Content uploaded by M. Sarnthein
Author content
All content in this area was uploaded by M. Sarnthein on Mar 22, 2018
Content may be subject to copyright.
Quaternary Science Reviews 19 (2000) 347}361
Abrupt onset and termination of the African Humid Period:
rapid climate responses to gradual insolation forcing
Peter deMenocal!,*, Joseph Ortiz!, Tom Guilderson", Jess Adkins!, Michael Sarnthein#,
Linda Baker!, Martha Yarusinsky!
!Lamont-Doherty Earth Observatory of Columbia University, Palisades, NY 10964, USA
"Center for Accelerator Mass Spectrometry, Lawrence-Livermore National Laboratory, Livermore CA 94551, USA
#Institute Fuer Geowissens Chafter, Universitaet Kiel, Kiel, Germany
Abstract
A detailed (ca. 100 yr resolution) and well-dated (18 AMS 14C dates to 23 cal. ka BP) record of latest Pleistocene}Holocene
variations in terrigenous (eolian) sediment deposition at ODP Site 658C o!Cap Blanc, Mauritania documents very abrupt,
large-scale changes in subtropical North African climate. The terrigenous record exhibits a well-de"ned period of low in#ux between
14.8 and 5.5 cal. ka BP associated with the African Humid Period, when the Sahara was nearly completely vegetated and supported
numerous perennial lakes; an arid interval corresponding to the Younger Dryas Chronozone punctuates this humid period. The
African Humid Period has been attributed to a strengthening of the African monsoon due to gradual orbital increases in summer
season insolation. However, the onset and termination of this humid period were very abrupt, occurring within decades to centuries.
Both transitions occurred when summer season insolation crossed a nearly identical threshold value, which was 4.2% greater than
present. These abrupt climate responses to gradual insolation forcing require strongly non-linear feedback processes, and current
coupled climate model studies invoke vegetation and ocean temperature feedbacks as candidate mechanisms for the non-linear
climate sensitivity. The African monsoon climate system is thus a low-latitude corollary to the bi-stable behavior of high-latitude deep
ocean thermohaline circulation, which is similarly capable of rapid and large-amplitude climate transitions. (1999 Elsevier Science
Ltd. All rights reserved.
1. Introduction
During the latest Pleistocene and early Holocene, the
now hyperarid Saharan desert was a verdant landscape
nearly completely vegetated with annual grasses and
shrubs (COHMAP Members, 1988; Jolly, 1998; Sar-
nthein, 1978). At that time, subtropical North Africa
was characterized by numerous large and small lakes
which supported abundant savannah and lake margin
fauna such as antelope, gira!e, elephant, hippopotamus,
crocodile, and human populations in regions that today
have almost no measurable precipitation (McIntosh and
McIntosh, 1983). The Holocene African Humid Period
occurred between ca. 9 and 6 cal. ka BP (Ritchie et al.,
1985; Roberts, 1998), but humid conditions had initially
commenced by ca. 14.5 cal. ka BP following full glacial
*Corresponding author. Tel.: 001-914-365-8483; fax: 001-914-365-
8165. Lamont}Doherty Earth Observatory, of Columbia University,
Geoscience 211, Palisades, NY 10964, USA.
E-mail address: peter@ldeo.columbia.edu (P. deMenocal)
hyperarid conditions during the latest Pleistocene (CO-
HMAP Members, 1988; Street and Grove, 1979; Street-
Perrot, et al., 1990; Sarnthein et al., 1982).
The early Holocene greening of North Africa has been
linked to an intensi"cation of the African monsoon due
to earth orbital changes which increased summer season
insolation forcing of the African monsoon. By 10}11 cal.
ka BP, summer insolation in the Northern Hemisphere
had risen to peak levels approximately 8% greater
than today due to earth's orbital precession which
gradually aligned the boreal summer solstice with
perihelion (Berger and Loutre, 1991). Monsoonal climate
is e!ectively a thermodynamic heat engine response
to seasonal radiation forcing (Webster, 1987). During
boreal summer, the North African land surface is
heated more e$ciently than the adjacent oceans and
this establishes a strong low pressure center over North
Africa. It is this land}sea pressure gradient which
largely drives the strong cyclonic in#ow of moist mari-
time surface winds and brings the heavy and highly
seasonal monsoonal precipitation during the boreal
summer months.
0277-3791/99/$ - see front matter (1999 Elsevier Science Ltd. All rights reserved.
PII: S 0 2 7 7 - 3 7 9 1 ( 9 9 ) 0 0 0 8 1 - 5
Climate modeling studies have abundantly
documented the strong sensitivity of monsoonal climate
to orbital changes in summer season insolation
(deMenocal and Rind, 1993; Kutzbach and Guetter,
1986; Kutzbach and Otto-Bliesner, 1982; Pokras and
Mix, 1987; Prell and Kutzbach, 1987). Orbital variations
cause changes in the seasonal distribution of incident
solar radiation, and resulting changes in summer season
insolation a!ect the strength of the summer monsoon.
Using an atmosphere-only climate model, Prell and Kut-
zbach (1987) found that precipitation over North Africa
increased by a factor of "ve over the proportional in-
crease in summer radiation. For example, North African
rainfall increased by approximately 40% (relative to the
control experiment) resulting from the 8% increase in
summer radiation forcing. These orbital variations in
monsoonal surface wind and precipitation "elds a!ected
not only African terrestrial climate, but also the surface
circulation of adjacent tropical and subtropical which are
dynamically linked to the associated monsoonal surface
wind "eld perturbations (Anderson and Prell, 1992;
Clemens et al., 1991; Clemens and Prell, 1990; McIntyre
et al., 1989; Mol"no and McIntyre, 1990b; Rossignol-
Strick, 1985; Tiedemann et al., 1994).
Despite the very high sensitivity of African monsoonal
precipitation to orbital insolation forcing in atmo-
sphere-only models, the calculated increases in the sur-
face hydrological budget were apparently insu$cient to
have established large standing lakes in simulations of
the early Holocene (BrostroKm et al., 1998; Coe and Bon-
an, 1997; Kutzbach, 1996). It is only when coupled ocean
model and responsive land surface (vegetation) model
elements are incorporated into the simulations that Afri-
ca can become su$ciently humid year-round to support
numerous large perennial lakes in the Sahara as the
geologic data document, such as Lake MegaChad which
spanned 330,000 km2(BrostroKm et al., 1998; Coe and
Bonan, 1997; Kutzbach, 1996). Parallel early Holocene
increases in surface ocean temperature, vegetation cover
(lower albedo), and moisture availability are positive
feedback ampli"cations which locally match or exceed
radiative e!ects due the initial insolation forcing alone
(BrostroKm et al., 1998; Coe and Bonan, 1997; Foley et al.,
1994; Hoelzmann et al., 1998; Kutzbach, 1996; Kutzbach
& Liu, 1997; Masson and Joussaume, 1997; Kohfeld and
Harrison, 1999 this issue). Land surface climate and
coupled ocean}atmosphere responses to boundary con-
dition changes are notoriously non-linear. In these more
complex model simulations, gradual boundary condition
forcing such as orbital insolation change commonly re-
sults in a recti"ed integrated climate response which is
rapid and abrupt (Claussen et al., 1998; Manabe and
Stou!er, 1995; Rahmstdorf, 1995).
How abrupt were the climate shifts associated with the
onset and termination of the African Humid Period?
Geologic evidence documenting the mid-Holocene ter-
mination of the African Humid Period suggests a rela-
tively abrupt shift toward more arid conditions which
occurred between ca. 5 and 6 cal. ka BP associated with
the gradual decline in summer season radiation (Gasse
and Van Campo, 1994; Petit-Maire and Guo, 1996;
Street and Grove, 1979; Street-Perrott and Harrison,
1984). Lake levels fell sharply (Street and Grove, 1979),
fossil pollen records from African lake basins show
a rapid retreat of mesic taxa (Lamb et al., 1995; Lezine,
1991; Lezine et al., 1990; Sarnthein et al., 1982) and
sedentary, lacustrine-tradition human populations in the
central and southern Sahara were rapidly replaced by
mobile, pastoralist-tradition cultures (McIntosh and
McIntosh, 1983). Terrestrial geologic records of the Afri-
can Humid Period are commonly incomplete due to
subsequent desiccation and erosion of paleolake basin
sequences, which furthermore can be di$cult to date due
to local hard water reservoir e!ects. With a few notable
exceptions, marine sediment sequences commonly lack
the temporal resolution needed to de"ne century-scale
climatic shifts. Our strategy here is to reconstruct a de-
tailed marine sediment record of the export of windborne
African dust and document its variability over the last
25 cal. ka BP.
2. Site location and climatic setting
We present a continuous, well-dated, and high-resolu-
tion (50}100 yr sampling) marine sediment record of Late
Pleistocene}Holocene changes in West African climate
which records both the timing and abruptness of the
onset and termination of the African Humid Period.
Ocean Drilling Program Site 658C was cored o!Cap
Blanc, Mauritania (20345@N, 18335@W, 2263m; Ruddiman
et al., 1988) during ODP Leg 108. This core has a high
accumulation rate (18 cm/kyr average) due to the dual
in#uences of high regional surface ocean productivity
and high supply of windblown Africa dust (Fig. 1) (Sar-
nthein and Tiedemann, 1989). Northeast Trade winds
parallel the Northwest African margin and resulting sur-
face water divergence promotes upwelling of cold, nutri-
ent-rich waters which support high surface productivity
and high biogenic particle #uxes to the sea#oor through-
out the year (Fischer et al., 1996).
Site 658C is ideally situated to monitor past variations
in northwest African climate. The site is positioned dir-
ectly below the axis of the summer African dust plume
which transports an estimated 400]106tons of mineral
aerosol dust annually from the sub-Saharan and Sahel
regions of Northwest Africa to the adjacent eastern sub-
tropical Atlantic (Fig. 1) (Tetsla!and Wolter, 1980; Pye,
1987). Mineralogic and geochemical data document the
dominant in#uence of these windborne African dusts on
the composition of marine sediments in the eastern
subtropical Atlantic, particularly along the Northwest
348 P. deMenocal et al. /Quaternary Science Reviews 19 (2000) 347}361
Fig. 1. Seasonal climatology of surface winds, rainfall, and atmospheric dust trajectories over subtropical West Africa. During boreal winter months
the land surface cools relative to the ocean and regional atmospheric circulation is dominated by the NE trade winds which advect African dust to the
eastern equatorial Atlantic. The winter African dust trajectory (stippled pattern) follows the NE}SW pattern of the transporting winter trade winds
(Pye, 1987). During boreal summer increased sensible heating over central North Africa drives the cyclonic in#ow of moisture-laden air from the
adjacent eastern equatorial Atlantic which brings sporadic but intense monsoon rains to the sub-Saharan and Sahel regions of West Africa. The
summer African dust plume (stippled pattern) results from strong surface turbulence associated with monsoonal frontal systems and entrained particles
are convectively lifted to mid-tropospheric levels and then transported westward by the African Easterly Jet (Pye, 1987; SchuKtz et al., 1981). Interannual
variations in African dust export are highly correlated to regional precipitation anomalies (Prospero, 1981; Prospero and Nees, 1977; Prospero and
Nees, 1986). The location of Ocean Drilling Program Site 658 o!Cap Blanc, Mauritania is shown.
P. deMenocal et al. /Quaternary Science Reviews 19 (2000) 347}361 349
African margin between 20 and 303N (Sarnthein et al.,
1982; Grousset et al., 1998; Kolla et al., 1979). During
boreal summer, frontal systems associated with the east-
erly-propagating jet generate strongly turbulent surface
winds which entrain mineral aerosols from immature
soils and desert sands (Pye, 1987). The combination of
turbulent suspension and vertical convection lifts aero-
sol-laden airmasses to the middle troposphere which are
then entrained into the African Easterly Jet and carried
westward (Fig. 1). African dust crosses the Atlantic in
roughly six days. In the vicinity of Site 658, the wind-
borne terrigenous sediment is composed of "ne silt-sized
subrounded quartz and feldspar grains with an asso-
ciated illitic matrix re#ecting the highly weathered dust
source areas of the African desert (Grousset et al., 1998;
SchuKtz et al., 1981).
Interannual increases in the export of African mineral
aerosols to the subtropical Atlantic marine boundary
layer are closely linked to negative annual precipitation
anomalies in the dust source areas of West Africa
(Middleton, 1987; Moulin et al., 1997; Prospero and
Nees, 1977). A continuously monitored atmospheric
aerosol sampling station in Barbados has been measur-
ing year-to-year variations in African dust loading for
many decades, and these data have documented greatly
increased aerosol dust loading associated with drought
conditions in Northwest Africa (Prospero and Nees,
1986). These drought periods re#ect reductions in the
strength and duration of the summer African monsoon
which is the dominant source of annual precipitation to
the region. Interannual variations in West African pre-
cipitation and atmospheric dust loading are evidently
linked to persistent sea-surface temperature anomalies in
the North Atlantic and tropical Atlantic sectors (Henn-
ing and Flohn, 1981; Druyan, 1987; Fontaine and Bigot,
1993; Moulin et al., 1997).
3. Analytical methods and age control
Core 1H from Site 658C was continuously subsampled
at 2 cm intervals using a polycarbonate spatula; the
sample interval is equivalent to between 50 and 150 yr
depending on the interval sedimentation rate. The 658C
depth scale used in this study was based on the cumulat-
ive section lengths which were recorded by ODP when
Core 1H was "rst split for analysis (Core 1H: Section
1 (153 cm), Section 2 (154 cm), and Section 3 (155 cm)).
We adopted this depth scale so that all analyses could be
positioned subsequently and unequivocally within the
context of the original archived sediments.
Samples were freeze-dried, weighed and a small por-
tion of each sample was powdered and measured for
carbonate and biogenic opal concentrations. Calcium
carbonate was measured in all samples (n"227) by
coulometry to within 0.3% absolute by weight, and bio-
genic opal was measured in every other sample using
the wet reduction spectrophotometric technique (Mor-
tlock and Froelich, 1989) which had an absolute pre-
cision of 0.48% by weight for these sediments. Organic
carbon was not measured for all samples but ranged
between 0.5 and 1.5% in these samples. The residual
terrigenous (detrital) sediment fraction was determined
by "rst interpolating the opal data to obtain opal percent
values for all samples, and residual terrigenous percent-
ages were calculated by di!erencing. Average carbonate,
opal, and terrigenous percentages at Site 658C were
42, 5, 53%, respectively (Fig. 2). Sediment wet bulk den-
sity was measured shipboard using the gamma-ray at-
tenuation porosity evaluator (GRAPE), and these data
were converted to dry bulk density using the shipboard
physical property data for Site 658 (Ruddiman et al.,
1988).
A benthic oxygen isotope stratigraphy was established
by analyzing between 5 and 8 picked individuals of
Cibicides wuellerstorxfrom the washed '150 lm frac-
tion of each sample. All samples were "rst sonicated in
dionized water then analyzed on the Wood Hole
Oceanographic Institution Finnigan MAT252 mass
spectrometer with the Kiel automated carbonate prep-
aration device. Samples were not roasted prior to analy-
sis. The external precision of the isotopic analyses was
$0.03&for d13C and $0.07&for d18O based on more
than 1200 measurements of NBS19 standard.
Age control was established by accelerator mass spec-
trometry (AMS) radiocarbon dating of picked, mono-
speci"c samples of the planktonic foraminifer
Globigerinoides bulloides over 18 levels spanning the last
23 cal. ka BP. Analyses were conducted at the Center for
Accelerator Mass Spectrometry (CAMS) facility at Law-
rence Livermore National Laboratory. Between 1000
and 1500 G. bulloides individuals were picked from the
'150 lm fraction of each sample and individual sam-
ples were sonicated in dionized water several times and
leached in dilute HCl prior to AMS radiocarbon analy-
sis. The precise reservoir correction to be applied for this
region is uncertain but should be larger than the surface
ocean average of 400 yr due to the upwelling of older
subsurface waters, so we used a conservative reservoir
correction of 500 yr which was subtracted from the raw
radiocarbon ages. Additional 14C ages measured on
G. inyata were consistently younger than ages obtained
for G. bulloides in the same sample (by 400}600 yr,
(Knaack, 1997; Papenfuss, 1999)), so the G. bulloides ages
listed in Table 1 must be considered maximum age esti-
mates. Corrected radiocarbon ages were calibrated to
calendar ages using the Calib 3.03 program (Stuiver and
Reimer, 1993); the radiocarbon dates and calibrated ages
(with $2 sigma ranges) are listed in Table 1. Dates and
ages referred to in this paper are calibrated calendar
ages (rounded to the nearest decade) unless otherwise
speci"ed.
350 P. deMenocal et al. /Quaternary Science Reviews 19 (2000) 347}361
Fig. 2. Sediment composition data and AMS radiocarbon age control from Site 658C. Samples were taken continuously at 2 cm intervals, which is
roughly equivalent to 50}150 yr based on the 18 cm/ka average sedimention rates at Site 658C. A brief hiatus between 14.8 and 17.2 cal. ka BP is
indicated by two closely spaced AMS radiocarbon dates at 324 and 328 cm (Table 1). Note the very abrupt changes in sediment composition which
occur at 326 cm (ca. 14.8 cal. ka BP), 260 cm (12.3 cal. ka BP), and 125 cm (ca. 5.5 cal. ka BP).
Table 1
AMS radiocarbon dating of Site 658C planktonic foraminifera. All values were measured at the Center for Accelerator Mass Spectrometry (CAMS) at
the Lawrence Livermore National Laboratory. Samples of between 1000 and 1500 picked, monospeci"c planktonic foraminifera (G. bulloides) were
sonicated in deionized water prior to analysis. Calibrated ages were rounded to the nearest decade (Stuiver and Reimer, 1993)
Site/core/type/sect./depth1 & 2 Depth (cm) Species 14
C age
(yr BP)
$error 14
C age
(yr), corr."
Calib. Age
(yr BP)
2-sigma calib.
age range
658C!1 H 1 16 18 17 G. bulloides 1130 50 630 600 549}653
658C 1 H 1 28 30 29 G. bulloides 1390 40 890 780 732}899
658C 1 H 1 52 54 53 G. bulloides 1600 40 1100 980 952}1057
658C 1 H 1 64 66 65 G. bulloides 2550 40 2050 1990 1944}2040
658C 1 H 1 74 76 75 G. bulloides 2100 40 1600 (1510) 1393}1555
658C 1 H 1 88 90 89 G. bulloides 4290 40 3790 4150 4088}4230
658C 1 H 1 100 102 101 G. bulloides 4220 40 3720 4030 3983}4092
658C!1 H 1 122 124 123 G. bulloides 5080 40 4580 5300 5098}5313
658C 1 H 2 4 6 158 G. bulloides 7100 40 6600 7470 7393}7518
658C 1 H 2 32 34 186 G. bulloides 7590 40 7090 7850 7826}7918
658C 1 H 2 48 50 202 G. bulloides 9280 60 8780 9710 9650}9885
658C 1 H 2 76 78 230 G. bulloides 10 520 40 10 020 11 210 11 047}11 646
658C 1 H 2 105 107 259 G. bulloides 11 780 40 11 280 13 190 13 122}13 260
658C 1 H 2 124 126 278 G. bulloides 11 780 40 11 280 13 190 13 122}13 260
658C!1 H 3 16 18 324 G. bulloides 13 030 40 12 530 14 690 14 518}14 873
658C!1 H 3 20 22 328 G. bulloides 14 860 60 14 360 17 210 17 103}17 320
658C 1 H 3 42 44 350 G. bulloides 16 490 80 15 990 18 870 18 749}18 992
658C!1 H 3 104 106 412 G. bulloides 19 960 80 19 460 22 910
!Dates used to constrain the `nominalaage model (see text).
"Radiocarbon ages after 500 year reserver corection.
P. deMenocal et al. /Quaternary Science Reviews 19 (2000) 347}361 351
Fig. 3. Biogenic carbonate, opal and terrigenous sediment percentage and #ux records (dashed lines, in g/cm2/ka) from Site 658C. Eighteen AMS
radiocarbon dates were used to constrain the carbonate, opal, and terrigenous percentage time series; age control points are indicated by "lled triangle
symbols to the left of the "gure. Radiocarbon ages were converted to calibrated calendar ages using the Calib 3.03 program (Stuiver and Reimer, 1993)
after applying a 500 yr reservoir correction to the raw 14C ages. Flux data were calculated using the `nominalaage model derived from a subset of
14C age control levels. Flux data are shown by dashed lines and the nominal age control points are shown by open triangles to the right of the "gure.
Note the abruptness of the 14.8, 12.3, and 5.5 cal. ka BP sediment composition transitions, each of which were completed within several centuries.
4. Results
The radiocarbon age model shown in Fig. 2 indicates
a very high average sedimentation rate at Site 658C of
approximately 18 cm/ka. The AMS 14C dates increase
monotonically with depth with the exception of one age
reversal at 89 cm (Fig. 2). The age control data do indi-
cate that sedimentation was continuous spanning the last
ca. 15 ka. However, a nondepositional or erosional hi-
atus is indicated for the interval from 17.20 to 14.80 cal.
ka BP by two closely dated samples at 328 cm (17.21 cal.
ka BP) and 324 cm (14.69 cal. ka BP). A hiatus at this
level was also suggested by early radiocarbon results by
Knaack (1997) and Papenfuss (1999). That this hiatus is
stratigraphically abrupt is supported by the step-like
increase bulk density and the sharp 0.7&increase
in d18O which occurs across this 328}324 cm interval
(Fig. 2). The hiatus appears to be bounded by these dates
and thus spans the "rst phase of the last deglaciation
between 14.80 and 17.20 cal. ka BP. The hiatus may have
been attributable to a coring artifact resulting from post-
recovery gas expansion (Ruddiman et al., 1988). There
are no other similarly sharp in#ections in the radiocar-
bon age model to suggest other hiatuses in the 658C
record.
The Site 658C composition data document rapid and
large-amplitude changes in sediment composition which
occurred during the latest Pleistocene and Holocene
(Figs. 2 and 3). Three major compositional shifts occur-
red over the last ca. 20 cal. ka BP which are centered at
depths of 324 cm (14.65 cal. ka BP), 260 cm (12.32 cal. ka
BP), and 125 cm (5.49 cal. ka BP) (Figs. 2 and 3). Work-
ing from the base of the Site 658C section upwards, there
is a dramatic increase in biogenic carbonate and opal
(and decrease in terrigenous (eolian) sediment) which
occurs between 324 cm and 126 cm (14.65}5.49 cal. ka
BP). This 14.65}5.49 cal. ka BP interval of lower ter-
rigenous percentages was punctuated by a brief but sharp
increase (decrease) in terrigenous (carbonate) sediment
concentrations between 284 and 257 cm (ca. 13.38}12.32
cal. ka BP) which is roughly contemporaneous with the
Younger Dryas period). The sediment compositional
352 P. deMenocal et al. /Quaternary Science Reviews 19 (2000) 347}361
Table 2
Mean sediment component accumulation rates at Site 658C
Age interval Carbonate (g/cm2/ka) Opal (g/cm2/ka) Terrigenous (g/cm2/ka) Total (g/cm2/ka)
0}5.2 ka 6.46 (34%) 1.06 (6%) 11.22 (60%) 18.75
5.3}14.8 ka 8.45 (49%) 1.04 (6%) 7.64 (45%) 17.12
17.2}26.0 ka 5.3 (37%) 0.54 (4%) 8.30 (59%) 14.15
timeseries shown in Fig. 3 were constrained using the
calibrated radiocarbon dates listed in Table 1.
4.1. Sediment yux variations at Site 658C
Were these abrupt compositional shifts attributed to
changes in terrigenous or carbonate sediment supply?
Using the calibrated radiocarbon age model listed in
Table 1 we computed the mass #ux (in g/cm2/ka) for each
constituent by computing the vector product of interval
sedimentation rates, fractional abundance of each con-
sitituent, and dry bulk density (Fig. 3). The actual sedi-
mentation rate pro"le at Site 658 is most likely much
more complex than can be captured by the 18 AMS
dates, so we computed the average component accumula-
tion rates for the three main intervals which show large
compositional di!erences: 26.0}17.2, 14.8}5.5, and 5.5}0
cal. ka BP (Table 2). This `nominalaage model is only
used to estimate large-scale changes in sediment in#uxes
of biogenic and terrigenous components summarized in
Table 2. It is immediately apparent that the Site 658C
sediment #uxes are much larger than typical pelagic
ocean sediments due to highly productive surface waters
and high supply of terrigenous eolian sediment. Average
terrigenous sediment #uxes at Site 658C are roughly 10
times larger than the more distal ODP Site 659 o!West
Africa (Tiedemann et al., 1994), which is consistent with
the log-normal mass-distance particle fallout relationship
for atmospheric aerosols (Pye, 1987; SchuKtz et al., 1981).
The abrupt increase in terrigenous percent which oc-
curred at ca. 5.5 cal. ka BP can be attributed to the nearly
50% increase in terrigenous sediment #ux across this
transition (Fig. 3). The in#ux of detrital, predominantly
African eolian sediment increased 47% from 7.6 to 11.2
g/cm2/ka, whereas the biogenic carbonate #ux decreased
24% across the 5.5 cal. ka BP transition. The abrupt
decrease in terrigenous percent which occurred near 14.8
cal. ka BP can mainly be attributed to an increase in the
supply of biogenic carbonate sediments which increase
59% from 5.31 to 8.5 g/cm2/ka; terrigenous #ux de-
creased 8% from 8.3 to 7.6 g/cm2/ka (Table 2). The
component #ux changes which we have calculated across
the 14.8 cal. ka BP transition must be considered prelimi-
nary and subject to modi"cation since the hiatus needs to
be more "rmly bounded by additional AMS dates. Al-
though the terrigenous #ux variations broadly mirror the
terrigenous percentage data (Fig. 3), changes in carbon-
ate #ux due to changes in productivity or preservation
are also important at Site 658C.
4.2. Abruptness of the sediment composition changes
These climatic transitions were very abrupt and were
evidently completed within a few centuries. These
transition times must be considered maximum estimates
as they were derived from radiocarbon dates of a non-
laminated marine sediment sequence, neither of which
have the temporal resolution to absolutely date precisely
decadal-century scale transitions. However, the high
sedimentation rates provide the necessary relative time
resolution for constraining century-scale climatic jumps
at Site 658C. Using all radiocarbon age control data, we
estimate that the ca. 5.5 cal. ka BP increase in terrigenous
(eolian) sediment supply occurred in less than four centu-
ries centered at 5490$190 yr BP (4780 14C yr BP). The
termination of the Younger Dryas arid period is dated in
this core at 12.3 cal. ka BP and evidently occurred in less
than two centuries (Fig. 3). The abrupt decrease in ter-
rigenous percent at ca. 14.8 cal. ka BP evidently occurred
rapidly as well and is centered at 14,785$90 yr BP
(12,630 14C yr BP). We cannot be certain at this point
that this 14.8 cal. ka BP transition is stratigraphically
distinct from the hiatus, but several lines of evidence
suggest that the hiatus occurs just below the transition.
A calibrated AMS radiocarbon date of 14,692 yr BP was
obtained for the 324 cm sample, which occurred stratig-
raphically within the transition (Figs. 2}4, Table 1), sug-
gesting that the transition itself is intact and distinct from
the hiatus. Furthermore, the 14.8 cal. ka BP date agrees
extremely well with the onset of humid conditions in
Africa derived from a variety of independent paleocli-
matic data sources (see below) (Sarnthein et al., 1994;
Johnson et al., 1996; Sowers and Bender, 1995; Street and
Grove, 1979).
4.3. The Site 658C record of eolian deposition and
comparison with African terrestrial paleoclimate records
The calibrated dates for the subtropical African cli-
mate shifts recorded at Site 658C agree with terrestrial
and marine paleoclimate records of the onset and termi-
nation of the African Humid Period and the Younger
P. deMenocal et al. /Quaternary Science Reviews 19 (2000) 347}361 353
Fig. 4. Comparison of the boreal summer (JJA) average insolation computed for 203N (Berger and Loutre, 1991) with the Site 658C benthic oxygen
isotope record (from analyses of C.wuellerstor,), and terrigenous (eolian) percentage and #ux records spanning the last 25 cal. ka BP. Note the onset
and termination of the African Humid Period in terms of the low eolian dust #ux at Site 658C between ca. 14.8 ka and 5.5 cal. ka BP associated with the
early Holocene rise in summer insolation forcing of the African monsoon. Atmospheric methane concentrations preserved in occluded ice bubbles and
the oxygen isotopic composition of glacial ice in the GISP2 Greenlandice core are also shown (Blunier et al., 1995; Dansgaard, 1993). The onset of the
African Humid Period was synchronous with the end of cold, glacial conditions in Europe, which occurred by ca. 14.5 cal. ka BP, corresponding to the
end of Heinrich event 1 in the North Atlantic (Blunier et al., 1995; Bond et al., 1993; Broecker et al., 1993; Dansgaard, 1993). The rapid onset of humid
conditions in the tropics is also recorded by the abrupt increase in atmospheric methane at ca. 14.7 cal. ka BP, which documents therapid expansion of
tropical wetland methane sources (Blunier et al., 1995). The Site 658C data con"rm terrestrial African paleoclimate records which document a brief (ca.
1 ka) interval of more arid conditions associated with the cool Younger Dryas Chronozone (Gasse et al., 1989; Gasse et al., 1990; Roberts et al., 1993;
Street-Perrott and Perrott, 1990; Williamson et al., 1993). The termination of the African Humid Period at 5.5 cal. ka BP coincides with the
mid-Holocene minimum in atmospheric methane concentration. The subsequent late Holocene methane rise has been attributed to the expansion of
boreal wetlands which were absent during the "rst stages of the deglaciation (Blunier et al., 1995). The timing of the terrigenous transitions at ca. 5.5
and 14.8 cal. ka BP are indicated on the JJA insolation curve ("lled symbols).
Dryas cool period. A survey of subtropical African
paleolakes documents that lake basins in the hyperarid
to arid regions began to "ll and expand near 14.5 cal. ka
BP (Street and Grove, 1979; Street-Perrott and Harrison,
1984; Gasse, 1999). Radiocarbon dates on piston cores
recovered from Lake Victoria "rmly place the onset of
lacustrine deposition there at 14,500 yr BP (12,400$70
14C yr BP) following late Pleistocene aridity and com-
plete desiccation (Johnson et al., 1996). Additionally, the
record of atmospheric methane variability recorded in ice
bubbles trapped in the Greenland ice cores documents
a very abrupt increase in global atmospheric methane
concentrations, probably re#ecting a dramatic increase
tropical wetland CH4production which commenced at
14.7 cal. ka BP (Fig. 4) (Blunier et al., 1995; Sowers and
Bender, 1995). The onset of the African Humid Period
was synchronous with the end of cold, glacial conditions
in Europe and the North Atlantic }Heinrich event
1*which occurred by ca. 14.5 cal. ka BP (Blunier et al.,
1995; Bond et al., 1993; Broecker et al., 1993; Dansgaard,
1993).
The Younger Dryas period is formally known in Euro-
pe and the North Atlantic as a return to cold, near-glacial
conditions between ca. 12.5 and 11.5 cal. ka BP (Broecker
354 P. deMenocal et al. /Quaternary Science Reviews 19 (2000) 347}361
et al., 1988; Mangerud, 1987; Roberts, 1998; Alley et al.,
1999 this issue). In subtropical Africa, the Younger Dryas
period has been associated with a sharp increase in re-
gional aridity which interrupted the African Humid Peri-
od (Gasse et al., 1989; Gasse et al., 1990; Roberts et al.,
1993; Street-Perrott and Perrott, 1990; Williamson et al.,
1993; Gasse, 1999 this issue). The record of terrigenous
(eolian) sedimentation at Site 658C documents a brief
(1 ka) interval of enhanced aridity (increased eolian
transport) which is dated between 13.4-12.3 cal. ka BP
(Figs. 3, 4). This event in Site 658C would appear to
slightly precede (by 800 yr) the well-constrained range of
the Younger Dryas event. As mentioned in the methods
section, our G. bulloides radiocarbon ages were consis-
tently older than ages determined for G. inyata within the
same samples (Knaack, 1997; Papenfuss, 1999), so the G.
bulloides ages must be considered as maximum age esti-
mates. Additionally, the marine 14C reservoir correction
was nearly twice as large as the modern value during
the Younger Dryas (700 yr vs. 400 yr) due to changes
in deep ocean circulation and related changes in ocean-
atmosphere radiocarbon partitioning (Austin et al., 1995;
Bard et al., 1994; Oeschger et al., 1980). Furthermore,
foraminiferal assemblage changes at Site 658C document
large increases in G. bulloides abundances over the
14.8}5.5 cal. ka BP interval (including the Younger
Dryas interval; P. deMenocal, unpub. data), which would
be consistent with increased regional upwelling and older
apparent 14C ages at this time. Together, these factors
would have been su$cient to increase the local apparent
reservoir age correction of the Younger Dryas event to
&800}900 yr at Site 658C.
Terrestrial paleoclimate records indicate that there
was an abrupt onset of more arid conditions in subtropi-
cal North Africa near 8.2 cal. ka BP (Alley et al., 1997;
Gasse and Van Campo, 1994). The development of more
arid conditions at this time spanned from subtropical
Africa to southeast Asia, and is contemporaneous with
a sharp, short-lived cooling event in the North Atlantic
and Greenland Alley et al., 1997; Gasse and Van Campo,
1994). At Site 658C, terrigenous sediment accumulation
increased sharply after ca. 8 cal. ka BP (Figs. 3, 4),
marking a transition toward more arid conditions. How-
ever, in contrast to the African and high-latitude paleocli-
mate records, the sediment transition at Site 658C was
not particularly abrupt nor short-lived. The Site 658C
record suggests that the 8.2 cal. ka BP event marked the
end of the most humid conditions in subtropical Africa,
with a gradual decline in the precipitation-evaporation
balance which culminated in the very abrupt onset of
much more arid conditions at 5.5 cal. ka BP. Planktonic
foraminiferal assemblage variations at the ca. 8 cal. ka
BP level at Site 658C indicate an abrupt and short-lived
cooling event which was the coolest period of the entire
Holocene, being nearly 6}83C cooler than modern values
(deMenocal, unpub. data).
The end of the African Humid Period was very abrupt,
having been completed within several centuries. The onset
of more arid conditions and the contraction of lake basins
evidently occurred across subtropical Africa between
6and5cal.kaBP(GasseandVanCampo,1994;JaKkel,
1979; Munson, 1981; Pachur and Altman, 1997; Petit-
Maire and Guo, 1996; Roberts, et al., 1994; Sarnthein,
1978; Talbot and Delbrias, 1980; Talbot, 1980; Gasse,
1999). The timing of the lake level lowerings associated
with this event are broadly clustered between 6-5 cal. ka
BP, but evidently show some basin-to-basin variability in
the timing of this shift to more arid conditions (Street-
Perrott and Harrison, 1984). The absence of a similarly
abrupt decrease in atmospheric CH4at this time has been
interpreted to re#ect the gradual Holocene emergence and
expansion of boreal wetlands which were absent during
the "rst stages of the deglaciation (Blunier et al., 1995). The
mid-Holocene minimum in atmospheric CH4concentra-
tionsnear5.5}5.0 cal. ka BP (Fig. 4) has been interpreted
to re#ect the loss of this subtropical methane source
(Blunier et al., 1995). The subsequent late Holocene meth-
ane rise has been attributed to the expansion of boreal
wetlands which were absent during the "rst stages of the
deglaciation (Blunier et al., 1995). Many studies have
shown that the primary aridi"cation event began between
6-5 cal. ka BP, followed by a brief return to humid condi-
tionsnear4cal.kaBP(seeMcIntoshandMcIntosh,1983;
Petit-Maire et al., 1987). The Site 658C terrigenous record
exhibits a subsequent, but short-lived reduction in ter-
rigenous sedimentation centered at 4.3}4.1cal.kaBP
which may be synchronous with this humid period (Fig. 4).
5. Discussion
Early investigations into the late Pleistocene}Holo-
cene evolution of African climate recognized the paleocli-
matic signi"cance of these abrupt arid-humid transitions
(Faure et al., 1963). The geologic evidence for the onset
and collapse of the African Humid Period is so striking
that these events have been termed `climatic crisesa,
acknowledging the sheer magnitude and spatial scale of
these climatic events (Rognon, 1983). The collective con-
tribution of decades of African paleoclimatic data now
bears witness to a century-scale replacement of a once
verdant, lake-dotted, and populated landscape by the
now-hyperarid desert largely devoid of vegetation and
life. Our results from Site 658C suggest that this and
other transitions occurred extremely abruptly, within
a few decades to centuries.
5.1. Timing of the African Humid Period relative to orbital
insolation forcing
One of the most striking features of the Site 658C
terrigenous record beyond the abruptness of the climate
P. deMenocal et al. /Quaternary Science Reviews 19 (2000) 347}361 355
transitions is the timing of the transitions themselves
relative to the orbital insolation forcing. Boreal summer
insolation was calculated for the boreal summer season
(JJA) at 203N, the latitude of maximum sensible heating
over subtropical West Africa (Webster, 1987; Berger and
Loutre, 1991). As shown in Fig. 4, the onset and termina-
tion of the African Humid Period near 14.8 and 5.5 cal.
ka BP occurred at nearly identical summer insolation
values of 470 W/m2, which is roughly 4.2% above the
modern value of 451 W/m2. This observation may indi-
cate a climate threshold response, whereby subtropical
African climate #ips abruptly between humid and arid
modes as summer radiation, the primary forcing of the
summer monsoon, passes a critical value. As described
below, fully coupled (ocean}atmosphere}vegetation) cli-
mate model simulations of the African monsoon suggest
that there are two stable solutions of the Saharan region
(vegetated vs. non-vegetated) to prescribed orbital insola-
tion boundary condition changes (see Claussen et al.,
1999; Brovkin et al., 1998). Furthermore, these climate
models also indicate that the arid-humid transition oc-
curs at the same orbital insolation threshold.
What changes in subtropical ocean}atmosphere cli-
mate dynamics led to such an abrupt climate response to
the gradual, periodic insolation forcing of the African
monsoon? Are these transitions re#ective of a threshold
response and, if so, what are the attractors? Climate
model simulations provide a useful context for studying
the dynamical processes that a!ect the seasonal timing
and amplitude of the monsoon, as well as its coupled
sensitivity to associated changes in regional vegetation
and surface ocean responses. Here, we consider climate
model investigations into the sensitivity of monsoonal
climate to prescribed changes in insolation forcing alone,
and then we assess the additional responses resulting
from associated (coupled) changes in surface ocean tem-
peratures, African vegetation and surface hydrologic
budgets, and, "nally, we discuss the response of fully
coupled ocean}atmosphere}vegetation models to the
initial orbital radiation forcing.
5.2. Atmosphere-only and coupled climate model
simulations of the mid-Holocene African monsoon
Early climate modeling e!orts using uncoupled, atmo-
sphere-only general circulation models (GCMs) indicate
that while the African monsoon system is very responsive
to insolation forcing, the response is strongly linear.
Using the NCAR-CCM1 climate model, Prell and Kut-
zbach (1987) found that the 8% relative increase in sum-
mer insolation due to earth orbital geometries at ca. 10
cal. ka BP produced a roughly 40% increase in subtropi-
cal African monsoonal precipitation (Fig. 5). This "ve-
fold gain between insolation forcing and monsoonal pre-
cipitation was found to be highly linear for the full late
Pleistocene range of orbital con"gurations (e.g. between
Fig. 5. The linear relationship between orbital insolation forcing and
simulated changes in African monsoonal precipitation using an atmo-
sphere-only climate model which does not employ coupled ocean
and/or vegetation feedbacks (from Prell and Kutzbach, 1987). The
monsoon is strongly linked to orbital changes in seasonal radiation
forcing, but there is no indication of non-linear or threshold responses
from these early atmosphere-only experiments. Strongly positive, non-
linear ampli"cations of the primary insolation forcing of monsoonal
climate are achieved in climate model simulations which explicitly
include coupled vegetation, land surface processes, and surface ocean
responses (see text).
!5to#15% of modern summer radiation). These
results are generally consistent with the paleoclimate
data, although the simulated P-E budget would not have
been su$cient to form and sustain the large, perennial
lakes which are known to have existed across the Sahara
and Sahel regions during the early to mid-Holocene
(BrostroKm et al., 1998; Coe and Bonan, 1997; Kutzbach,
1996). A paradigm which arose from these early simula-
tions is that monsoonal circulation is essentially a heat
engine which responds linearly to prescribed radiation
changes.
5.3. Ewects of vegetation and land-surface feedbacks on
African monsoonal climate
Subsequent climate modeling studies have explicitly
examined how monsoonal climate sensitivity is a!ected
by formally incorporating coupled vegetation and/or
coupled ocean model elements. Using an asynchronously
coupled ocean}atmosphere GCM, Kutzbach and Liu
(1997) found that the increased mid-Holocene summer
radiation also increased surface ocean temperatures
(by 0.43C) which further increased summer African
monsoonal precipitation by 25% over uncoupled,
atmosphere-only simulations. Similar conclusions have
been described by Hewitt and Mitchell (1998) using the
356 P. deMenocal et al. /Quaternary Science Reviews 19 (2000) 347}361
Hadley Centre synchronously coupled ocean-atmo-
sphere GCM (HADCM2). In these experiments, the
monsoon was both greatly intensi"ed due to increased
moisture advection from the warmer surface ocean, and
rains penetrated much deeper into the North African
interior (and later into the year) relative to the uncoupled
simulations (Hewitt and Mitchell, 1998; Kutzbach and
Liu, 1997). Although these experiments were not able to
produce the monsoon intensi"cation required to match
mid-Holocene paleoclimate data (e.g. abundant peren-
nial lakes throughout the present-day expanse of the
Sahara Desert), the surface ocean feedbacks were positive
and large.
A separate set of climate model studies have explicitly
considered the e!ects of prescribed and coupled changes
in African vegetation and surface water on the strength of
the African monsoon; these feedbacks were also positive
and large (Brovkin et al., 1998; Coe and Bonan, 1997;
Foley, 1994; Hoelzmann et al., 1998; Kutzbach, 1996).
Idealized studies have shown that prescribed changes in
African vegetation can increase monsoonal precipitation
to levels which are roughly equal to the increases due to
orbital radiation forcing alone. Kutzbach (1996) em-
ployed a coupled atmosphere}vegetation GCM in series
of experiments to explore how changes in vegetation and
soil type a!ected the response of the African monsoon to
insolation forcing. He found that the prescribed replace-
ment of desert with grassland vegetation, and of desert
soils with loamy soils resulted in separate precipitation
increases of 6 and 10%, respectively, as compared to
the radiation-only precipitation increase of 12%. Coe
and Bonan (1997) investigated the e!ect of prescribed
mid-Holocene increases in surface water coverage
(lakes and wetlands) over northern Africa and found,
too, that the monsoonal circulation was enhanced due to
associated changes in the surface albedo and latent heat
budgets. BrostroKm et al. (1998) found that prescri-
bing mid-Holocene land-surface conditions in their
coupled atmosphere}land surface climate model pro-
longed the monsoon season by two months and caused
it to penetrate nearly 300 km further into north
Africa.
5.4. Abrupt mid-Holocene shift in subtropical
African climate simulated by a fully coupled
atmosphere}ocean}vegetation climate model
All of the climate models described thus far are too
complex to be run as transient experiments for many
thousands of model years to quantify time-dependent
changes in the coupled ocean}atmosphere}vegetation re-
sponses to the initial insolation forcing. Ganopolski et al.
(1998) report the application of the Climate and Bio-
sphere (CLIMBER) model, a zonally averaged, coupled
ocean}atmosphere model with an equilibrium vegetative
subsystem. The model is computationally e$cient and it
can be used to explore transient changes in ocean-
atmosphere-vegetation feedbacks over many millennia.
The model has very coarse (103]513) spatial resolution
and consists of a 2.5-dimensional dynamical}statistical
atmospheric model coupled to a multibasin, zonally
averaged ocean model with responsive sea-ice, and a
terrestrial vegetation parameterization that responds
to growing degree days and precipitation. Fractional
vegetation types are permitted within gridboxes so that
the e!ective resolution of the vegetative model is greater
than the model resolution (Ganopoloski et al., 1998).
The model was later modi"ed (CLIMBER2) to include
dynamic (diagnostic) vegetation responses to surface
climate changes (Brovkin et al., 1998; Claussen et al.,
1999).
The CLIMBER2 model was con"gured for a transient
simulation of the last 9000 yr of the Holocene, being
forced only by the gradual changes in incident summer
season radiation due earth orbital variations (Claussen
et al., 1999). CLIMBER2 simulates an abrupt termina-
tion of the African Humid Period centered at 5450 yr BP
in response to the gradual drop in insolation from
9000 yr. BP to the present (Fig. 6a}c). The humid-arid
transition occurs abruptly, within several centuries (Fig.
6b,c). The timing of the climate model transition was
stable, occurring at 5440$30 yr, based on an average of
10 separate simulations that di!ered only in the initial
conditions but were otherwise forced by the same insola-
tion series (Claussen et al., 1999). The timing and shape of
this modeled transition compare favorably with the
abrupt mid-Holocene increase in eolian dust supply at
Site 658C centered at 5490$190 yr BP (4780 14C yr BP)
(Fig. 6d).
The abruptness of the Claussen et al. (1999) modeled
climate transition was attributed to the vegetation sensi-
tivity to changing precipitation "elds associated with the
gradually decreasing insolation forcing of the monsoon
(Claussen et al., 1999). Atmosphere-only, and coupled
ocean}atmosphere simulations that lack responsive veg-
etation exhibited smooth climatic responses to the grad-
ual orbital forcing (Claussen et al., 1999). The abrupt
climate response to gradual insolation forcing in this
model can be attributed to the positive feedbacks which
link changes subtropical vegetation, albedo, and precipi-
tation. The gradually decreasing monsoonal precipita-
tion leads to decreases in vegetation cover which raise the
surface albedo as desert sands become increasingly
prevalent, thereby reducing the e$ciency of the initial
radiation forcing of the monsoon. As demonstrated with
fully coupled model experiments, the individual e!ects of
vegetation and land-surface feedbacks can far exceed the
fundamental radiation sensitivity of monsoonal climate
(BrostroKm et al., 1998; Brovkin et al., 1998; Coe and
Bonan, 1997; Foley, 1994; Hewitt and Mitchell, 1998;
Hoelzmann et al., 1998; Kutzbach, 1996; Kutzbach and
Liu, 1997).
P. deMenocal et al. /Quaternary Science Reviews 19 (2000) 347}361 357
Fig. 6. Average Northern Hemisphere summer insolation (JJA, in W/m2) used to force the coupled ocean}atmosphere}vegetation CLIMBER2 model
for the 9 000 yr BP to present transient simulation of Holocene changes in subtropical African climate (Claussen et al., 1999) (a). In the fully coupled
model, annual Saharan precipitation (b) and fractional Saharan vegetation cover (c) decreases very abruptly at 5440$30 yr BP (based on an ensemble
average of 10 separate transient simulations; (Claussen et al., 1999)). The abrupt African climate responses were attributed to the highly non-linear
(positive) feedbacks linking progressive decreases in regional precipitation, vegetation cover loss, and increasing albedo (Claussen et al., 1999). Note the
remarkable similarity between the shape and timing of the abrupt mid-Holocene termination of the African Humid Period at 5.5 cal. ka BP for the
climate model simulation (b), (c) and the Site 658C record of West African eolian dust supply variations (d).
6. Conclusions
The record of terrigenous (eolian) sedimentation at
West African Site 658C documents the very abrupt onset
and termination of the African Humid period near 14.8
and 5.5 cal. ka BP, respectively. These transitions appar-
ently occurred within several decades to centuries, and
were much too rapid to be driven by a simple linear
358 P. deMenocal et al. /Quaternary Science Reviews 19 (2000) 347}361
responses to gradual insolation forcing. Current climate
modeling results suggest two (positive) feedback pro-
cesses that may have been important factors contributing
to the abruptness of these transitions: Coupled vegeta-
tion-albedo feedback and surface ocean temperature}
moisture transport feedback. These processes can `rec-
tifyathe sinusoidal orbital forcing signal to produce the
square-wave climate response signal we see at Site 658C
and in terrestrial African paleoclimate records of abrupt
climatic change (Figs. 4 and 6). The timing of the
transitions themselves relative to insolation forcing indi-
cates that they occurred when summer insolation crossed
a threshold value of 470 W/m2, which is roughly 4.2%
above modern values. The abruptness of these climate
transitions at Site 658C suggest that this recti"cation
process is both very e!ective and very stable. These
feedback processes interact with primary (e.g. radiation)
forcing to lock subtropical African climate into one of
two (vegetated vs. non-vegetated) stable equilibrium cli-
matic solutions (Claussen et al., 1998). As such, the sub-
tropical monsoon climate system may be viewed as
a low-latitude corollary to the well-documented, bi-
stable behavior of high-latitude deep ocean thermohaline
circulation (Broecker, 1994; Broecker, et al., 1985;
Manabe and Stou!er, 1988; Rahmstdorf, 1995; Stocker
et al., 1992).
Acknowledgements
The authors would like to thank John Miller and
Paula Weiss from the Ocean Drilling Program for their
help in the detailed sampling Site 658C. Helpful
comments and signi"cant input to this paper were
contributed by Bob Anderson, Andre Berger, Martin
Claussen, Claude Hillaire-Marcel, George Kukla, and
David Rind. Comments by two anonymous reviewers
greatly improved the "nal manuscript. This project was
supported by the Marine Geology and Geophysics divis-
ion of the National Science Foundation. This is LDEO
contribution number 5961.
References
Alley, R.B., Mayewski, P.A., Sowers, T., Stuiver, M., Taylor, K.C.,
Clark, P.U., 1997. Holocene climatic instability: a prominent, wide-
spread event 8200 yr ago. Geology 25 (6), 483}486.
Alley et al. 1999. The Younger Dryas cold interval as viewed from
central Greenland Quaternary Science Reviews, 19, this issue.
Anderson, D.M., Prell, W.L., 1992. The structure of the southwest
monsoon winds over the Arabian Sea: observations, simulations,
and marine geological evidence. Journal of Geophysical Research
97 (C10), 15481}15487.
Austin, W.E., Bard, E., Hunt, J.B., Kroon, D., Peacock, J.D., 1995.
The 14C age of the Icelandic Vedde Ash: Implications for
Younger Dryas marine reservoir corrections. Radiocarbon 37 (1),
53}62.
Bard, E., Arnad, M. Mangerud, J. Pateine, M., Labeyne, L., Duprat, J.,
Melieres, M.A., Sonstegaard, E., Duplessy, J.-C., 1994. The North
Atlantic atmosphere-sea surface 14C gradient during the Younger
Dryas climatic event. Earth and Planetary Science Letters 126,
275}287.
Berger, A., Loutre, M.F., 1991. Insolation values for the climate of the
last 10 million years. Quaternary Science Reviews 10, 297}317.
Blunier, T., Chappellaz, J., Schwander, J., Stau!er, B., Raynaud, D.,
1995. Variations in atmospheric methane concentration during the
Holocene epoch. Nature 374, 46}49.
Bond, G., Broecker, W.S., Johnsen, S., McManus, J., Labeyrie, L.,
Jouzel, J., Bonani, G., 1993. Correlations between climate records
from North Atlantic sediments and Greenland ice. Nature 365,
143}147.
Broecker, W., Bond, G., McManus, J., 1993. Heinrich eventstriggers of
ocean circulation change? In: Peltier, W.R. (Ed.), Ice in the Climate
System. Springer-Verlag, Berlin, pp. 161}166.
Broecker, W.S., 1994. An unstable superconveyor. Nature 367, 414}415.
Broecker, W.S., Andree, M., Wol#i, W., Oeschger, H., Bonani, G.,
Kennett, J., Peteet, D., 1988. The chronology of the last deglaci-
ation: Implications to the cause of the Younger Dryas event.
Paleoceanography 3 (1), 1}19.
Broecker, W.S., Peteet, D.M., Rind, D., 1985. Does the ocean-atmo-
sphere system have more than one stable mode of operation?.
Nature 315, 21}26.
BrostroKm, A., Coe, M., Harrison, S.P., Gallimore, R., Kutzbach, J.E.,
Foley, J., Prentice, I.C., Behling, P., 1998. Land surface feedbacks
and paleomonsoons in northern Africa. Geophysical Research Let-
ters 25 (19), 3615}3618.
Brovkin, V., Claussen, M., Petoukhov, V., Ganopolski, A., 1998. On the
stability of the atmosphere-vegetation system in the Sahel/Sahara
region. Journal of Geophysical Research 103 (D4), 31613}31624.
Claussen, M., Brovkin, V., Ganopolski, A., Kubatzki, C., & Petoukhov,
V. (1998). Modeling global terrestrial vegetation-climate interac-
tion. Philosophical Transactions of the Royal Society of London,
353 (B).
Claussen, M., Kubatzki, C., Brovkin, V., Ganopolski, A., Hoelzmann,
P., Pachur, H.-J. (1999). Simulation of an abrupt change in Saharan
vegetation in the mid-Holocene. Geophysical Research Letters,.
Clemens, S., Prell, W., Murray, D., Shimmield, G., Weedon, G., 1991.
Forcing mechanisms of the Indian Ocean monsoon. Nature 353,
720}725.
Clemens, S.C., Prell, W.L., 1990. Late Pleistocene variability of Arabian
Sea summer monsoon winds and continental aridity: Eolian records
from the lithogenic component of deep-sea sediments. Paleoceanog-
raphy 5, l09}145.
Coe, M., Bonan, G., 1997. Feedbacks between climate and surface water
in northern Africa during the middle Holocene. Journal of Geo-
physical Research 102 (D10), 11087}11101.
COHMAP Members (1988). Climatic changes of the last 18,000 years:
Observations and model simulations, Science, 241, 1043}1052.
Dansgaard, W., et al., 1993. Evidence for general instability of past
climate from a 250-kyr ice core record. Nature 364, 218}220.
deMenocal, P.B., Rind, D., 1993. Sensitivity of Asian and African
climate to variations in seasonal insolation, glacial ice cover, sea-
surface temperature, and Asian orography. Journal of Geophysical
Research 98 (4), 7265}7287.
Druyan, L.M., 1987. GCM studies of the African monsoon. Climatic
Dynamics 2, 117}126.
Faure, H., Manguin, E., Nydal, R., 1963. Formations lacustres du
Quaternaire supierieur du Niger oriental: diatomites et a(ges ab-
solus. Bulletin de Bureau de Researches de Geologie et Minieres 3,
41}63.
Fischer, G., Donner, B., Ratmeyer, V., Davenport, R., Wefer, G., 1996.
Distinct year-to-year particle #ux variations o!Cap Blanc during
1988-1991: Relation to d18O-deduced sea-surface temperatures and
trade winds. Journal of Marine Research 54, 73}98.
P. deMenocal et al. /Quaternary Science Reviews 19 (2000) 347}361 359
Foley, J., Kutzbach, J., Coe, M., Levis, S., 1994. Feedbacks between
climate and boreal forests during the Holocene epoch. Nature 371,
52}53.
Foley, J.A., 1994. The sensitivity of the terrestrial biosphere to climatic
change: A simulation of the middle Holocene. Global Biogeochemi-
cal Cycles 8 (4), 505}525.
Fontaine, B., Bigot, S., 1993. West African rainfall de"cits and sea-
surface temperatures. International Journal of Climatology 13,
271}285.
Ganopoloski, A., Kubatzki, C., Claussen, M., Brovkin, V., Petoukhov,
V., 1998. The in#uence of vegetation-atmosphere-ocean interaction
on climate during the mid-Holocene. Science 280, 1916}1919.
Gasse, F., LeHdeHe, V., Massault, M., Fontes, J.C., 1989. Water level
#uctuations of Lake Tanganyika in phase with oceanic changes
during the last glaciation and deglaciation. Nature 342, 57}59.
Gasse, F., TeHhet, R., Durand, A., Gibert, E., Fontes, J.C., 1990. The
arid-humid transition in the Sahara and Sahel during the last
deglaciation. Nature 346, 141}146.
Gasse, F., Van Campo, E., 1994. Abrupt post-glacial climate events in
West Asia and North Africa monsoon domains. Earth and Planet-
ary Science Letters 126, 435}456.
Gasse et al. 1999. Hydrological changes in the African tropics since
the Last Glacial Maximum. Quaternary Science Reviews 19, this
issue.
Grousset, F.E., Parra, M., Bory, A., Martinez, P., Bertrand, P., Shim-
mield, G., Ellam, R.M., 1998. Saharan wind regimes traced by the
Sr-Nd isotopic composition of subtropical Atlantic sediments: last
glacial maximum vs. today. Quaternary Science Reviews 17, 395}409.
Henning, D., Flohn, H., 1981. Some aspects of evaporation and sensible
heat #ux of the tropical Atlantic. Contributions to Atmospheric
Physics 53 (3), 430}441.
Hewitt, C.D., Mitchell, J.F.B.H., 1998. A fully coupled GCM simulation
of the climate of the mid-Holocene. Geophysical Research Letters
25 (3), 361}364.
Hoelzmann, P., Jolly, D., Harrison, S.P., Laarif, F., Bonne"lle, R.,
Pachur, H.-J., 1998. Mid-Holocene land surface conditions in north-
ern Africa and the Arabian Peninsula: A data set for the analysis of
biogeochemical feedbacks in the climate system. Global Bi-
ogeochemical Cycles 12, 35}52.
JaKkel, D., 1979. Run-o!and #uvial formation processes in the Tibesti
Mountains as indicators of climatic history in the central Sahara
during the late Pleistocene. Palaeoecologia Africana 11, 13}44.
Johnson, T.C., Scholz, C.A., Talbot, M.R., Kelts, K., Ricketts, R.D.,
Ngobi, G., Beuning, K., Ssemmanda, I., McGill, J.W., 1996. Late
Pleistocene desiccation of Lake Victoria and rapid evolution of
cichlid "shes. Science 273, 1091}1093.
Jolly, D., et al., 1998. Biome reconstruction from pollen and plant
macrofossil data for Africa and the Arabian Peninsula at 0 and 6 ka.
Journal of Biogeography 25, 1007}10027.
Knaack, J.J., 1997. Eine neue Transferfunktion zur Rekonstruktion der
PalaKoproduktivitaKt aus Gemeinschaften mariner Diatomeen. Beri-
chte-Reports, Geol.-PalaKont. Inst. Univ. Kiel 83, 1}118.
Kohfeld and Harrison, 1999. How well can we simulate past climates?
Evaluating the models using global palaeoenvironmental data sets.
Quaternary Science Reviews, 19 this issue.
Kolla, V., Biscaye, P.E., Hanley, A.F., 1979. Distribution of quartz in
late Quaternary Atlantic sediments in relation to climate. Quater-
nary Research 11, 261}277.
Kutzbach, J.E., 1996. Vegetation and soil feedbacks on the response of
the African monsoon to orbital forcing in the early to middle
Holocene. Nature 384, 623}626.
Kutzbach, J.E., Guetter, P.J., 1986. The in#uence of changing orbital
parameters and surface boundary conditions on climate simulations
for the past 18,000 year. Journal of Atmospheric Science 43, 1726}1759.
Kutzbach, J.E., Liu, Z., 1997. Response of the African monsoon to
orbital forcing and ocean feedbacks in the middle Holocene. Science
278, 440}444.
Kutzbach, J.E., Otto-Bliesner, B.L., 1982. The sensitivity of the African-
Asian monsoonal climate to orbital parameter changes for 9000
years BP in a low-resolution general circulation model. Journal of
Atmospheric Science 39, 1177}1188.
Lamb, H.F., Gasse, F., Benkaddour, A., El Hamouti, N., van der Kaars,
S., Perkins, W.T., Pearce, N.J., Roberts, C.N., 1995. Relation be-
tween century-scale Holocene arid intervals in tropical and temper-
ate zones. Nature 373, 134}137.
Lezine, A.M., 1991. West African paleoclimates during the last climatic
cycle inferred from an Atlantic deep-sea pollen record. Quaternary
Research 35, 456}463.
Lezine, A.-M., Casanova, J., Hillaire-Marcel, C., 1990. Across an early
Holocene humid phase in western Sahara: Pollen and isotope
stratigraphy. Geology 18, 264}267.
Manabe, S., Stou!er, R.J., 1988. Two Stable Equilibria of a Coupled
Ocean-Atmosphere Model. Journal of Climate 1, 841}866.
Manabe, S., Stou!er, R.J., 1995. Simulation of an abrupt climate change
induced by freshwater input to the North Atlantic Ocean. Nature
378, 165}167.
Mangerud, J., 1987. The Allerod/Younger Dryas boundary. In: Berger,
W.H., Labeyrie, L.D. (Eds.), Abrupt Climatic Change. D. Reidel
Publishing Company, Dordrecht, pp. 163}171.
Masson, V., Joussaume, S., 1997. Energetics of the 6000-yr BP atmo-
spheric circulation in Boreal summer, from large-scale to monsoon
areas: A study with two versions of the LMD AGCM. Journal of
Climate 10, 2888}2902.
McIntosh, S.K., McIntosh, R.J., 1983. Current directions in West Afri-
can prehistory. Annual Review of Anthropology 12, 215}258.
McIntyre, A., Ruddiman, W.F., Karlin, K., Mix, A.C., 1989. Surface
water response of the equatorial Atlantic Ocean to orbital forcing.
Paleoceanography 4 (1), 19}55.
Middleton, N., 1987. Deserti"cation and wind erosion in the western
Sahel: the example of Mauritania. School of Geography. University
of Oxford, Research Paper 40, 1}26.
Mol"no, B., McIntyre, A., 1990b. Precessional forcing of nutricline
dynamics in the equatorial Atlantic. Science 249, 766}769.
Mortlock, R.A., Froelich, P.N., 1989. A simple method for the rapid
determination of biogenic opal in pelagic marine sediments. Deep-
Sea Research 36, 1415}1426.
Moulin, C., Lambert, C., Dulac, F., Dayan, U., 1997. Control of atmo-
spheric export of dust from North Africa by the North Atlantic
oscillation. Nature 387, 691}694.
Munson, P.J., 1981. A late Holocene (c. 4500-2300) climatic chronology
for the southwest Sahara. Palaeoecologia Africana 13, 53}60.
Oeschger, H., Welten, M., Eicher, U., Moll, M., Riesen, T., Siegenthaler,
U., Wegmuller, S., 1980. 14C and other parameters during the
Younger Dryas cold phase. Radiocarbon 22 (2), 299}310.
Pachur, H.-J., Altman, N. (1997). The Quaternary (Holocene, ca. 8000a
BP). In H. Schandelmeier, & Reynolds, P.-O. Paleogeographic-
paleotectonic atlas of northeastern Africa, Arabia, and adjacent
areas Late Neo-Proterozoic to Holocene. Rotterdam: Balkema.
Papenfuss, T., 1999. Glazial-interglaziale Variation von Plankton-
produktivitaet und Naehrsto!gehalten im tropisch-subtropischen
Ostatlantik im Abbild der Barium-Gehalte. Berichte-Reports, Inst.
fuer Geowissenschaften, Univ. Kiel 5, 1}119.
Petit-Maire, N., Guo, Z., 1996. Mis en evidence de variations climat-
iques Holocenes rapides, en phase dans les deserts actuels de Chine
et de Nord de l'Afrique. Sciences de la Terre et des Planetes 322,
847}851.
Petit-Maire, N., Fabre, J., Carbonel, P., Schulz, E., Aucour, A.M., 1987.
La depression de Taoudenni (Sahara malien) a l'Holocene.
Geodynamique 2, 61}67.
Pokras, E.M., Mix, A.C., 1987. Earth's precession cycle and Quaternary
climatic change in tropical Africa. Nature 326, 486}487.
Prell, W.L., Kutzbach, J.E., 1987. Monsoon variability over
the past 150,000 years. Journal of Geophysical Research 92,
8411}8425.
360 P. deMenocal et al. /Quaternary Science Reviews 19 (2000) 347}361
Prospero, J., 1981. Arid regions as sources of mineral aerosols in the
marine atmosphere. Geological Society of America Special Publica-
tions 186, 71}86.
Prospero, J.M., Nees, R.T., 1977. Dust concentration in the atmosphere
of the equatorial North Atlantic; possible relationship to Sahelian
drought. Science 196, 1196}1198.
Prospero, J.M., Nees, R.T., 1986. Impact of North African drought and
el Nino on mineral dust in the Barbados trade winds. Nature 320,
735}738.
Pye, A.C., 1987. In: Eolian Dust and Dust Deposits. Acamedic Press,
New York.
Rahmstdorf, S., 1995. Bifurcations of the Atlantic thermohaline circula-
tion in response to changes in the hydrological cycle. Nature 378,
145}150.
Ritchie, J.C., Eyles, C.H., Haynes, C.V., 1985. Sediment and pollen
evidence for an early to mid-Holocene humid period in the eastern
Sahara. Nature 330, 645}647.
Roberts, N., 1998. In: The Holocene Oxford, Blackwell, 316 pp.
Roberts, N., Lamb, H.F., El Hamouti, N., Barker, P., 1994. Abrupt
Holocene hydroclimatic events: Palaeolimnological evidence from
Northwest Africa. In: Pye, A.C.M.a.K. (Ed.), Environmental Change
in Drylands: Biogeographical and Geomorphological Perspectives.
Wiley, London, pp. 163}175.
Roberts, N., Taieb, M., Barker, P., Danati, B., Icole, M., Williamson, D.,
1993. Timing of the Younger Dryas event in East Africa from
lake-level changes. Nature 366, 146}148.
Rognon, P., 1983. Essai de de"nition et typologie des crises climatiques.
Bull. Inst. Geol. Bassin Aquitaine 34, 151}164.
Rossignol-Strick, M., 1985. Mediterranean quaternary sapropels,
an immediate resonse of the African Monsoon to variation of
insolation. Palaeogeography, palaeoclimatology, palaeoecology 49,
237}263.
Ruddiman, W., Sarnthein, M., Baldauf, J., 1988. Proceedings of the
Ocean Drilling Program Leg 108, U. S. G. P. O!., Ed., pp. 1073.
Sarnthein, M., 1978. Sand deserts during the glacial maximum and
climate optimum. Nature, 272, 43}46.
Sarnthein, M., Erlenkeuser, H., Zahn, R., 1982. Termination I:
The response of continental climate in the subtropics as recorded
in deep-sea sediments. Bull. Inst. du Bassin d'Aquitaire 31,
393}407.
Sarnthein, M., Tiedemann, R., 1989. Toward a high-resolution stable
isotope stratigraphy of the last 3.4 million years: Sites 658 and 659
o!northwest Africa. In: Ruddiman, W., Sarnthein, M. (Eds.), Pro-
ceedings of the ocean drilling program, scienti"c results, Ocean
Drilling Program. College Station, TX, pp. 167}185.
SchuKtz, L., Jaenicke, R., Pietrek, H., 1981. Saharan dust transport of the
North Atlantic Ocean. Geological Society of America Bulletin
Special Paper 186, 87}100.
Sowers, T., Bender, M., 1995. Climate records covering the last deglaci-
ation. Science 269, 210}214.
Stocker, T.F., Wright, D.G., Broecker, W.S., 1992. The in#uence of
high-latitude surface forcing on the global thermohaline circulation.
Paleoceanography 7, 529}541.
Street, F.A., Grove, A.T., 1979. Global maps of lake-level #uctuations
since 30,000 years BO. Quaternary Research 12, 83}118.
Street-Perrot, F.A., Mitchell, J.F.B., Marchand, D.S., Brunner, J.S.,
1990. Malankovitch and albedo forcing of the tropical monsoon:
A comparison of geological evidence and numerical simulations for
9000 yr BP. Transactions of the Royal Society of Edinburgh (Earth
Science) 81, 407}427.
Street-Perrott, F.A., Harrison, S.A., 1984. Temporal variations in
lake levels since 30,000 yr BP }An index of the global hydrological
cycle. In: Hansen, J.E., Takahashi, T. (Eds.), Climate Processes and
Climate Sensitivity. Amer. Geophys. Union, Wash. DC, pp.
118}129.
Street-Perrott, F.A., Perrott, R.A., 1990. Abrupt climate #uctuations in
the tropics: the in#uence of Atlantic Ocean circulation. Nature 343,
607}612.
Stuiver, M., Reimer, P.J., 1993. Extended 14C data base and revised
CALIB 3.0 14C age calibration program. Radiocarbon 35, 215}230.
Talbot, M.A., Delbrias, G., 1980. A new late Pleistocene-Holocene
water level curve for Lake Bosumtwi, Ghana. Earth Planetary
Science Letters 47, 336}344.
Talbot, M.R., 1980. Environmental responses to climatic change in the
West African Sahel over the past 20,000 years. In: Williams, M.A.,
Faure, H. (Eds.), The Sahara and the Nile. Balkema, Rotterdam, pp.
527}582.
Tetzla!, G., Wolter, K., 1980. Meteorological patterns and the trans-
port of mineral dust from the North African continent. Paleoecol.
Africa 12, 31}42.
Tiedemann, R., Sarnthein, M., Shackleton, N.J., 1994. Astronomic time-
scale for the Pliocene Atlantic d18O and dust #ux records of ODP
Site 659. Paleoceanography 9 (4), 619}638.
Webster, P.J., 1987. The variable and interactive monsoon. In: Fein,
J.S., Stephens, P.L. (Eds.), Monsoons. Wiley, New York, pp.
269}330.
Williamson, D., Taieb, M., Damnati, B., Icole, M., Thouveny, N., 1993.
Equatorial extension of the Younger Dryas event: rock magnetic
evidence from Lake Magadi (Kenya). Global and Planetary Change
7, 235}242.
P. deMenocal et al. /Quaternary Science Reviews 19 (2000) 347}361 361
