 (preceding interstadial 11). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) precipitation during HE 5 was (i) either not significantly reduced (relative to that of the interstadial winter, see Fig. 5A), resulting in an increase in the relative contribution of winter precipitation to annual rainfall, (ii) or winter precipitation was reduced as well but was accompanied by lighter values in the d 18 O of precipitation. A modeling study simulating Heinrich event-like conditions suggests an increase in the relative contribution of winter precipitation in Gulf of Guinea coastal area supporting the first alternative explanation (Chang et al., 2008). Whether the observation we made in HE 5, interstadial 12, and stadial 11 can be extrapolated to other Heinrich events and stadials/interstadials is subject to speculation. We argue, however, that Heinrich events and stadials having similar magnitudes of change and duration to those investigated may have experienced comparable seasonal contrasts in sea surface conditions. The extent of southward ITCZ displacement and shift in the amount of seasonal precipitation may have played a crucial role in modulating moisture source and rainout that affect the oxygen isotope composition of precipitation. Because of the strong link that existed between northern high latitude climate and the equatorial West African monsoon (Fig. 3), it is instructive to look at proposed seasonality of northern high latitude climate during the last glacial. While the African monsoon is a summer phenomenon, temperature drops in the range of 15 C in the northern high latitude during the last glacial (Cuffey and Clow, 1997) have been interpreted to reflecting mainly extreme winter conditions (Atkinson et al., 1987; Denton et al., 2005; Kelly et al., 2008). Model (Barnett et al., 1988) and observational data (Kripalani et al., 2003; Zhao and Moore, 2004) indicate that winter and spring snow depth on the Tibetan Plateau has a strong effect on the ensuing Indian Summer Monsoon precipitation and Tropical Easterly Jet that is considered as critical for the development of West African monsoon precipitation (Nicholson, 2009). By analogy, the above described linkages may provide a viable mechanism for explaining the close co-variation of monsoon and high latitude climate during the last glacial, as suggested by previous studies (Denton et al., 2005; Weldeab et al., 2011). With regard to the extent of ITCZ displacement, the markedly poor correlation between d 18 C indiv and d 18 O indiv (Fig. 5B) and the absence of large SST changes (Weldeab, in press) indicate that during stadials and Heinrich events no significant changes in the direction of seasonal trade winds occurred over the Gulf of Guinea. This observation together with evidence of weak monsoon suggests that during Heinrich events/stadials seasonal migration of the ITCZ was strongly contracted. The southernmost position of the ITCZ, however, remained north of the eastewest trending coastal area. This interpretation is consistent with previous findings suggesting that during the Younger Dryas the mean summer position ITCZ remained north of the Sanaga and Nyong basins (Weldeab et al., 2011). 6. Summary and conclusion The results and discussion of this study provide detailed account of past West African monsoon variability as summarized below. 1. This study reveals a detailed record of the impact of southern and northern high latitude climates on the hydrology of equatorial West Africa throughout the last glacial. Heinrich events and stadials are paralleled by rapid drop in monsoon precipitation. The northern high latitude climate exerted the most dominant control on the monsoon precipitation. However, thermal fluctuation of the southern high latitudes also influenced the onset and pace of some events, as revealed in the local salinity proxy (Ba/Ca). 2. During Heinrich events we note a marked shift in oxygen isotope ratio ( 18 O/ 16 O) of precipitation toward higher values, indicating changes in the moisture source, evapo-precipitation, seasonality, and/or amount effect. This finding has an important implication for assessing the severity of precipitation changes based solely on the d 18 O record. 3. d 18 O analysis of individual tests (d 18 O indiv ) of surface-dwelling planktonic foraminifers (G. ruber pink) suggests that during Heinrich events the seasonal contrast of sea surface conditions was strongly weakened relative to that of interstadials. Furthermore, the distribution of d 18 O indiv during HE 5, stadial, and interstadial 12 suggests that the weakening of seasonal contrast during HE 5 was mainly due to changes in summer monsoon precipitation that results in an increase in the relative contribition of winter rainfall to the annual precipitation. 4. Poor correlation between d 18 O indiv and d 13 C indiv indicates that during Heinrich events and stadials no significant changes in the direction of trade winds occurred over eastewest trending coastal area. This result is consistent with the idea that the

28 S. Weldeab / Quaternary Science Reviews 40 (2012) 21e29 mean summer position of ITCZ remained over the eastewest trending coastal area of Nigeria and Cameroon. The most probable mechanism that can account for the millennial-scale variability of equatorial West African monsoon precipitation and its link to high latitude climate perturbations is a large-scale shift of mean summer position of the ITCZ. Cooling of summer temperature over Greenland in the range of 3.9e6.6 C, as suggested by Kelly et al. (2008) for the Younger Dryas, may have instigated a southward displacement of wind systems including the NE trade winds. Based on the growing lines of evidence that millennial-scale drops in temperature over the northern high latitude during Heinrich events and stadials reflect largely extreme winter cooling (Denton et al., 2005; Kelly et al., 2008), the climate link between northern high latitude and West African monsoon could have been established via winter preconditioning that affect low latitude atmospheric circulation in the ensuing summer. Snow depth over the Tibetan Plateau is considered to have impacted not only the Indian summer monsoon (Barnett et al., 1988; Kripalani et al., 2003; Zhao and Moore, 2004) but also the African monsoon via the Tropical Easterly Jet (Barnett et al., 1988; Nicholson, 2009). Though the northern high latitude climate exerted the most dominant control on equatorial West Africa hydrology, we identify pacing and timing in the monsoon record that appear to be linked to southern high latitude thermal oscillation, lending support to previous proxy-based (Mulitza and Rühlemann, 2000; Brown et al., 2007) and modeling (Broccoli et al., 2006) works. Regardless of the details of the linking mechanism, this study demonstrates that millennial-scale interhemispheric climate oscillations had a strong impact on the strength, moisture source, and seasonal contrast of West African monsoon precipitation. Acknowledgment We thank Dorothy Pak, Alex Simms, Jim Kennett, and David Lea for discussion and suggestion that helps to improve a previous version of this paper. We also thank Georges Paradis for ICPeMS operation. Insightful and constructive comments of two anonymous reviewers are highly acknowledged. I am grateful to UCSB for the generous start-up package that facilitated the analysis of the data presented here. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.quascirev.2012.02.014. References Adegbie, A.T., Schneider, R., Röhl, U., Wefer, G., 2003. Glacial millennial-scale fluctuation in central African precipitation recorded in terrigenous sediment supply and fresh water signals offshore Cameroon. Palaeogeography Palaeoclimatology Palaeoecology 197, 323e333. Antonov, J.I., et al., 2010. World Ocean Atlas 2009. In: Levitus, S. (Ed.), NOAA Atlas NESDIS 69. Salinity, vol. 2. U.S. Government Printing Office, Washington, D.C., 184 pp. Atkinson, T.C., Briffa, K.R., Coope, G.R., 1987. Seasonal temperatures in Britain during the past 22,000 years, reconstructed using beetle remains. Nature 325 (6105), 587e592. Barnett, T.P., Duemenil, L., Schlese, U., Roeckner, E., 1988. The effect of Eurasian snow cover on global climate. Science 239, 504e507. Bemis, B.E., Spero, H., Bijma, J., Lea, D.W., 1998. Reevaluation of oxygen isotope composition of planktonic foraminifera: experimental results and revised paleotemperature equations. Paleoceanography 13, 150e160. Berger, A., Loutre, M.F., 1991. Insolation values for the climate of the last 10 million years. Quaternary Science Reviews 10 (4), 297e317. Blunier, T., Brook, E.J., 2001. Timing of millennial-scale climate change in Antarctica and Greenland during the last glacial period. Science 291, 109e112. Broccoli, A.J., Dahl, K.A., Stouffer, R.J., 2006. Response of the ITCZ to Northern Hemisphere cooling. Geophysical Research Letters 33 (1), L01702. Brown, E.T., Johnson, T.C., Scholz, C.A., Cohen, A.S., King, J.W., 2007. Abrupt change in tropical African climate linked to the bipolar seesaw over the past 55,000 years. Geophysical Research Letters 34, L20702. Burns, S.J., Fleitmann, D., Matter, A., Kramers, J., Al-Subbary, A.A., 2003. Indian Ocean Climate and an absolute chronology over Dansgaard/Oeschger Events 9 to 13. Science 301, 1365e1367. Cai, Y., et al., 2006. High-resolution absolute-dated Indian Monsoon record between 53 and 36 ka from Xiaobailong Cave, southwestern China. Geology 34, 621e624. Carroll, J., Falkner, k.k., Brown, E.T., Moore, W.S., 1993. The role of the Ganges- Brahmaputra mixing zone in supplying barium and 226Ra to the Bay of Bengal. Geochimica et Cosmochimica Acta 57 (13), 2981e2990. Castañeda, I.S., et al., 2009. Wet phases in the Sahara/Sahel region and human migration patterns in North Africa. Proceedings of the National Academy of Sciences of the United States of America 106 (48), 20159e20163. Chang, P., et al., 2008. Oceanic link between abrupt changes in the North Atlantic Ocean and the African monsoon. Nature Geoscience 1 (7), 444e448. Chiang, C.H., Biasutti, M., Battisti, D.S., 2003. Sensitivity of the Atlantic intertropical convergence zone to last glacial maximum conditions. Paleoceanography 18. doi:10.1029/2003pa000916. Cuffey, K.M., Clow, G.D., 1997. Temperature, accumulation, and ice sheet elevation in central Greenland through the last deglacial transition. Journal of Geophysical Research 102, 26383e26396. Dansgaard, W., et al., 1993. Evidence for general instability of past climate from a 250-kyr ice-core record. Nature 364, 218e220. demenocal, P., et al., 2000. Abrupt onset and termination of the African humid period: rapid climate responses to gradual insolation forcing. Quaternary Science Reviews 19, 347e361. Denton, G.H., Alley, R.B., Comer, G.C., Broecker, W.S., 2005. The role of seasonality in abrupt climate change. Quaternary Science Reviews 24 (10e11), 1159e1182. Dupont, L.M., Jahns, S., Marret, F., Ning, S., 2000. Vegetation change in equatorial West Africa: time-slices for the last 150 ka. Palaeogeography, Palaeoclimatology, Palaeoecology 155 (1e2), 95e122. Edmond, J.M., Boyle, E.D., Drummond, D., Grant, B., Mislick, T., 1978. Desorption of Barium in the plume of the Zaire (Congo) River. Netherlands Journal of Sea Research 12, 324e328. Fairbank, R.G., Sverdlove, M., Free, R., Wiebe, P.H., Be, W.H., 1982. Vertical distribution and isotopic fractionation of living planktonic foraminifera from the Panama Basin. Nature 298, 841e844. Fleitmann, D., et al., 2011. Holocene and Pleistocene pluvial periods in Yemen, southern Arabia. Quaternary Science Reviews 30 (7e8), 783e787. Fleitmann, D., et al., 2007. East African soil erosion recorded in a 300 year old coral colony from Kenya. Geophysical Research Letters 34 (4), 5. Giresse, P., Aloisi, J.-C., Kuete, M., Monteillet, J., Ngueutchoua, G., 1995. Quaternary sedimentary deposits on the Cameroon shelf: characterization of facies and late quaternary shorelines. Quaternary International 29e30, 75e82. Heinrich, H., 1988. Origin and consequences of cyclic ice rafting in the Northeast Atlantic Ocean during the past 130,000 years. Quaternary Research 29 (2), 142e152. Hemleben, C., Spindler, M., Anderson, O.R., 1989. Modern Planktonic Foraminifera. Springer-Verlag, New York, 363 pp. Hönisch, B., et al., 2011. Planktic foraminifers as recorders of seawater Ba/Ca. Marine Micropaleontology 79 (1e2), 52e57. Johnson, T.C., et al., 2002. A high-resolution paleoclimate record spanning the past 25,000 years in southern East Africa. Science 296 (5565), 113e132. Johnson, T.C., Brown, E.T., Shi, J., 2011. Biogenic silica deposition in Lake Malawi, East Africa over the past 150,000 years. Palaeogeography, Palaeoclimatology, Palaeoecology 303 (1e4), 103e109. Kelly, M.A., et al., 2008. A 10Be chronology of lateglacial and Holocene mountain glaciation in the Scoresby Sund region, east Greenland: implications for seasonality during lateglacial time. Quaternary Science Reviews 27 (25e26), 2273e2282. Koutavas, A., demenocal, P., Olive, G.C., Lynch-Stieglitz, J., 2006. Mid-Holocene El NiñoeSouthern Oscillation (ENSO) attenuation revealed by individual foraminifera in eastern tropical Pacific sediments. Geology 34, 993e996. Kripalani, R.H., Kulkarni, A., Sabade, S.S., 2003. Western Himalayan snow cover and Indian monsoon rainfall: a re-examination with INSAT and NCEP/NCAR data. Theoretical and Applied Climatology 74 (1), 1e18. Kudrass, H.R., Hofmann, A., Doose, H., Emeis, K., Erlenkeuser, H., 2001. Modulation and amplification of climatic changes in the Northern Hemisphere by the Indian summer monsoon during the past 80 k.y. Geology 29 (1), 63e66. Kutzbach, J.E., Liu, Z., 1997. Response of the African monsoon to orbital forcing and ocean feedbacks in the Middle Holocene. Science 278, 440e443. Lea, D.W., Spero, H., 1994. Assessing in the reliability of the paleochemical tracers: barium uptake in the shells of planktonic foraminifera. Paleoceanography 9, 445e452. LeGrande, A.N., Schmidt, G.A., 2006. Global gridded data set of the oxygen isotopic composition in seawater. Geophysical Research Letters 33, L12604. LeGrande, A.N., Schmidt, G.A., 2009. Sources of Holocene variability of oxygen isotopes in paleoclimate archives. Climate of the Past 5, 441e455. Lewis, S.C., LeGrande, A.N., Kelley, M., Schmidt, G.A., 2010. Water vapour source impacts on oxygen isotope variability in tropical precipitation during Heinrich events. Climate of the Past 6, 325e343. Liu, Z., Wang, Y., Gallimore, R., Notaro, M., Prentice, J.C., 2006. On the cause of abrupt vegetation collapse in the North Africa during the Holocene: climate variability vs. vegetation feedback. Geophysical Research Letters 33. doi:10.1029/ 2006GL028062.

S. Weldeab / Quaternary Science Reviews 40 (2012) 21e29 29 Martin, P.A., Lea, D.W., 2002. A simple evaluation of cleaning procedures on fossil benthic foraminiferal Mg/Ca. Geochemistry Geophysics Geosystems 3. doi:10.1029/2001gc000280. McManus, J.F., Francois, R., Gherardi, J.-M., Keigwin, L.D., Brown-Leger, S., 2004. Collapse and rapid resumption of Atlantic meridional circulation linked to deglacial climate changes. Nature 428, 834e837. Moore, S.M., 1997. High fluxes of radium and barium from the mouth of the Ganges- Brahmaputra River during low discharge suggest a large groundwater source. Earth and Planetary Science Letters 150, 141e150. Mulitza, S., Rühlemann, C., 2000. African monsoonal precipitation modulated by interhemispheric temperature gradients. Quaternary Research 53, 270e274. Mulitza, S., et al., 2008. Sahel megadroughts triggered by glacial slowdowns of Atlantic meridional overturning. Paleoceanography 23. Ngueutchoua, G., Giresse, P., 2010. Sand bodies and incised valleys within the Late Quaternary SanagaeNyong delta complex on the middle continental shelf of Cameroon. Marine and Petroleum Geology 27 (10), 2173e2188. Nicholson, S.E., 2009. A revised picture of the structure of the monsoon and land ITCZ over West Africa. Climate Dynamics 32, 1155e1171. Niedermeyer, E.M., et al., 2010. Orbital- and millennial-scale changes in the hydrologic cycle and vegetation in the western African Sahel: insights from individual plant wax delta D and delta C-13. Quaternary Science Reviews 29 (23e24), 2996e3005. Paillard, D., Labeyrie, L., Yiou, P., 1996. Macintosh program performs time-series analysis. EOS Transaction AGU 77, 379. Pausata, F.S.R., Battisti, D.S., Nisancioglu, K.H., Bitz, C.M., 2011. Chinese stalagmite [delta]18o controlled by changes in the Indian monsoon during a simulated Heinrich event. Nature Geoscience 4 (7), 474e480. Rohling, E.J., 2007. Progress in paleosalinity: Overview and presentation of a new approach. Paleoceanography 22 (3), PA3215. Rohling, E.J., et al., 2009. Controls on the East Asian monsoon during the last glacial cycle, based on comparison between Hulu Cave and polar ice-core records. Quaternary Science Reviews 28 (27e28), 3291e3302. Sangen, M., Neumann, K., Eisenberg, J., 2011. Climate-induced fluvial dynamics in tropical Africa around the last glacial maximum? Quaternary Research 76 (3), 417e429. Schmidt, G.A., 1999. Error analysis of paleosalinity calculations. Paleoceanography 14, 422e429. Schmidt, G.A., Bigg, G.R., Rohling, E.J., 1999. Global Seawater Oxygen-18 Database. http://data.giss.nasa.gov/o18data/. Scholz, C.A., et al., 2007. East African megadroughts between 135 and 75 thousand years ago and bearing on early-modern human origins. Proceedings of the National Academy of Sciences 104 (42), 16416e16421. Scholz, C.A., Talbot, M.R., Brown, E.T., Lyons, R.P., 2011. Lithostratigraphy, physical properties and organic matter variability in Lake Malawi Drillcore sediments over the past 145,000 years. Palaeogeography, Palaeoclimatology, Palaeoecology 303 (1e4), 38e50. Shanahan, T.M., et al., 2006. Paleoclimatic variations in West Africa from a record of late Pleistocene and Holocene lake level stands of Lake Bosumtwi, Ghana. Palaeogeography, Palaeoclimatology, Palaeoecology 242 (3e4), 287e302. Siddall, M., et al., 2003. Sea-level fluctuations during the last glacial cycle. Nature 423 (6942), 853e858. Sinha, A., et al., 2005. Variability of Southwest Indian summer monsoon precipitation during the Boiling-Allerod. Geology 33 (10), 813e816. Spero, H., 1998. Life history and stable isotope geochemistry of planktonic foraminifera. In: Norris, R.D., Corfield, R.M. (Eds.), Isotope Paleobiology and Paleoecology. Paleontological Society Papers, Pittsburgh, Pennsylvania, pp. 7e36. Stager, J.C., Mayewski, P.A., Meeker, L.D., 2002. Cooling cycles, Heinrich event 1, and the desiccation of Lake Victoria. Palaeogeography, Palaeoclimatology, Palaeoecology 183, 169e178. Stager, J.C., Ryves, D.B., Chase, B.M., Pausata, F.S.R., 2011. Catastrophic drought in the Afro-Asian monsoon region during Heinrich Event 1. Science 331 (6022), 1299e1302. Timm, O., Kohler, P., Timmermann, A., Menviel, L., 2010. Mechanisms for the onset of the African humid period and Sahara greening 14.5e11 ka BP. Journal of Climate 23 (10), 2612e2633. Timmermann, A., Krebs, U., Justino, F., Goosse, H., Ivanochko, T., 2005. Mechanisms for millennial-scale global synchronization during the last glacial period. Paleoceanography 20. Tjallingii, R., et al., 2008. Coherent high- and low-latitude control of the northwest African hydrological balance. Nature Geoscience 1 (10), 670e675. Turner, D.R., Dickson, A.G., Whitfield, M., 1980. Watererock partition coefficients and the composition of natural waters e a reassessment. Marine Chemistry 9 (3), 211e218. Vidal, L., labeyrie, L., Cortijo, E., Arnold, M., Duplessy, J.C., Michel, E., Becque, S., Van Weering, T.C.E., 1997. Evidence for changes in the North Atlantic Deep Water linked to meltwater surges during the Heinrich events. Earth and Planetary Science Letters 146, 13e27. Waelbroeck, C., 2002. Sea-level and deep water temperature changes derived from benthic foraminifera isotopic records. Quaternary Science Research 21, 295e305. Wang, Y.J., et al., 2001. A high-resolution absolute-dated late Pleistocene monsoon record from Hulu Cave, China. Science 294 (5550), 2345e2348. Weldeab, S. Impact of last glacial atmospheric CO 2 oscillation and Heinrich events on tropical Atlantic sea surface temperatures. Paleoceanography, under review. Weldeab, S., Frank, M., Stichel, T., Haley, B., Sangen, M., 2011. Spatio-temporal evolution of the West African monsoon during the last deglaciation. Geophysical Research Letters 38 (13), L13703. Weldeab, S., Lea, D.W., Schneider, R.R., Andersen, N., 2007a. 155,000 years of West African monsoon and ocean thermal evolution. Science 316, 1303e1307. Weldeab, S., Lea, D.W., Schneider, R.R., Andersen, N., 2007b. Centennial scale climate instabilities in a wet early Holocene West African monsoon. Geophysical Research Letters 34, L24702. Weldeab, S., Schneider, R.R., Kölling, M., Wefer, G., 2005. Holocene African droughts relate to East-equatorial Atlantic cooling. Geology 33, 981e984. doi:10.1130/ G21874.1. Zhao, H., Moore, G.W.K., 2004. On the relationship between Tibetan snow cover, the Tibetan plateau monsoon and the Indian summer monsoon. Geophysical Research Letters 31 (14), L14204.