Newsletters on Stratigraphy, Vol. 47/3 (2014), 263282

Published online July 2014; published in print September 2014

Article

Chronological constraints on Pleistocene sapropel

depositions from high-resolution geochemical
records of ODP Sites 967 and 968
T. Y. M. Konijnendijk*, M. Ziegler, and L. J. Lourens
With 6 figures, 3 tables and 5 appendices
Abstract. We completed and merged an existing high-resolution X-Ray Fluorescence (XRF) data set derived from Ocean Drilling Program (ODP) site 968 with new data from adjacent ODP site 967. An astronomical age model spanning the last 1.05 Myr was constructed for the spliced record using the highly linear
relation between the elemental ratio of titanium and aluminum in the sediment (Ti/Al) and insolation. This
rendered detailed ages for sapropel deposition in the Eastern Mediterranean. Our results imply major revisions of previous sapropel age models below MIS 11, with changes of up to two precession cycles in the interval ~ 450850 ka of the ODP 967 sapropel chronology. Based on the Ti/Al age model, we find that color
reflectance used as an indicator for sapropels in this interval of ODP site 967 and 968 is highly incongruent with the Ti/Al proxy as well as insolation forcing during periods of minima in the 405 kyr eccentricity cycle. Our findings indicate that the use of color reflectance as a proxy for so-called ghost sapropels in
these intervals is not reliable. Lastly, we demonstrate the presence of a strong obliquity signal in the Ti/Al
record that lags obliquity forcing by 4 0.7 kyr). This time lag is only marginally longer than the adopted
2.7 1.1 kyr for the dominantly precession-tuned age model, and is hence much shorter than the generally
accepted ice sheet response time to insolation forcing. This suggests that the obliquity-bound changes in Ti/Al
are not glacial controlled and most likely reflect changes in low-latitude climate oscillations, such as the monsoon.
Key words. sapropel, Mediterranean, astronomical tuning

1.

Introduction

linked to increased discharge from the river Nile, resulting from increased amount of rainfall in the Niles
catchment area through an intensified North African
Monsoon (NAM) (Rossignol-Strick 1983). Sapropels
occur mainly in the Eastern Mediterranean and are
used as tools for correlating sediment cores and for
studying Mediterranean climate and ocean circulation
(e. g. Lourens 2004).
The link between sapropel formation and the precession-bound periods of high summer insolation at

Sapropel deposition occurs periodically in the Mediterranean Sea (Rossignol-Strick 1983, Hilgen 1991,
Lourens et al. 1996). This process is paced by the precession cycle of the Earths rotational axis, and is
thought to occur as a result of stagnating circulation in
the Mediterranean basin through increased stratification. A freshwater cap at the surface of the Mediterranean Sea during sapropel formation is commonly

Authors address:
Department of Earth Sciences, Faculty of geosciences, Utrecht University, Budapestlaan 4, 3584 CD Utrecht, The Netherlands.
* Corresponding author: [email protected]
2014 Gebrder Borntraeger, Stuttgart, Germany
DOI: 10.1127/0078-0421/2014/0047

www.borntraeger-cramer.de
0078-0421/2014/0047 $ 5.00
eschweizerbart_xxx

264 T. Y. M. Konijnendijk et al.

2.1

the Northern Hemisphere allows us to establish accurate age depth relationships of sediment cores and
land-based sections. This method is the principle behind the astronomically tuned time scale for the late
Neogene (Hilgen et al. 1993, Lourens 2004). Accordingly, the stratigraphic midpoints of sapropels are
tuned to maxima in the June 21st 65 N insolation
curve with a lag of 3000 years (3 kyr) to obtain an age
model (e. g. Hilgen 1991, Emeis et al. 2000, Lourens
2004). The 3 kyr time lag is based on 14C AMS dating
of the Early Holocene sapropel S1, the midpoint of
which falls at 8.5 ka 3 kyr behind the corresponding
insolation maximum at 11.5 ka. This lag is assumed to
be inherent to the processes behind sapropel formation
and is commonly kept constant when tuning sapropels
to insolation (Lourens et al. 1996, Emeis et al. 2000).
The astronomical tuning of sapropels to insolation
can be tested by comparing the resulting age model
to that derived by other methods. For instance Lourens
et al. (1996) validated their sapropel tuning by comparing planktic foraminiferal stable oxygen isotope
results with those of benthic foraminifera derived
from the open ocean. In addition, Lourens (2004) and
Ziegler et al. (2010) evaluated the late Pleistocene
sapropel chronology against the results of amongst
others Langereis et al. (1997) by using tephras, magnetochronology and a number of other proxies. In this
paper, we elaborate on the tuning strategy of Ziegler et
al. (2010) by extending the high-resolution geochemical data sets of ODP Site 968 and merging it with new
data of ODP Site 967 over the past 1050 ka. Unfortunately, the interval at ODP site 968 below core A3 is
heavily disturbed, forcing us to incorporate the adjacent site, ODP 967, for a continuous record. The tuning approach is based on elemental ratios in the bulk
sediment composition, which reflect regional climate.
The resulting data provide an alternative, independent
age model for the cores involved and will be used to
test and evaluate previous sapropel chronologies (e. g.
Emeis et al. 2000, Lourens 2004).

2.

Revised splice 967 section

The splice of ODP 967 (appendix I and II) is a revision

of the original splice created by Sakamoto et al.
(1998). The revised composite depths (rmcd) that are
used in this paper will thus differ from previously
published data of ODP 967. To enable comparison, we
list shipboard meters below sea floor (mbsf) as well as
rmcd here. The topmost 30 mbsf contains only minor
deviations. These are of little consequence because
we incorporated only data of ODP 968 in this part of
the composite (section 2.2). The changes to the initial
splice become more substantial, however, below
30 mbsf. Notably, below the interval of core B4 ending at 29.38 mbsf our splice reverts to core C4 at
29.56 mbsf (33.05 rmcd), rather than core A4, which
was heavily sampled in prior studies. At 36.10 mbsf
(39.72 rmcd) our splice jumps from core C4 to core B5
at 35.68 mbsf. Here it rejoins the initial splice for a
short interval, though at 40.56 mbsf (44.56 rmcd) it
jumps to core C5 at 39.52 mbsf, 260 cm above the
point of the initial splice.

2.2

Composite section 967/968

In order to merge the existing data obtained from

ODP 968 with the new data of ODP 967, the shipboard
color reflectance (CR) data was used to align the
cores. The reflectance values at 550 nm of the separate
splices were plotted on the revised composite depth
scale of ODP 967 (Appendix I and Fig. 1). The wavelength 550 nm is typically indicative of organic matter
content and is therefore useful in detecting nutrient
cycles (Emeis et al. 1996). This proxy exhibits ample
variability that is largely similar up to minute details in
both cores due to their proximity.
Tie points, such as sapropels, were identified (Appendix III). The ODP 968 data was transferred to a
new depth scale by fitting characteristic features to
ODP 967 and applying linear interpolation between tie
points (Fig. 1). In a next step we identified the interval
in ODP 967 that covered the gap between ODP 968
cores A2 and A3, and sampled it with sufficient overlap to tie the records of ODP 967 and 968 together.
Characteristic features in Ti/Al and Ba/Al were used to
determine the splice tie points between data from ODP
967 and 968. Similarly, the lowest part of ODP 968 A3
was identified in ODP 967 and sampled with overlap.
This resulted in a single, continuous record through the
combination of the two sites. The result is a splice of
the combined ODP cores (Fig. 1).

Material and Methods

Our study uses data from ODP Sites 967 and 968 from
the Eastern Mediterranean. ODP Site 967 (34 04 N,
32 43 E) was drilled south of Cyprus; near the Eratosthenes seamount at a water depth of 2554 m. ODP
968 (34 20 N, 32 45 E) was drilled on the Eratosthenes seamount at a water depth of 1961 m (Emeis et
al. 1996).

eschweizerbart_xxx

Chronological constraints on Pleistocene sapropel depositions

s7

s8

s9

s10

sb

s11 s12

s13

s14 s15

265

s16 s17

ODP 967
Colour reflectance
at 550 nm (%)

10
15
20
25
30
35

ODP 968
Colour reflectance
at 550 nm (%)

12
16
20
24
28
32

10

12

14

16

18

20

22

24

26

28

RMCD2013
(equivalent)

Fig. 1. The color reflectance records of ODP cores 967 and 968, set to core 967 RMCD (equivalent). The sampled interval
in core 967 to complete the existing 968 record is indicated with bright red (grey in print version) markers. Sapropels defined by Emeis et al. (2000) are marked with grey bars.

2.3

XRF measurements

The combined splice of ODP 967 and 968 was sampled at 2 cm intervals. The samples of 35 gram were
freeze-dried and then ground and homogenized. The
powdered sediment was placed in a furnace and
heated stepwise to 100 C, 450 C, 550 C, 800 C and
1000C. At each interval the relative weight loss was
measured to obtain residual water content, organic
matter content and carbonate minerals, respectively.
The residue of this treatment was then molten into
glass beads for XRF scanning. 600 mg of the sample
powder was mixed with 6000 mg lithium tetraborate
(Li2B4O7, Spektromelt), and fused to glass beads. The
beads were analyzed by a Philips PW 2400 X-ray
spectrometer. Analytical precision was determined by
parallel analysis of one international standard (ISE
921) and one in-house standard to be better than 2 %
for Al and Ti.

3.

Age model

3.1

0340 ka

To create a detailed chronology for the composite ODP

967/968 record we made stepwise improvements to
the initial age model of Lourens et al. (2004), who correlated sapropel midpoints to their inferred 65 N summer insolation maxima, including a fixed 3-kyr time
lag. We followed Ziegler et al. (2010), who dated the
youngest part of the record in detail by correlating the
remarkable similarity between variations in the ratio
of titanium versus aluminum (Ti/Al) in ODP 968 to the
speleothem oxygen isotope (18Ospeleothem) records of
the Sanbao/Hulu caves (Wang et al. 2008). The Ti/Al in
the sediments at ODP sites 967 and 968 is considered
to be determined by two major sources: Nile River
suspended matter and windblown dust (Wehausen and
Brumsack 2000, Lourens et al. 2001). Because the sus-

eschweizerbart_xxx

266 T. Y. M. Konijnendijk et al.

pended particles that contain titanium will be heavier,
they will be preferentially deposited close to the river
mouth. The river derived sediment in ODP site 967 and
968 will therefore have been depleted in titanium before deposition at the drill sites, and the titanium in the
sediment will therefore be predominantly of aeolian
origin. Windblown dust is estimated to contribute between 65% and 95% of the titanium in the sediment
(Lourens 2001). The aeolian input depends on the
availability of dust for windblown transport in the
Northern Sahara. This is mainly controlled by vegetation cover of the surface (Ehrmann et al. 2013) in a tight
relation with North African precipitation (Claussen et
al. 2013). The observed Ti/Al changes at ODP sites 967
and 968 are therefore linked to monsoonal-induced
humidity changes in Central to North Africa.
The 18Ospeleothem record of Wang et al. (2008) reflects primarily insolation-driven monsoon activity in
East Asia and is yet one of the best U-Series dated climate records covering the past 340 ka. As such, this
record can be used to constrain leads and lags in the

climate system to astronomical forcing and hence to

test the assumed 3-kyr time lag between monsoon intensity and the formation of Mediterranean sapropels.
Figure 2 shows the correlation between the Ti/Al
and color reflectance records of ODP 968 to the U/Thdated 18O record of the SB-Hulu speleothems (Ziegler et al. 2010). The tuning is straightforward for the
upper 340 kyr. Frequency and cross-spectral analysis
using Analyseries (Paillard et al. 1996) reveals that the
23 kyr signal in the speleothem record, and hence the
Ti/Al record, lags 21 June 65 N insolation by about
2.7 1.1 kyr, which means the 3-kyr lag adopted by
Lourens et al. (1996) is within error.
Rossignol-Strick and Paterne (1997) suggested the
3-kyr time lag between the S1 and the Holocene insolation optimum was partly explained by the occurrence
of the Younger Dryas cold spell, delaying deglaciation
warming and triggering cold, dry conditions in the
Mediterranean region at that time adversely affecting vegetation development. Ziegler et al. (2010) accordingly argued that the phase lag found in Late

ODP 967/968
Ti/Al

0.06

0.07

0.08

10
15
20
25

18Ocalcite
Speleothem record

-12

ODP 967/968
Colour reflectance
at 550 nm (%)

0.09

30

-10

-8

-6

-4

50

100

150

200

250

300

350

400

450

Age (ka)

Fig. 2. The composite Ti/Al record, tuned to the U/Th dated speleothem record of Cheng et al. (2009). Dashed lines indicate tie points between the records. Black dots in the speleothem record indicate U/Th dates.

eschweizerbart_xxx

Chronological constraints on Pleistocene sapropel depositions

267

Table 1 The combined splice of ODP 967 and ODP 968 cores. Rmcd depth values are according to the revised splice for
ODP 967 as presented in appendix I and II. For ODP 968, the rmcd values represent their ODP 967 analog depth.
Site

Hole

Core

Sect

Top
(cm)

Depth
(mbsf)

968

1w

968

968

Site

Hole

Core

Sect

Top
(cm)

Depth
(mbsf)

Depth
(rmcd2013)
ODP 967
(equivalent)

10

0.1

1w

80

6.8

968

2w

54

5.94

5.39

2w

38

13.88

968

2w

144

14.34

12.36

968

2w

80

18.8

967

2H

30

14.3

16.14

968

3H

18.54

967

2H

22

15.72

17.62

968

3H

CC

20

28.34

967

3H

16

23.46

24.70

967

4H

114

25.44

967

3H

114

27.44

28.70

967

4H

58

29.38

967

4H

106

29.56

33.05

967

5H

38

35.68

967

4H

10

36.1

39.72

967

5H

74

40.54

967

5H

39.52

44.50

967

6H

114

45.94

967

5H

148

45.48

51.10

967

6H

94

50.24

0.71

54.95

Pleistocene monsoon response to insolation results

from North Atlantic cold spells, associated with the
so-called Heinrich events. They reasoned that the lag
would most likely be smaller or absent in the Early
Pleistocene as the ice caps were smaller and the cold
spells were less intense and less frequent. The timing
of this proposed shift is still largely uncertain.
A Ca/Sr-based proxy for ice rafted debris (IRD) of
the North Atlantic ODP Site U1308 showed for instance that the occurrence of Heinrich events were
most likely negligible signal prior to ~ 650 ka, whereas proxies for IRD low in carbonates (of different
source area), such as Si/Sr, revealed however that
Heinrich events occurred until at least ~ 1350 ka
(Hodell et al. 2008). At present we cannot exclude that
this mechanism might have influenced monsoon climates throughout the Pleistocene and we therefore
choose to keep the 2.7 kyr lag constant for the tuning
of our record prior to 340 ka.

insolation time series (Laskar et al. 1993) with a fixed

2.7 kyr time lag (Fig. 3). For this purpose, we first performed a Blackman-Tukey spectral analysis on the
Ti/Al series in the depth domain using the Analyseries
program (Paillard et al. 1996) on the lower part of the
record ( 15 rmcd). The analysis revealed a dominant
frequency at 1.70 0.2 cycles/m, which we attributed
to the imprint of the precession cycle. Using a Gaussian filter, we extracted this precession-related frequency and determined the successive minimum and maximum values to tie them to their inferred insolation extremes (Table 2). In some cases we chose as tie points
the intervals of maximum rate of change, in particular
when the interference between obliquity and precession results in exceptionally broad insolation maxima
(e. g. at 540 ka; Fig. 3). Tuning of the Ti/Al record to
65 N June 21st insolation curve is overall straightforward, especially in intervals with pronounced eccentricity.

3.2

3.3

Age model construction

for 3401050 ka

Turbidites

For our final time series, we eliminated a number of

turbiditic layers in the sediments aged between 650
to 950 ka (appendix IV). These layers, though not always apparent from the core surface, are identifiable

Prior to 340 ka, our new astronomically-tuned age

model was constructed by correlating the precessionrelated variations in the Ti/Al record to the June 65 N

eschweizerbart_xxx

eschweizerbart_xxx

Ti/Al
in depth domain

0.09

0.08

0.07

0.06

i-cycle

0.09

0.08

0.07

0.06

300

28

30

32

38

S*2

400

34 36

S10

40

42

44

18

48

500

46

Sb

50

S11

20

52

54

S13

S12

24

600

56

58

60

62

S*3 S*4 S14 S15 Sc

22

66

68

Age (ka)

700

64

S16
6 S17

T T

26

70 72

RMCD 2013

76 78

800

74

28

80

30
T

82

S*5

86

TT

S*7

900

84

S*6

32
T

88

90

36

94

96

S21 S22

1000

92

S18 S19 S20

0

34

98

S23

100

1100

102

400

440

480

520

560

600

Fig. 3. The Ti/Al record matched to 65 N June 21st insolation. The filtered dominant signal in the depth domain (dashed line; top) was used to mathematically determine midpoints or, alternatively, maximum slope. Dashed lines connect the tie points in the records from depth to time domain. Variations
in the record are tied to the insolation target curve with a 2.7 kyr lag. Where the dashed lines cross the insolation curve, closed dots indicate a midpoint
was used, open dots for slope. Grey vertical bars at the top indicate turbidites. Note that in the bottom record, turbidites are filtered out of the data.

Ti/Al
in time domain

16

insolation
(65N June 21st) (W/M2)
Sedrate
cm/kyr

14

268 T. Y. M. Konijnendijk et al.

Chronological constraints on Pleistocene sapropel depositions

Table 2 Tie points used to define the age/depth relation for
the combined splice.
Depth (rmcd2013)
ODP 967 (equivalent)
0
1.06
1.25
4.18
4.44
4.53
4.77
5.19
5.26
5.37
5.42
7.16
7.77
9.41
9.65
9.84
10.67
11.10
11.47
11.65
11.88
12.38
13.88
14.94
15.36
15.73
16.0
17.9
18.6
19.4
19.8
20.2
20.7
21.9
22.4
22.7
23.3
24.0
24.5
25.1
25.7
26.7
27.6
28.2
29.2
29.7
30.3
30.7
31.1
31.9
32.8
33.3
33.8
35.1
36.2

Target age

Correlated feature

5.1
10.2
77.3
83.9
86.6
91
99.6
104
105.9
109.7
121.4
129.5
165.5
169.8
178.5
191.9
198.5
209.5
213
224.1
239
288
317
332
340
354.5
401.5
431.2
460.0
475.0
494.0
513.0
542.5
562.5
574.8
596.0
617.0
646.0
667.0
689.0
728.0
758.0
780.0
812.0
821.5
840.0
857.2
871.7
897.2
922.0
933.0
952.6
997.2
1048.0

top s1
bot s1
top s3
bot s3
top prec s3
bot prec s3
top s4
bot s4
top prec s4
bot prec s4
top s5
bot s5
top s6
dip s6
bot s6
top s7
bot s7
top s8
dip s8
bot s8
mid s9
mid s*
onset i-cycle 28
mid s10
onset i-cycle 30
mid i-cycle 32
end i-cycle 38
mid i-cycle 41
mid i-cycle 44
end i-cycle 46
end i-cycle 48
mid i-cycle 49
onset i-cycle 50
mid i-cycle 53
mid i-cycle 54
mid i-cycle 56
mid i-cycle 58
mid i-cycle 60
mid i-cycle 62
mid i-cycle 64
end i-cycle 68
mid i-cycle 71
end i-cycle 74
onset i-cycle 76
mid i-cycle 78
end i-cycle 80
end i-cycle 82
mid i-cycle 83
onset i-cycle 85
end i-cycle 87
mid i-cycle 88
end i-cycle 90
mid i-cycle 94
mid i-cycle 98

269

through sudden and extreme enrichment (depletion) of

the elements Ti, Si, Zr (Ca), and to a lesser extent, Fe,
Na, K, and P. This enrichment is the result of a sorting
effect favoring the relatively larger, more angular
grains associated with windblown sediments and
therefore aeolian elements, at this site.
The turbidites seem to occur predominantly during
precession minima, when Ti/Al values are generally
low. Such relations between astronomical forcing and
turbiditic systems have been documented before in this
basin (Postma et al. 1993, Weltje and de Boer 1993).
Often, they coincide with sapropels though not exclusively so. On the other hand all turbiditic layers
coincide with the glacial terminations (appendix IV)
as documented by Lisiecki and Raymo (2005), suggesting that they were triggered by sudden changes in
glacio-eustatic sea level. In particular, the onset of recurrent turbiditic layers after 950 ka (MIS 25/26) may
be related to the Mid Pleistocene Transition (MPT)
when global ice volume and associated sea level
changes became larger. The absence of turbiditic layers after 650 ka, i. e. around the end of the MPT, is
most likely related to the predominant use of data from
ODP 968 in the composite after 650 ka, the depositional depth of which is ~ 600 m shallower than that of
ODP 967. As a result, ODP 968 may be situated above
the typical depth where the mass movements occur on
the slopes of the Eratosthenes seamount.

4.

Evaluation of the Late Pleistocene sapropel Chronology

Our new Ti/Al tuning results in a highly detailed

chronology for events, such as sapropel depositions
and tephra layers that can be compared to other studies. Table 3 gives an overview of the new ages for the
sapropel midpoints compared to those given by Emeis
et al. (2000), Langereis et al. (1997) and Lourens
(2004). The differences are small or negligible down
to ~ 400 ka (see also Ziegler et al. 2010). A first major
deviation occurs during the eccentricity minimum
around ~ 400 ka, where we correlate sapropel b in
ODP 967 (Emeis et al. 2000) to one insolation cycle
(i-cycle) older, resulting in an age difference of
~ 12 kyr due to the relatively short precession period at
that time. Sapropel b in ODP Site 967 has been correlated to SAP 8 in ODP Site 964 (Emeis et al. 2000) and
S11 in KC01B (Lourens 2004), and is deposited at the
onset of MIS 11. The S11 was initially correlated to the
insolation maximum at 407 ka (Lourens et al. 1996),

eschweizerbart_xxx

270 T. Y. M. Konijnendijk et al.

since the maximum summer insolation value is much
stronger than that at 419 ka. The Ti/Al record however allows little room for such an interpretation: the
sharp gradient in insolation around 404 ka is reflected
very clearly in the Ti/Al ratio (Fig. 4). Adhering to the
proposed age of 407 ka for sapropel b would put the
Ti/Al record in antiphase with insolation, for which
there are no plausible arguments. Moreover, it would
compress the record above this interval, and stretch it
below, resulting in unlikely changes in sedimentation
rate. Hence, our tuning suggests that the age of the
S11, SAP 8 and Sapropel b is one i-cycle older than
initially proposed or that the deposition of these sapropels was not time equivalent.
Our new sapropel chronology of ODP 967/968 is
consistently older than proposed by Emeis et al. (2000)
below this point down to ~ 860 ka. In particular, most
sapropels should be tuned one, and in a few cases two,
i-cycles older than proposed by Emeis et al. (2000).

s8

s9

s10

sb

s*1

Our new tuning is largely consistent however with the

revised chronology used in the ATNTS2004 (Lourens
et al. 2004) and that of ODP 964 (Lourens 2004). The
only exception is S13, related to i-cycle 48 according
to Emeis et al. (2000). The latter interval, which appears to be a double sapropel with S12 just above it
(Fig. 4), seems to correspond to an obliquity maximum
rather than precession related forcing. The interference
between the two orbital parameters creates a broad insolation maximum here around 530 ka. S13, correlated to Sa in KC01 or SAP10 in ODP 964 as proposed
by Lourens et al. (2004), is not related to i-cycle 52.
The interval between 700950 ka does not contain
any sapropels and has proven problematic to tune. A
more thorough discussion on this interval will be presented in the next section. Prior to 950 ka, the sapropel
chronology of Emeis et al. (2000) is in agreement with
that of Lourens et al. (2004) and compares well with
the ages we obtained for S19 through S22.

s11 s12 *2 *3 s14 s15

s13 s s

sc

s16 s17

s*4 s*5

s*6

s18 s19 s20 s21

s22 s23

Ti/Al
ODP 967/968

0.06

0.07

0.08

0.09

insolation
(65N June 21st) (W/M2)

560

520

480

440

200

300

400

500

600

700

800

900

1000

1100

Age (ka)

Fig. 4. The extended composite Ti/Al record, tuned to 21 june 65 N insolation (Laskar et al. 1993). The dashed line is the
combination of 21 kyr and 41 kyr variability filtered out of the Ti/Al record using Analyseries (Paillard et al. 1996). Sapropels as defined by Emeis et al. (2000), set on the Ti/Al age model, are marked with grey bars.

eschweizerbart_xxx

5.1

Incongruent behavior of the proxies

26

27

28

29

30

31

32

33

34

The Ti/Al proxy has enabled us to make a new tuning

for the combined record of ODP cores 967 and 968.
As a consequence the age model for the sapropels in
these cores has shifted significantly in large parts of
the record. This new tuning is robust and the age estimates are more accurate than before. However, there
are some incongruencies between the different proxies
in the interval between 700950 ka, which contains
very few sapropel deposits. The relatively weak insolation changes in this time interval of low eccentricity
creates difficulties with tuning because proxies lack a
characteristic pattern to tie to insolation. Emeis et al.
(2000) use so called red intervals as indicator for
missing or so-called ghost sapropels for tuning to insolation. However, the color reflectance proxy in this
interval behaves capriciously and contradictory to the
Ti/Al record (Fig. 5). Of these proxies, Ti/Al has the
simplest operational mechanics, and generally seems
to be much more linearly related to insolation (Fig. 4).
Therefore we argue that the Ti/Al tuning renders the
most reliable age model.
Our Ti/Al based ages for sapropels or red intervals
in this period, as identified by Emeis et al. (2000),
often diverge significantly from their estimates
(Table 3). For example, one of the red intervals tuned
by Emeis et al. (2000) to i-cycle 74 appears much more
related to a maximum in obliquity than to a minimum
in precession. A similar situation seems to hold for
the red interval S*5, tuned to i-cycle 82 by Emeis et
al., that occurs during an insolation minimum but coincides with an interval of high obliquity. Many other
red intervals defined by Emeis et al. in the period 700
850 ka appear, according to our age model, unrelated
to insolation and have as such not been dated.

Obliquity influence
in the Ti/Al proxy record

Ti/Al
ODP 967/968
eschweizerbart_xxx

30

25

20

15

0.09

0.08

0.07

0.06

25

Another intriguing fact is the relatively large influence

of obliquity found in the Ti/Al record (Fig. 6). Its influence has been recognized before in the sapropel record,
e. g. by Lourens et al. (1996). They constructed a curve
by generating a normalized precession curve and detracting a normalized obliquity signal divided by two
(P-12T curve). This is very similar to the 65 N June 21
insolation curve in terms of relative power and has a
power spectrum that is nearly identical. The added benefit of the P-12T curve is that you can separately con-

Color Reflectance
at 550nm (%)

5.2

Fig. 5. The Ti/Al record and color reflectance record of ODP site 967 plotted over depth. The two proxies are especially incongruent between 2934 RMCD.

Discussion

RMCD2013

5.

271
35

Chronological constraints on Pleistocene sapropel depositions

Table 3 Comparison of sapropel chronologies. The columns on the left present the chronology of Emeis et al. (2000) for ODP 967 sapropel midpoints and
those of red intervals identified by Emeis et al. (2000). The middle columns present the chronology of Lourens (2004) for Eastern Mediterranean
sapropels. The presented data represent sapropels from ODP 964 as defined by Emeis et al. (2000), KC01, and KC01B, and are correlated to the
ODP 967 sapropels in the columns to the left. The columns on the right present the data of this study. This is an extension of the data of Ziegler
et al. (2010), which consists of all data up to 967968-s10. The lithological boundaries and midpoints of sapropels are as defined by Emeis et al.
(2000). The nomenclature follows Emeis et al. (2000), with a prefixed 967968 to avoid confusion with the classification in KC01 and KC01B.
An exception is s (~ 288 ka), which is renamed to match its KC01 counterpart. Several of the red intervals identified by Emeis et al. (2000) are
renamed as an s* followed by a number, though not all are recognized as a response to insolation (n. i.).

272 T. Y. M. Konijnendijk et al.

eschweizerbart_xxx

Chronological constraints on Pleistocene sapropel depositions

trol the phasing of obliquity power and precession

power in the target curve. A different phase for the different factors creates a different interference pattern in
the target curve. Lourens et al. (1996) used this feature
to test which phasing for obliquity best matches the
thick/thin alternations of the sapropel record. They find
that a phase lag of 3 kyr for tilt has the closest match.
That is close to the ~ 4 0.7 kyr lag between the Ti/Al
record and obliquity found in this study by cross spectral analysis. Although some influence of global climate cannot be excluded, this relatively short lag indicates that the 41 kyr signal in the record is not due to a
teleconnection with high boreal latitudes. The time lag
of ice volume response to insolation is ~ 7 kyr (Lisiecki
and Raymo 2005). Instead, it implies that the obliquity
power with its associated time lag found in these monsoon proxies is an intrinsic part of this low latitude climate system.
The intriguing possibility of low latitude climate
control of obliquity has been discussed before. In an
earlier study into the dust content of ODP 967, performed by Larrasoaa et al. (2003), magnetic measurements are used to find the hematite content of the
sediment. Arguably this proxy works much in the same
fashion as our Ti/Al record. The 3 Myr record of Larrasoana et al. (2003) extends well into the 41 kyr world
of before the Mid Pleistocene Transition (0.60.9 Ma).
The spectral analysis reveals strong, even dominant influence of obliquity in their record (Larrasoana et al.
2003), also linking low dust input to high obliquity.
They argue that high obliquity might enhance the dust
flux through an increased meridional temperature gradient in the Southern Hemisphere during austral winter. Additionally, they pose that a critical watershed is
activated during periods of higher obliquity, making
the NAM susceptible to a threshold-type response to
what would otherwise seem a minor latitudinal change
in the African summer monsoon penetration (Larrasoaa et al. 2003). This seems possible, although the
highly linear relationship between our Ti/Al record
and insolation speaks against the involvement of a
threshold.
Another possible low latitude mechanism explaining
a direct influence of obliquity on the low latitude is proposed by Lourens and Reichart (1996). They reasoned
that the northward shift of the ITCZ during monsoon
season expands the Hadley circulation in summer,
which would reach from the tropic of Capricorn at 23 S
and the ITCZ, perhaps as far as 23 N during strong
monsoons. The strength of this monsoon is regulated by
the contrast between the insolation at 23 S and 23 N.

273

Lourens and Reichart argue that a target curve of the

insolation differences between 23 N and 23 S introduces a significant amount of power in the obliquity
range and matches the thick/thin pattern of sapropels
much better than the monsoon index of RossignolStrick (1983). This idea, based on the functioning of
the present day monsoon system, has recently gained
support from modeling results exploring the effect of
obliquity on low latitude climate (Bosmans 2014).

6.

Conclusions

We constructed a high-resolution composite record

by combining data from ODP 967 with the adjacent
site 968. Using X-Ray Fluorescence we produced a
high resolution geochemical dataset. The high level
of agreement between the Ti/Al ratio (proxy for monsoon activity) and the Sanbao/Hulu caves (Cheng et al.
2009) allows us to establish the relation between North
African Monsoon activity and insolation. Using this
relation we create a new and highly accurate age model for the composite record and the sapropels therein
through orbital tuning of the Ti/Al record.
Our age model for the sapropels in ODP 967/968 is
different from the previously published chronology of
Emeis et al. (2000). In most cases our chronology
agrees with the ATNTS2004 (Lourens et al. 2004) and
the chronological framework of Lourens (2004) for
ODP 964. In the interval between 700950 ka where
changes in insolation forcing are generally small and
sediments are devoid of true sapropel formation the
Ti/Al proxy behaves very differently from the color
reflectance values, or red intervals that are otherwise
used to mark missing sapropels. In these cases of conflicting proxy data the Ti/Al proxy is given preference
because it has the simplest operational mechanics and
linearly resembles the insolation curve. The different
behavior of this proxy leads to major revisions of the
existing age model of Emeis et al. (2000).
Throughout our record we find significant influence
of obliquity. Cross spectral analysis shows a lag of
~ 4 0.7 kyr to obliquity maxima, pointing at a direct
response to insolation rather than a teleconnection to
the high latitudes, with a time lag that is intrinsic to the
climate system and only minor influences from other
sources (i. e. ice volume). A possible mechanism introducing obliquity power in these low latitudes is using
the gradient between 23 N and 23 S June 21 insolation, as opposed to just low latitude northern hemisphere insolation as the driver for monsoon strength.

eschweizerbart_xxx

274 T. Y. M. Konijnendijk et al.

0.006

41
kyr

23
kyr

19
kyr

0.004
40K

0.002

20K

BT power insolation
65N June 21st

BT power Ti/Al

60K

Arctan Coherency
Ti/Al vs insolation

0.8

95%

0.6

90%
80%

0.4

0.2

1.5

Phase (rad)
Ti/Al vs insolation

1.0
0.5
0.0
-0.5
-1.0
-1.5

Fig. 6. Top panel: power spectrum of the tuned Ti/Al record, enveloped by the confidence limits at 80, 90, an 95%, and
the power spectrum of the target insolation curve in red (grey in printing version). The distribution of power is similar, with
a dominant spectral peak at 23 kyr and smaller ones at 41 and 19 kyr reflecting the linear behavior of Ti/Al in response
to insolation. Middle panel: coherency between the Ti/Al record and insolation. Only at the dominant frequencies does the
coherency reach the 95% confidence boundary. Bottom panel: Phase difference between the Ti/Al record and insolation, in
radials. The high coherency translates into a narrow uncertainty envelope at the dominant spectral frequencies. The phase
difference corresponds to a time lag of 3.9 kyr, 2.5 kyr, and 1.7 kyr for the orbital periods of 41 kyr, 23 kyr, and 19 kyr,
respectively.

eschweizerbart_xxx

Chronological constraints on Pleistocene sapropel depositions

275

Laskar, J., Joutel, F., Boudin, F., 1993. Orbital, precessional, and insolation quantities for the Earth from 20 MYR
to 10 MYR. Astronomy and Astrophysics 270, 522533.
Lisiecki, L. E., Raymo, M. E., 2005. A PliocenePleistocene
stack of 57 globally distributed benthic D18O records.
Paleoceanography 20, PA1003.
Lourens, L. J., Reichart, G.-J., 1996. G. Low-latitude forcing
of glacial cycles. G.-J. Reichart (PhD thesis), Geologica
Utraiectina 154, 153168.
Lourens, L. J., 2004. Revised tuning of Ocean Drilling Program Site 964 and KC01B (Mediterranean) and implications for the 18O, tephra, calcareous nannofossil, and
geomagnetic reversal chronologies of the past 1.1 Myr.
Paleoceanography 19, PA3010.
Lourens, L. J., Wehausen, R., Brumsack, H. J., 2001. Geological constraints on tidal dissipation and dynamical
ellipticity of the Earth over the past three million years.
Nature 409, 10291033.
Lourens, L. J., Antonarakou, A., Hilgen, F. J., van Hoof,
A. A. M., Vergnaud-Grazzini, C., Zachariasse, W. J., 1996.
Evaluation of the Plio-Pleistocene astronomical timescale.
Paleoceanography 11, 391413.
Paillard, D., Labeyrie, L., Yiou, P., 1996. Macintosh Program performs time-series analysis. Eos Transactions
American Geophysical Union 77, 379.
Postma, G., Hilgen, F. J., Zachariasse, W. J., 1993. Precession punctuated growth of a late Miocene submarine fan
lobe on Gavdos (Greece). Terra Nova 5, 438444.
Rossignol-Strick, M., 1983. African monsoons, an immediate
climate response to orbital insolation. Nature 304, 4649.
Rossignol-Strick, M., Paterne, M., 1997. A synthetic pollen
record of the eastern Mediterranean sapropels of the last
1 Ma: implications for the time-scale and formation of
sapropels. Marine Geology 153, 221237.
Sakamoto, T., Janecek, T., Emeis, K.-C., 1998. Continuous
sedimentary sequences from the eastern Mediterranean
Sea: composite depth sections 160, 3759.
Wang, Y., Cheng, H., Edwards, L. R., Kong, X., Shao, X.,
Chen, S., Wu, J., Jiang, X., Wang, X., An, Z., 2008.
Millennial- and orbital-scale changes in the East Asian
monsoon over the past 224,000 years. Nature 451, 1090
1093.
Wehausen, R., Brumsack, H. J., 2000. Chemical cycles in
Pliocene sapropel-bearing and sapropel-barren eastern
Mediterranean sediments. Palaeogeography, Palaeoclimatology, Palaeoecology 158, 325352.
Weltje, G., de Boer, P. L., 1993. Astronomically induced paleoclimatic oscillations reflected in Pliocene turbidite deposits on Corfu (Greece): implications for the interpretation of higher order cyclicity in ancient turbidite systems.
Geology 21, 307310.
Ziegler, M., Tuenter, E., Lourens, L. J., 2010. The precession
phase of the boreal summer monsoon as viewed from the
eastern Mediterranean (ODP Site 968). Quaternary Science Reviews 29, 14811490.

Supplementary data are available in the Pangaea

database under doi 10.1594/PANGAEA.831712
Acknowledgements. This project was nancially supported by NWO-ALW (project number 865.10.001). We
kindly acknowledge the Ocean Drilling Program in general,
and Walter Hale and the Bremen Core Repository specifically for the samples used in this study. Thanks to two
anonymous reviewers for valuable comments.
This work was initiated by L. J. Lourens. Lab work and
analyses were performed by M. Ziegler and T. Y. M. Konijnendijk. T. Y. M. Konijnendijk wrote the manuscript, with
contributions of the other authors.

References
Bosmans, J. H. C., 2014. A model perspective on orbital forcing of monsoons and Mediterranean climate using ECEarth (Utrecht Studies in Earth Sciences 055). Utrecht:
UU Depts. of Physical Geography and Earth Sciences.
Cheng, H., Edwards, R. L., Broecker, W. S., Denton, G. H.,
Kong, X., Wang, Y., Zhang, R., Wang, X., 2009. Ice Age
Terminations. Science 326, 248252.
Claussen, M., Bathiany, S., Brovkin, V.& Kleinen, T., 2013.
Simulated climate-vegetation interaction in semi-arid regions affected by plant diversity. Nature Geoscience 6
(11), 954958.
Ehrmann, W., Seidel, M., Schmiedl, G., 2013. Dynamics of
Late Quaternary North African humid periods documented in the clay mineral record of central Aegean Sea sediments. Global and Planetary Change 107, 186195.
Emeis, K. C., Robertson, A., Richter, C., and the shipboard
scientific party, 1996. Proceedings of the Ocean Drilling
Program, Initial Reports. ODP, College Station (TX).
Emeis, K., Sakamoto, T., Wehausen, R., Brumsack, H. J.,
2000. The sapropel record of the eastern Mediterranean
Sea results of Ocean Drilling Program Leg 160. Palaeogeography, Palaeoclimatology, Palaeoecology 158, 371
395.
Hilgen, F., 1991. Astronomical calibration of Gauss to Matuyama sapropels in the Mediterranean and implication for
the geomagnetic polarity time scale. Earth and Planetary
Science Letters 104, 226244.
Hilgen, F., Lourens, L., Berger, A., Loutre, M. F., 1993.
Evaluation of the astronomically calibrated time scale for
the late Pliocene and earliest Pleistocene. Paleoceanography 8, 549565.
Langereis, C., Dekkers, M., de Lange, G. J., Paterne, M.,
Sandvoort, P. J. M., 1997. Magnetostratigraphy and astronomical calibration of the last 1.1 Myr from an eastern
Mediterranean piston core and dating of short events in the
Brunhes. Geophysical Journal International 129, 7594.
Larrasoana, J., Roberts, A., Rohling, E. J., Winklhofer, M.,
Wehausen, R., 2003. Three million years of monsoon
variability over the northern Sahara. Climate Dynamics
21, 689698.

Manuscript received: March 10, 2014; rev. version accepted:

May 5, 2014.

eschweizerbart_xxx

276 T. Y. M. Konijnendijk et al.

Appendix I
Color reflectance for the composite splice of ODP 967 (top graph). The core intervals used from the individual
holes are indicated in black. Unused intervals are in grey. The overall correlation between cores is very high on
the revised depth scale (2013). To construct the splice we attempted to minimize switches between holes and
maximize the length of the intervals used from a single core. Motivations to switch can be either the end of a
core segment or deviating CR values.

Composite

CR 550 nm

12
20
28

Hole A1

A2

Hole B1

B2

Hole C1

C2

Hole D1

8
Depth RMCD2013

eschweizerbart_xxx

10

12

14

Chronological constraints on Pleistocene sapropel depositions

CR 550 nm

12

Composite

20
28
A3

A2

B2
B3

C3

C2

10

12

14

16

18

20

22

24

Depth RMCD2013

Composite
CR 550 nm

12
20
28
A4

A3

B3

B4

C4
C3

20

22

24

26

28
Depth RMCD2013

eschweizerbart_xxx

30

32

34

277

278 T. Y. M. Konijnendijk et al.

Composite
CR 550 nm

12
20
28
A4
A5

B4

B5

C4
C5

30

32

34

36

38

40

42

44

Depth RMCD2013

Composite

CR 550 nm

12
20
28

A5

B6
B5

C6

C5

40

42

44

46

48
Depth RMCD2013

eschweizerbart_xxx

50

52

54

Chronological constraints on Pleistocene sapropel depositions

Appendix II
Revised splice for ODP 967 (2013).
site

hole

core

type

section

top

mbsf

rmcd2013

967
967
967
967
967
967
967
967
967
967
967
967
967
967
967
967
967
967
967
967

D
D
A
A
B
B
C
C
A
A
B
B
C
C
B
B
C
C
B
B

1
1
1
1
2
2
2
2
3
3
4
4
4
4
5
5
5
5
6
6

H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H

1
5
4
5
1
4
1
7
1
6
1
4
1
6
2
5
2
5
2
5

2
6
144
138
108
44
16
52
100
114
114
58
106
10
38
74
2
148
114
94

0.02
6.06
5.94
7.38
6.38
10.24
9.66
19.02
19.80
27.44
25.44
29.38
29.56
36.10
35.68
40.54
39.52
45.48
45.94
50.24

0.020
5.915
5.915
7.382
7.382
11.324
11.324
21.039
21.039
28.699
28.699
33.050
33.050
39.722
39.722
44.497
44.497
51.097
51.097
54.949

eschweizerbart_xxx

279

280 T. Y. M. Konijnendijk et al.

Appendix III
Tie points used to fit the composite of ODP 968 to the revised ODP 967 splice and merge the data into a continuous dataset.
968 tie points

967 tie points

correlated
feature

Hole

Core

Sec

Top
(cm)

Depth
(mbsf)

Hole

core

type

section

top

mbsf

967composite

Bottom S1
Top S3
Bottom s3
Bottom s4
Bottom s5
Precursor s5
Bottom s6
Bottom s7
Bottom s8
Top s9
Top S
Top s10

Bottom s11

Top s12
Mid s12

Mid sa

Top sb
Bottom sb

A
A
A
A
A
B
B
B
B
B
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A
A

1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3

H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H

1
1
4
4
5
2
2
4
5
6
4
5
6
7
CC
1
1
1
2
3
3
3
3
4
4
4
4
5
5
5
6
6
6
CC

6
86
92
128
94
44
92
106
102
76
38
128
106
80
18
72
100
148
140
32
100
102
134
20
40
104
112
44
64
94
116
128
148
20

0.06
0.86
5.42
5.78
6.94
7.34
7.82
10.96
12.42
13.66
13.88
16.28
17.56
18.80
18.99
19.22
19.50
19.98
21.40
21.82
22.50
22.52
22.84
23.20
23.40
24.04
24.12
24.94
25.14
25.44
27.16
27.28
27.48
28.34

D
D
D
D
D
B
B
B
B
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
A
A
A
A
A
A
A
A
A

1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3

H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H

1
1
3
4
4
1
2
3
4
1
1
2
3
4
4
5
5
5
6
6
6
6
6
7
7
1
1
2
2
2
3
3
3
4

68
126
118
20
98
134
2
42
16
68
116
110
114
44
56
70
92
122
30
58
96
98
134
2
22
124
130
52
68
106
102
112
128
34

0.68
1.26
4.18
4.7
5.48
6.64
6.82
8.72
9.96
10.18
10.66
12.1
13.64
14.44
14.56
16.2
16.42
16.72
17.3
17.58
17.96
17.98
18.34
18.52
18.72
20.04
20.1
20.82
20.98
21.36
22.82
22.92
23.08
23.64

0.68
1.26
4.18
4.70
5.48
7.76
7.94
9.84
11.08
11.88
12.36
13.80
15.34
16.14
16.26
17.90
18.12
18.42
19.00
19.28
19.66
19.68
20.04
20.22
20.42
20.96
21.02
21.74
21.90
22.28
23.74
23.84
24.00
24.56

eschweizerbart_xxx

Chronological constraints on Pleistocene sapropel depositions

281

Appendix IV
Table of the turbidites in ODP 967 as inferred from the Ti/Al record. Note that each of the identified turbidites
is indicated in the top part of Fig. 3.
Turbidites

mbsf

rmcd2013

site

core

sect.

top

bot

top

bot

top

bot

MIS

Age (ka)

967
967
967
967
967
967
967
967
967
967
967

A3
A3
A3
A3
B4
B4
B4
B4
B4
C4
C4

5
5
5
6
1
3
3
3
4
2
2

17
23
91
71
131
7
109
147
3
5
62

20
41
99
101
145
15
125
149
15
11
66

24.97
25.03
25.71
27.01
25.61
27.37
28.39
28.77
28.83
30.05
30.62

25.00
25.21
25.80
27.31
25.75
27.45
28.55
28.79
28.95
30.11
30.66

26.23
26.28
26.96
28.27
28.91
30.85
31.98
32.40
32.45
33.51
34.06

26.25
26.46
27.05
28.57
29.06
30.94
32.16
32.42
32.60
33.60
34.09

17/18
17/18
18.1
19/20
20
21/22
23
23/24
23/24
25
25/26

709.0
714.0
736.9
785.7
804.7
864.9
902.2
911.9
915.3
942.1
960.9

eschweizerbart_xxx

282 T. Y. M. Konijnendijk et al.

Appendix V
Shipboard core photo of ODP 967 core A4 (http://iodp.tamu.edu/janusweb/imaging/photo.shtml)

eschweizerbart_xxx