aus dem Fachbereich Geowissenschaftender Universität Bremen
Nr.135
Schmieder, Frank
MAGNETIC CYCLOSTRATIGRAPHYOF SOUTH ATLANTIC SEDIMENTS
~~~L--___________________ \
Berichte, Fachbereich Geowissenschaften, Universität Bremen, Nr. 135,82 Seiten, Bremen 1999
ISSN 0931-0800
Die "Berichte aus dem Fachbereich Geovvissenschaften" werden in unregelmäßigen Abständen vom
Fachbereich 5, Universität Bremen, herausgegeben.
Sie dienen der Veröffentlichung von Forschungsarbeiten, Doktorarbeiten und wissenschaftlichen Beiträgen. die
im Fachbereich angefertigt vvurden.
Die Berichte können bei:
Frau Gisela BoeIen
Sonderforschungsbereich 26 I
Universität Bremen
Postfach 330 440
D 28334 BREMEN
Telefon: (49) 421218-4124
Fax: (49) 421218-3116
angefordert werden.
Zitat:
Schmieder. F.
Magnetic Cyclostratigraphy of South Atlantic Sediments.
Berichte, Fachbereich Geowissenschaften, Universität Bremen, Nr. 135,82 Seiten, Bremen, 1999.
ISSN 0931-0800
Magnetic Cyclostratigraphy
of South Atlantic Sediments
Dissertation
zur Erlangung des
Doktorgrades der Naturwissenschaften
im Fachbereich Geowissenschaften
der Universität Bremen
vorgelegt von
Frank Schmieder
Bremen 1999
Tag des Kolloquiums:
12.2.1999
Gutachter:
Prof. Dr. U1rich BleilProf. Dr. Gerold Wefer
Prüfer:
Prof. Dr. Volkhard SpießProf. Dr. Heinrich Miller
"So ergibt sich ... ein Bild des Eiszeitalters, welches sich mit den Ergebnissen der geologischen Forschung vollständig deckt. Mit dieser erfreulichen Schlußfolgerung soll nichtdas Bekenntnis unterdrückt werden, daß durch die vorliegende Lösung des Eiszeitenproblems ein neues Problem entstanden ist, nämlich jenes, warum der säkulareBestrahlungsgang, welcher sich durch alle Zeugnisse des Quartärs so unzweideutig zuerkennen gab, während der vorhergehenden geologischen Zeiten in Europa nicht zusolchen Vereisungen geführt habe wie während der letzten 800 Jahrtausende."
Milutin Milankovitch, 1936
Contents
Page
Chapter 1 Introduction
Milankovitch theory and cyclostratigraphy
Main objectives
Individual studies
References
3
5
8
Chapter 2 Using rock magnetic proxy records for orbital tuning and extended 11
time series analyses into the super- and sub-Milankovitch bands
T. von Dobeneck and F. Schmieder
Chapter 3 Mid-Pleistocene climate transition: initiation, interim state and termi- 44
nal event
F. Schmieder, T. von Dobeneck and U. Bleil
Chapter 4 Cycles, trends and events of Pleistocene sedimentation in the oligotro- 49
phic subtropical South Atlantic Ocean
F. Schmieder, T. von Dobeneck and U. Bleil
Chapter 5 Terrigenous flux in the Rio Grande Rise area during the past 1500 ka: 69
evidence of deepwater advection 01' rapid response to continental rain-
fall patterns?
F.X. GingeIe, F. Schmieder, 1'. von Dobeneck, R. Petschick and C. Rühlemann
Chapter 6 Summary and perspectives 81
Introduction
Milankovitch theory and cyclostratigraphy
The biggest breakthrough in understanding the
Pleistocene ice ages was the discovery that climatic
changes are strongly inf1uenced by orbitally induceel
variations in solar insolation received on Earth. AI
though the seminal idea has been expressed earlier
this theary is inseparably associated with the name
of Milutin Milankovitch (1879-1958). He was the
first to compute in great mathematical detail the ele
ments of Earth' sorbit and their effect on changes
in insolation (Figure 1). His complete astronomical
theory of Pleistocene ice ages (e.g., Milankovitch,
1936) fundamentally supported the basic idea that
cold summers in northern high latitudes are neces
sary for the builel up of]arge ice shields (for a his
torical review ofthe astronomical theory ofpaleo
climate see Milankovitch, 1936; Berger, 1988 and
Schwarzacher, 1993a).
Variations in insolation mainly result from
changes in obliquity of the Earth's rotational axis
anel in precession ofthe equinoxes (Figure 2). Hays
et a1. (1976) showed that the corresponeling perioels
around 41 (obliquity), 23 and 19 kyr (precession)
can be identifieel in oxygen isotope recorels of ma
rine sediments tagether with the near 100 kyr icc
age cycle dominating the late Pleistocene. For each
of these periods thc coherency betwecn orbital and
isotopic signals is statistically highly significant
(lmbrie et a1., 1984; Figure 3). The discovery that
paleoclimatic variations can be traced to exactly
preelictable periodical fluctuations of the Earth 's
orbit opened formerly unequaled possibilities far
detailed age modeling. Toelay, the once visionm'y
idea of cyclostratigraphy is realized by quantita
tively corrclating pa1eoclimate signals or filtered
components to target recards constructed from or
bital variations. During the last two decades this
'astronomical tuning' has proven to be one of the
most powerful tools for high resolution chrollOstrati-
Jonrlousende
qoc!,,{}(}(J !J5(} $00 tJ:IJ. t5~ T 7(}(J 550 i 550
1, ~ ~~ ~r ~ i ~ I t~ tU\? T 0
J:~ ,~,~.~,'\!~~ ·~~~~tts:~~~II@;i9;-c-·~, ~j~~;~ij;i::~ -j\I~ YI\' 'nv v ',J.. AA,"'\'!l1 f\.... ! ,Vr\.'II'fr-J r W\J'MVr\. i~rvT"AV r-.. V7\
~ 7ßO :'" y~; -y :tt-~~-,v.. --I\-, -t.~ \l T f--·\; -1 ~ ~, :" VI ~ .Y-75° I 1l 11 JY F 1'l J'JJ J1![ JX Ka.t Xl X!a .n .IlIl X1Y
Pliozün? OOll{](J·fiueil wnz-fiszeil Minde/-fisuil ffiß-fiszeif Wiif'fT1-füzeilFig. 16. StrahlungBdiagramm von M. Milsllk{)"it('h
.4bsfiindi' der Ab/agu{j{7gen in f'f!!a/ivtn gwhgisckn J#rfen. 1 2 3 q 5 6 7 8 910 11 I1CL 12 13 N 1$ J6i ~ ~~~H----Hf-~~':(
':'~i[-tl ~1'" ... ---- tt--- :f--------..---.- - a-:-lUY''/----/---Pl' '-MIi-tl~H1':fil4-·...,
'" C' --'11 • • •. '!'.......- ------lll---f-l'------IIII'--Iil!... ,,1,'":::..; 'l ?iJ _.-~ I -------1--~·-·- ~ . ~l , , 1-- ; 11' w
~] 3IJ .--' , .~--+-- -1- -----lil----11h------+---H---I,i+-,-:---l,
~~ qo!--~ : ~ I I ' .. ------:----.;------ ' -0-----r-~lro_:+.~.~ I I ~ 1 I i ( --r-'! T I !I \.~ ",>?f----r ---t-----+--r-r I I -+-----·-·------rw;r;;;;~m ii il~ &7f--1 ~-j-+++---._-~ --1·,-+----··· , I' ::.='-----!--~,~TI~,-l'<:; 70 A !i 01 Oll 0lJ[ 51 5l[ MI MIl 111 IIIlIlla. WI W1l WlI!
~er Sfarrfenbtrg lJonou·Sftldim 6ünz·Sfadien Hindel-SfCldien 1I1f)-J'lodien Würm-J'ladl",SdxJIfer ,>dKJ/Ier
Fig. 17. Stratigraphisches Diagmmm "Oll 13. Eb,,!
Figure 1. Milutin Milankoviteh's radiation estimates far latitude 50,55 and 65°N for the past 1000 kyr (top) eompared to astratigraphie interpretation ofEuropean iee ages (Würm, Riß, Mindel, Günz, Donau) by B. Eber! (bottom). From Köppen &Geiger (1936).
Figure 2. Insolation received on Earth depends on severalelements ofthe Earth's (E) orbit around the sun (S) varyingdue to gravitational effects of the planets and the moon.Eccenh'icitye = .j(;'-!/ / ([ defines the elliptical shape oftheorbit (presently e = 0.016). Obliquity (E, presently 23°27')measures the angle between the Earth's axis ofrotation (SN)and the perpendicular to the ecliptic (SQ) and is mainlyresponsible for the seasonal contrasts and the latitudinalgradient ofinsolation. WW and SS denote winter and summersolstice, respectively, y indicates the vernal equinox. Theinclination E ofthe Earth's axis follows a complex precessionalmovement along the cone shaped surface that produces amigration of the equinoxes scaled to the longitude 01' theperihel ion co. Today, the Earth is about at the perihelion (P,closest approach between Earth and Sun) during the northernhemisphere winter solstice. From Berger (1988).
graph1', especiall1' for the late Pliocene and Pleisto
cene but b1' no means restricted to these epochs (e.g.,
Schwarzacher, 1993a, b; Shackleton et al. , 1995;
D'Argenio et al., 1998).
The tard1' triumph of the Milankovitch theor1'
over long-standing criticism (see Berger, 1988) is
documented impressivel1' in the delayed acceptance
ofthe orbitall1' tuned age ofthe Brunhes/Matu1'ama
boundary. Although earlier proposed b1' Johnson
(1982), the age of 780-790 ka became widel1' ac
cepted onl1' after the 'second attempt' ofShackleton
et al. (1990) and Hilgen (1991). Cande and Kent
(1992) for example revised their magnetic polarit1'
time scale in 1995 (Cande and Kent, 1995) and made
it consistent with astrochronolog1', also in older
sections of the record. In the meantime, Wilson
2
N Q
Chapter 1
(1993) showed that astronomical calibration resldts
in a more concordant sea-floor spreading history
when applied to spacings of magnetic anomal1'
pattern in the Pacific Ocean. Still assuming an age
of 730 ka for the Brunhes/Matu1'ama boundar1'
and therefore somewhat inaccurate in thc older part
- Imbrie et al. (1984) published their epoch-making
SPECMAP time scale for the last about 800 k1'r and
started a s1'stematic documcntation of the Earth' s
paleoclimatic history which since then has been
extended and refined graduall1' by astronomicall1'
calibrated 8 180 time scales (e.g., Ruddiman et al.,
1986; Martinson et al. , 1987; Ruddiman et al. , 1989;
Raymo et al., 1989; Shackleton et al. , 1990;
Tiedemann et al. , 1994; Berger and Jansen, 1994a).
The ratio ofox1'gen isotopes IBO and 1(,0 has been
applied as a paleoclimatic indicator far quite some
time (e.g., Emiliani, 1955). In most cases it is mainly
a measure of global ice volUll1e (e.g., Shackleton,
1967). In the context of cyclostratigraph1' ox1'gen
isotopes are b1' far the most frequentl1' used and
best studied prox1'. But measuring detailed 8 1BO
records is very time consuming and involves an
immense laborator1' work. In low sedimentation
rate, predominantl1' terrigenaus and/or deposits
below the CCD (carbonate campensation depth)
detailed anal1'ses may be seriousl1' hampered as it
is necessar1' to coHect a sufficient number ofcalcare
ous tests from each sampie. Consequentl1', mainl1'
cores from regions with high (carbonate) sedimen
tation rates are used to achieve detailed paleo
climatic records. This strateg1' excludes large parts
ofthe worlds oceans holding valuable paleoenviron
mental infonnation.
The outlined restrictions ofox1'gen isotope stra
tigraph1' led marine scientists to explore adequate
alternatives. Since the discover1' that paleoclimatic
patterns are often mirrored b1' ph1'sical properties
like magnetic susceptibilit1' (e.g., Robinson, 1990;
Tm-duno et al. , 1991) and GRAPE (Gamma Ra1'
Introduction 3
-- - - -- 6'1'0 VAR 1I - - - ETP VAAff -+-+-+-+ COHE RENCY
Figure 3. Coherency and variance spectra calculated fromrecords of clil11atic and orbital fluctuations spanning the last780 kyr (Il11brie et a1., 1984). The ETP signal cOl11bines normalized variations of~ccentricity, obliquity ([iIt) ancl ~recession. 8 180 is the unsl1100thed SPECMAP stack. From Bergeret a1. (1984).
10
-co
"oci~
.c
100 41 1
,/\ /\ ?' I_~ I g• -" '\ I .'. \ I .... ' ( \ I",' "'-', I,' \1 I,' " \ I 1 ;':" \ ,'J/ \1 1/ \ \ I \ 1 V)
\i I \';;""\f '\V,.... \, I, ,-~ BA"DW!DT>i I ~\1 11 I I '! "1', I cu
o I 11 11 \l, ~ .", " I~z. 11 \( I, "'" "'-.::., ". '"i;! ',' 1,1 '~--' - .... ,,~/~ ~\:- :,~, ~w 0.9 .....
6 "---+-+----:r-+-----t~-;:t_-r -j-",/5"'%--:-SiQ. level (wh) I >U 0.8 r -1
07 I
~~ -.-~~001 0.02 003 004 005 0.06 007 0.08 009 0 I100 50.0 33.3 250 200 167 143 12.5 \ I: \00
Attenuation Porosity Evaluator) density records
(e.g., Herbert and Mayer, 1991; Grützner et a1.,
1997) numerous chronostratigraphies have been es
tablished on basis ofthese continuously measurable
proxies (e.g., Shackleton et a1., 1995; Chi and
Mienert, 1996; ShackJeton and Crowhurst, 1997;
Bickert et a1., 1997). The measurements are very
rapid and nondestructive and the high resolution
reachable makes these parameters especially
suitable both for the analyses of high frequency
variability and low sedimentation rate deposits.
Main objectives
During recent years Marine Geopbysics in tbe De
partment ofGeosciences at Bremen University suc
cessfully establisbed magnetic cyclostratigrapby as
a very efficient dating tool for marine sediment se
quences in addition to long-proven magnetostrati
graphic technique (e.g., Bleil and Gard, 1989;
Nowaczyk, 1991; Thießen, 1993; Nowaczyk et a1.,
1994; Frederichs, 1995; Gersondeeta1., 1997). The
age models for tbe sediment series presented in this
thesis are all based on orbitally tuned high-resolu
tion magnetic susceptibility records, partly C0111
bined with oxygen isotope data 01' based on Quater
nary magnetostratigraphies. Advantages, difficulties
and restrictions of using rock magnetic records for
dating purposes are discussed in full detail in Chap
tel' 2.
Although today Milankovitch's theory ofa link
age between Earth's climate and orbital variations
is widely accepted and successfully applied chrono
stratigraphically, the history ofPleistocene climate
variability still holds l1Umerous unresolved puzzles.
To resolve its cyclicity, located within the main but
also in the adjacent frequency bands above and
below the principal Milankovitch frequencies, re
quires particularly adapted statistical methods.
Extensions of time series analyscs into the super
and sub-Milankovitch bands are described in Chap
tel' 2.
The central paleoceanographic topicofthis thesis
(Chapters 2, 3 and 4) is the mid-Pleistocel1e tran
sition (MPT) of tbe global climate system (Pisias
and Moore, 1981; Prell, 1982; Ruddiman et a1. ,
1989, Berger and Wefer, 1992; Berger and Jansen,
1994b), the change from a predominantly 41 kyr
cyclicity in early Pleistocene to the late Quaternary
100 kyr ice age cycles. Variance in climate indices
at periods of precession of the equinoxes (with
periods in the range of 19 to 24 kyr) and obliquity
01' tilt (with major cycles around 41 kyr) can be
explained in the framework ofMilankovitch theory
as linear responses to changes in solar insolation
(e.g., Imbrie et a1., 1992). In contrast, the origin of
the 100 kyr ice age cycle dominating the late Pleis
tocene (Figure 3) calls for more complex explana
tions because the direct influence of eccentricity
(with major periods at about 95 and 124 kyr in ad
dition to a somewhat stronger 413 kyr component)
on insolation is by far too small to produce the cor
responding climate style (e.g., Imbrie et a1., 1993).
4 Chapter 1
~:~ ~~=loowMo 100 200 300 400
.------- t ---------..ECCEHffilCrTY BAND
HOOkyr)
]1------------OBUQlJITY BAND
(......11<yr)15
-15
~t~]-~~6180
o 100 200 300 400AGE (ko)
7S
...~
-15
-1
8"-0
Figure 4. The 100 kyr cycle problem. Partitioning radiation (top, data from Berger, 1978) and climate time series (bottom,data from Imbrie et al., 1984) into their dominant periodic cOl11ponents reveals a large discrepancy in the eccentricity band.While the three frequencies contribute to the insolation signal in proportion to 1,0.2 and 0.02, the respective ratios for climatevariations al110unt to 1, 2.5 and 11. From Il11brie et al. (1993).
This discrepancy between driving force and re
sponse is illustrated in Figure 4.
Besides its pure existence, another enigma is the
more or less abrupt onset of the 100 kyr cycle ap
proximately 650 kyr aga (e.g., Berger et aL, 1994;
Mudelsee and Schulz, 1997). Build up ofmajor ice
shields on the northem hemisphere already started
at about 2.5 Ma (e.g., Shackleton et al., 1984). While
late Pliocene to early Pleistocene paleoclimate
records reveal a mainly obliquity and precession
related variance (e.g., Raymo et al. , 1989; Ruddiman
et al. , 1989; Bloemendal and deMenocal, 1989),
variations with aperiod near 100 kyr are the primary
rhytbm oflate Pleistocene climatic change although
insolation vm1ations remained almost identical. The
exact timing of the MPT and the question whether
the transition from 41 to 100 kyr cyclicity was gra-
dual ('mid-Pleistocene evolution' ,Ruddiman et al. ,
1989) or abrupt ('mid-Pleistocene revolution', Ber
ger and Wefer, 1992) were subject ofmultiple stud
ies and attempts to model this shift with different
statistical tec1miques (e.g., Maasch, 1988; DeBlonde
and Peltier, 1991; Park and Maasch, 1993; Mudelsee
and Schulz, 1997; Mudelsee and Stattegger, 1997;
Clark and Pollard, 1998). Of the diverse explana
tions proposed for the 100 kyr cycle, models invok
ing a pC02
controlled insolation threshold for the
melting oflarge ice shields (e.g., Saltzman and Ver
bitsky, 1993; Berger and Jansen, 1994b) currently
yield the best approximation of8180 signal pattem.
As illustrated in Figure 5, these models accomplish
a reconstruction ofthe MPT timing and the develop
ment of 100 kyr cycles by introducing a decreasing
atmospheric pC02level (Raymo, 1997; Paillard, 1998).
Introduction 5
~
0)()'1
Q
400 Z
~3-'"
.....c:
450 ..:<I'\l-
, 500 ~2llJc::',
g
""""",,,j, , I ,:, , , I , ,', , I , ; , ,I ,,' I ,:, , , I ;, , ,I '" I , ,,I ;" I:
Wisconsin lilinoian Kansan NebraakanWürm Riss Mindei Gim
I I IV V VII3
'"Il \
I0 4 I
'" \ (-<0
I I
-.~\ I
5 -'2
-200 o 200 400 600Aga (kyr)
800 1000 1200
Figure 5. Benthie oxygen isotope data from Paeifie ODP Site 849 eompared to insolation at 65°N on Jllly 21, The horizontaldashed line in the alsO reeord indieates the level belO\v whieh 'exeess 100 kyr iee' is observed during glaeials. The vertiealdotted lines mark the midpoint ofterminations. Assllming a long-term drawdown ofthe atmospherie pC0
2level, olltlined by
the sloping line in the insolation reeard, results in a ehanging threshold far the melting of large iee shields. During latePleistoeene a suffieiently high insolation is only attained at times of positive interaetion of obliquity and preeession at the endof the blaek shaded regions, Beeallse the amplitude of preeession is modlliated by eeeentrieity, 100 kyr eycles are produeed,eoineiding with several of the tenninations. From Raymo (1997).
With a few exceptions (e.g., Raymo et al., 1997)
so far all investigations of the MPT were based on
oxygen isotope records. The rock magnetic view
presented in this thesis adds some completely new
aspects for the understanding of this fundamental
enigma of Quaternary climate evolution. Aseries
of sediment sequences from submarine ridges of
the subtropical South Atlantic Ocean located in a
water depth alternately affected by North Atlantic
Deep Water (NADW) and Lower Circumpolar Deep
Water (LCDW) have been studied which recorded
the paleoceanographic history in a region of strate
gie importance for the global thermohaline circula
tion (e.g., Berger and Wefer, 1996).
Individual studies
This thesis comprises four manuscripts which are
in press or have been submitted for publication. The
studies were performed in the Marine Geophysics
Section of the Department of Geosciences in the
framework of the Graduiertenkolleg 221 "Stoff
Flüsse in marinen Geosystemen", associated with
the Sonderforschungsbereich 261 "Der Südatlantik
im Spätquartär: Rekonstmktion von Stoffhaushalt
und Stromsystemen". Both research projects are
funded by the Deutsche Forschungsgemeinschaft.
The work has been supervised by Prof. Dr. Ulrich
Bleil.
A stratigraphie synthesis of twelve sediment
cores covering the deep subtropical South Atlantic
Ocean was established using magnetostratigraphic
and cyclostratigraphic methods. This regional
6 Chapter 1
chronostratigraphic framework fonDs the basis for
various research strategies and scientific co-opera
tions - most intensely with Dr. Tilo von Dobeneck,
co-author of all four papers.
A first, methodological treatise intro duces
advanced signal analytical techniques enabling
paleoceanographic interpretations in the super- to
sub-Milankovitch ranges. Their immense potential
is exemplified by susceptibility records from the
subtropical and western tropical South Atlantic, the
latter dated by von Dobeneck (1998).
In the course of the mid-Pleistocene climate
transition (MPT) the dominant periodicity of the
climate system response changed from 41 to
100 kyr. Rock magnetic investigations identify this
period as an interim state of reduced carbonate
accumulation in thc subtropical South Atlantic. The
second paper interprets the MPT as aseparate cli
mate state ofreduced NADW influence, tem1inated
by an ocean-wide event at 530 ka.
The third manuscript compares the temporal evo
lution ofclimatic cycles in environmentalmagnetic
proxy records to driving orbital variations. Recipro
cal exchange of spectral energy between the 41 and
100 kyr bands suppOlis models invoking a pC02
controlled insolation threshold for the occurrence
of 100 kyr glaciation cycles in late Pleistocene.
The final study is concerned with the tenigenous
input in the Rio Grande Rise area and primarily
based on clay mineral investigations by the first
author, Dr. Franz X. Gingeie. My own contribution
to this study consisted in establishing age models
for the cores and statistical analyses ofclay mineral
data. Additional, previously unpublished results
were provided by Dr. Rainer Petschick (clay miner
als) and Dr. Carsten Rühlemann (oxygen isotopes).
In the following the four manuscripts are briefly
summarized. All data used in the thesis are archived
in the information system PANGAEA/SEPAN
(www.pangaea.de).
T. von Dobeneck and F. Schmieder
Using rock magnetic proxy records for orbital
tuning and extended time series analyses
into the super- and sub-Milankovitch bands
In: G. Fischer and G. Wefer
Use ofProxies in Paleoceanography: Examples
Fom the South Atlantic
Springer-Verlag, in press.
This paper outlines the methodical background of
the thesis. Two case studies are presented, both
based on a several sediment seq ucnces dated by or
bital tuning of their high-resolution magnetic sus
ceptibility records. Extended time series analyses
focus on the statistical evaluation ofdifferent pa1eo
climatic aspects documented in the frequency bands
above and below the main Milankovitch cycles. The
Subtropical §outh Atlantic §usceptibility (SUSAS)
stack extents back to 1.5 Ma and thus provides ideal
conditions to study various superstructures, e.g.,
amplitude modulations of proxy responses. Com
paratively elevated sedimentation rates in the Ceara
Rise §usceptibility (CEARIS) stack allow an exami
nation ofhigh-frequency sub-Milankovitch pheno
mena. Paleoceanographic implications are only
briefly discussed in this publication.
F. Schmieder, T. von Dobeneck and U. Bleil
Mid-Pleistocene climate transition: initiation~
interim state and terminal event
submitted to Nature
The article addresses a fundamental enigma ofQua
temary climate evolution which recently has been
subject of several inspiring scientific publications.
So far, most investigations of the mid-Pleistocene
climate transition were based on oxygen isotope
records reflecting global ice volume. The analysis
ofrock magnetic parameters presented here focuses
Introduction 7
on changes in deep water chemistry. The data sug
gest that the mid-Pleistocene c1imate transition from
about 920 to 640 ka does not represent a gradual
change from a 41 to a 100 kyr world, but rather a
third, quite different c1imate state confined by brief
shifts at both boundaries. Moreover, many of the
sediment series studied document an unusual ter
minal event at the end of the mid-Pleistocene cli
mate transition, possibly related to other yet unex
plained findings in l1umerous paleoclimatic records
at around this time.
F. Schmieder, T. von Dobeneck and U. Bleil
Cycles, trends and events of Pleistocene
sedimentation in the oligotrophie subtropieal
South Atlantic Ocean
to be submitted to Paleoceanography
The submarine ridges ofthe South Atlantic Ocean
provide ideal opportunities to study paleoclimati
cally driven temporal and spatial altemations of
NADW and LCDW as they intersect the vertically
fluctuating broad transition zone separating these
two deep water masses during glacial as weil as
interglacial times. Sediment series recovered at
twelve sites on these ridges were investigated for
cyclic fluctuations and trends of Quaternary depo
sition. The results imply a linkage between changes
ofpredominant cyclicity, long-term trends in carbo
nate accumulation and a paleoceanographic event
at about 530 ka documented at severallocations.
F.X. GingeIe, F. Schmieder, T. von Dobeneck, R.
Petschick and C. Rühlemann
Terrigenous flux in the Rio Grande Rise area
du ring the past 1500 ka:
evidence of deepwater advection or rapid
response to continental rainfall patterns?
Paleoceanography, 14, 84-95, 1999
The main objective of this study is to investigate
the usability of kaolinite as a tracer for NADW in
the Rio Grande Rise area. A comparison of kao
linite/chlorite ratios in surface sediment sampIes
from the mid-Atlantic Ridge and the Rio Grande
Rise region reveals that in the continental realm the
Rio Doce (Brazil) is a major source of kaolinite in
marine sediments whereas the supply of kaolinite
by NADW is of minor importance. Time series
analyses evince cyclic variations of kaolinite/
chlorite ratios in one ofthe SUSAS cores coherent
with global ice volume in the 41 and 100 kyr bands.
They are interpreted to have recorded fluctuations
in discharge ofthe Rio Doce and to mÜTor humidity
conditions in the South American hinterland.
8 Chapter 1
References
Berger, A, Long-term variations ofcaloric insolation resultingfrom Ealih's orbital elements, Quat. Res., 9, 139-167, 1978.
Berger, A., J. Imbrie, J. Hays, G. Kuk1a and B. Saltzman, Mi1ankovitch and Climate Part 1, Reide1, Dordrecht,510 pp, 1984.
Berger, A., Milankovitch theory and climate, Rev. Geophys.,26, 624-657, 1988.
Berger, W.B. ancI E. Jansen, Fourier stratigraphy: spectra1ga in adjustment of orbital ice mass models as an aid indating late Neogene deep-sea sediments, Pa/eoceanography, 9, 693-703, 199421.
Berger, W.B. and E. Jansen, Mid-P1eistocene climate shiftthe Nansen connection, in: The po/ar oceans and theil"roje in shaping the global environment, edited by O.M.Johannessen, R.D. Muench and lE. Overland, Geophys.Monogr. Ser., 85, AGU, Washington, pp 295-311, 1994b.
Berger, W.H. and G. Wefer, Klimageschichte aus Tiefseesedimenten, Naturwissenschaften 79, 541-550, 1992.
Berger, W.B., M.K. Yasuda, T Bickert, G. Wefer, T. Takayama, Quaternary time sca1e for the Ontong Java Plateau:Milankovitch template for Ocean Drilling Program Site806, Ge%gy, 22, 463-467, 1994.
Berger, W.H. and G. Wefer, Central themes ofSouth At1anticcirculation, in The South At/antic: present {[nd past circu/ation, edited by G. Wefer, W.H. Berger, G. Siedler andDJ. Webb, Springer, Berlin, pp 1-11, 1996.
Bickert, T, W.B. Curry and G. Wefer, Late Pliocene to Holocene (2.6-0 Ma) western equatorial Atlantic deep-watercirculation: inferences from benthic stable isotopes, Proc.
ODP, Sei. Resu/ts, 154,239-254, 1997.B1eil, U. and G. Gard, Chronology and corre1ation ofQuater
nmy magnetostratigraphy and nannofossil biostratigraphyin Norwegian-Greenland Sea sediments, Geol. Rundschau,78, 1173-1187, 1989.
Bloemendal, land P. deMenocal, Evidence for 21 change inthe periodicity oftropical climate cycles at 2.4 Myr fromwhole-core magnetic susceptibility measurements, Nature,342,897-900, 1989.
Cande, S.C. and D.V. Kent, A new geomagnetic polarity timescale for the Late Cretaceous and Cenozoic, 1. Geophys.Res., 97, 13917-13951, 1992.
Cande, S.c. and D.V. Kent, Revised calibration of the geomagnetic polarity timescale for the Late Cretaceous andCenozoic,1. Geophys. Res., 100,6093-6095,1995.
Chi, 1. and l Mienert, Linking physical property records ofQuaternary sediments to Heinrich events, Mal". Ge%gy,131,57-73,1996.
Clark, P.u. and D. Pollard, Origin of the middle P1eistocenetransition by ice sheet erosion ofregolith, Pa/eoceanography, 13, 1-9, 1998.
D'Argenio, B.D., AG. Fischer, G.M. Richter, G. Longo, N.Pe10si, F. Molisso and M.L. Duarte Morais, Orbital cyclicity in the Eocene ofAngola: visua1 and image-time seriesanalysis compared, Earth Planet. Sei. Lett., 160, 147-161,1998.
DeBlonde, G. and W.R. Peltier, A one-dimensional model ofcontinental ice volume fluctlrations through the Pleistocene: implications for the origin of the Mid-Pleistoceneclimate transition, 1. Clim., 4, 318-344, 1991.
Emiliani,C., Pleistocene temperature, 1. Geol., 63, 538-578,1955.
Frederichs, T, Regionale und altersabhängige Variation ge·steinsmagnetischer Parameter in marinen Sedimenten derArktis, Bel". Po/arforschung, 164, 212 pp, 1995.
Gersonde, R., FT Kyte, U. Bleil, B. Diekmann, l1\ Flores.K. Gohl, G. Orah1, R. Hagen, G. Kulm, F.J. Sierro, D.Völker, A. Abelmann and J.A Bostwick, Gcologicalrecord and reconstruction of the late Pliocenc irnpact ofthe Eltanin asteroid in the Southern Ocean, Nature. 390,357-363, 1997.
Grützner, l, F.C. Bassinot and l Mienert, High-resolutioncompressional-wave velocity measurements in Pleistoceneseeliments ofthe Ceani Rise (western equatoriaJ Atlantic):implications for orbital driven seelimentary cycles, Proc.ODP, Sci Resu/ts, 154, 135-149, 1997.
Hays, J.D., l Imbrie and NJ. Shackleton, Variations in Earih'sorbit: pace-maker ofthe ice ages, Scienee 194, 1121-1132,1976.
Ilerbert, TD. and L.A Mayer, Long c1imatic time series fromsediment physical property measurements, 1. Sediment.Petrol., 61,1089-1108, 1991.
Hilgen, F..T., Extension of the astronomically calibrated(polarity) time sca1e to the Miocene/Pliocene bounelary,Earth Planet. Sei. Lett., 107, 349-368, 1991.
Imbrie, J., 1D. Hays, D.O. Martinson, A. Mclntyre and A.C.Mix, The orbital theory of Pleistocene c1imate: supportfrom 21 revised chronology of the marine 8 180 recorel, inMilankovitch amI Climate Part 1, edited by ABerger, lImbrie, 1 Hays, G. Kukla and B. Saltzman, Reide1,Dordrecht, pp 269-305, 1984,
Imbrie, l, ABerger, E.A Boyle, S.c. C1emens, A. Duffy,W.R, Howard, G. Kukla, l Kutzbach, D.G. Martinson,A. McIntyre, AC. Mix, B. Molfino, lJ. Morley, L.c.Peterson, N.G. Pisias, W.L. Prell, M.E. Raymo, N.J.Shackleton and lR. Toggwei1er, On the structllre anelorigin of major glaciation cyc1es 1. Linear responses toMilankovitch forcing, Pa/eoceanography, 7, 701-738,1992.
Imbrie, l, ABerger, E.A. Boy1e, S.c. C1emens, A. Duffy,W.R. Howard, G. Kuk1a, 1 Kutzbach, D.G. Martinson,A. McIntyre, A.C. Mix, B. Molfino, lJ, Morley, L.c.Peterson, N.G. Pisias, W.L. Prell, M.E. Raymo, N.1Shackleton and lR. Toggwei1er, On the strucrure andorigin of major glaciation cycles 2. The 100,000-yearcycle, Paleoceanography, 8,699-735,1993.
Johnson, R.G., Brunhes-Maruyama magnetic reversal dateelat 790,000 yr B.P. by marine-astronomica1 cone1ations,Quat. Res., 17, 135-147, 1982.
Köppen, W. and R. Geiger, Handbuch eier Klimatologie,Verlag von Gebrüder Bornträger, Berlin, 1936.
Maasch, K.A., Statistical detection of the miel-Pleistocenetransition, Clim. Dyn., 2,133-143, 1988.
lntroduction 9
Martinson, D.G., N.G. Pisias, 1.D. Hays, J. Imbrie, T.c. MooreJr. and N..T. Shackleton, Age dating and thc orbital theory ofthe ice ages: development of a high-resolution 0 to 300,000year chronostratigraphy, Quat. Res., 27, 1-29, 1987.
Milankovitch, M., Die astronomische Theorie der Klimaschwankungen, in Handhuch der Klimatologie, edited byKöppen, \\l and R. Geiger, Verlag von Gebrüder Bornträge~Berlin, 1936.
Mudelsee, M. and M. Schulz, The mid-Pleistocene climatetransition: onset of 100 ka cycle lags ice volume build-upby 280 ka, Eorth. Planet. Sei. Lett., 151, 117-123, 1997.
Mudelsee, M. and K. Stattegger, Exploring the structure ofthe mid-P1eistocene revolution with advanced methods oftime-series analysis, Geo1. Rundschau, 86, 499-5 11, 1997.
Nowaczyk, N" I-Iochauf1ösende Magnetostratigraphie spätquartärer Sedimente arktischer Meeresgebiete, Bel'. Polorforschung, 78,187 pp, 1991.
Nowaczyk, N.. T.W. Frederichs, A. Eisenhauer and G. Gard,Stratigraphy of late Quaternary sediments from the Yermak Plateau, Arctic Ocean: evidence for four geomagnetic polarity events within the last 170 kyr ofthe Brunhesepoch, Geophys. J /nt, 117,453-471,1994.
Paillard, D., The timing of Pleistocene glaciations from asimple multiple-state c1imate model, Nature, 391, 1998.
Park, 1. and K.A. Maasch, P1io-P1eistocene time evolution ofthe 1OO-kyr cycle in marine paleoclimate records, J Geophys. Res, 98, 447-461,1993.
Pisias, N.G. and TC. Moore, The evolution of Pleistocenec1imate: a time series approach, Earth Planet Sci. Lett.,52,450-458, 1981.
Prell, W.L., Oxygen and carbon isotope stratigraphy for theQuaternary of Hole 502B: evidence for two nlOdes ofisotopic variability, Initial Rep. DSDP, 68,455-464, 1982.
Raymo, M.E., The timing of major climate transitions,Paleoceanography, 12,577-585, 1997.
Raymo, M.E., W.F. Ruddiman, 1. Backman, B.M. Clementand D.G. Martinson, Late Pliocene variation in northernhemisphere ice sheets and North Atlantic deep \vatercirculation, Paleoceanography, 4, 413-446, 1989.
Raymo, M.E., D.W. Oppo and W. Curry, The mid-Pleistoceneclimate transition: a deep-sea carbon isotopic perspective,Paleoceanography, 12, 546-559, 1997.
Robinson, S.G., Applications for who1e-core magnetic susceptibility measurements of deep-sea sediments, Proc.GDP, Sci Results, 115,737-771,1990.
Ruddiman, W.F., M.E. Raymo and A. McIntyre, Matuyama41,000-year cyc1es: North Atlantic Ocean and northernhemisphere ice sheets, Earth. Planet Sci. Lett., 80, 117129,1986.
Ruddiman, W.F., M.E. Raymo, D.G. Martinson, B.M.C1ement and l Backman, P1eistocene evolution: northernhemisphere ice sheets and North At1antic ocean, Paleoceanography, 4,353-412, 1989.
Sa1tzman, B. and M.Y. Verbitsky, Multiple instabilities andmodes ofglacia1 rhythmicity in the Plio-Pleistocene: a general theory oflate Cenozoic c1imatic change, Clim. Dyn.,9,1-15,1993.
Schwarzacher, W., Cyc10stratigraphy and the IVlilankovitchtheory, Elsevier, Amsterdam, 225 pp, 1993a.
Schwarzacher, W., Milankovitch cyc1es in the pre-Pkistocenestratigraphie record: a review, in High resolution stmtigmphy, edited by E.A. Hailwood and R.B. Kidd. Geul Soc.Spec. PuM, 70. 187-194, 1993b.
Shackleton, N..T., Oxygen isotope analyses and P1eistocenetemperatures re-assessed, Nature, 215, J5-17, 1967.
Shack1eton, NJ., J. Backman, H. Zimmermann, D.V. Kent,M.A. Hall, D.G. Roberts, D. Schnitker, lei. Ba1dauf, A.Desprairies, R. Homrighausen, P. Hudcllestun, 1.B. Keene,A.J. Kaltenback, K.A.O. Krumsiek, A.C. Morton, lW.Murray and 1. Westberg-Smith, Oxygen isotope calibration ofthe onset of ice-rafting and history of glaciation inthe North Atlantic region, Nature, 307, 620-623, 1984.
Shackleton, N.J., A. Berger, W.R. Peltier, An alternative astronomical ca libration of the lower P1eistocene tirnesca1ebased on ODP Site 677, Trans. Royal Soc. Edinburgh.·Em·th Sci, 81, 251-261,1990.
Shackleton, N.J., S. Crowhurst, T Hagelberg, N.G. Pisias andD.A. Schneider, A new Late Neogene time scak: application to Leg 138 sites, Proc. GDP, Sci. Re.l'ult.l', 138,73101,1995.
Shackleton, N.J. and S. Crowhurst, Sediment f1uxes based onan orbital tuned time scale 5 Ma to 14 Ma, site 926, Proc.GDP, Sci. Re.l'ults, 154,69-82, 1997.
Tarduno, lA., L.A. Mayer, R. Musgrave and ShipboardScientific Party, High resolution, whole-core magneticsusceptibility data from Leg 130, Ontong Java Plateau,Proc. GDP, /nit. Rep., 130, 541-548, 1991.
Thießen, W., Magnetische Eigenschaften von Sedimenten desöstlichen Südatlantiks und ihre paläozeanographische Relevanz, Ph.D. thesis, Bel'. FB Geow. Univ. Brernen, 170 pp,1993.
Tiedemann, R., M. Sarnthein and N..T. Shack1eton, Astronomietimescale for the Pliocene Atlantic 8 1sO and dust f1uxrecords ofOcean Drilling Program site 659, Paleoceanograph}',9, 619-638,1994.
von Dobeneck, T, Ana1ytical strategies in environmentalmagnetism - a South Atlantic perspective, Habilitation thesis,FB Geowissenschaften, Univ. Bremen, 1998.
Wilson, D.S., Confirmation of the astronomica1 calibrationofthe magnetic po1arity timesca1e from sea-f100r spreadingrates, Nature, 364, 788-790, 1993.
10
Using Rock Magnetic Proxy Records for Orbital Tuning and ExtendedTime Series Analyses into the Super- and Sub-Milankovitch Bands
1. von Dobeneck* and F. Schmieder
Fachbereich Geowissensch{4ten, Universität Bremen, Pos(lach 33 0440,D-28334 Bremen, Germany
*corresponding author (e-mai1):doheneck(ilJuni-bremen.de
Abstl'act: High-resolution rock magnetic pro:\.)' records of marine sediments, in partieular magneticsusceptibility logs, delineate variations of sediment litholol:,TJ' and mirror climatic and oceanographicchanges of different duration. Most commonly, Milankovitch cyclicity resulting from orbital forcingof carbonate dissolution and terrigeno\ls sedimentation prevails. Extracted by bandpass filtering,these signal components can serve for multiple core correlation, cyclostratigraphic analyses andorbital tuning. Phase relations between astronomical obliquity and precession cycles and theirequivalents in rock magnetic records depend on regional sedimentological settings. Two case studiesare developed to demonstrate specific aims, strategies, strengths and restrictions of rock magnetictime series analyses and their extension into the super- and sub-Milankovitch bands. In the firstexample, twelve Pleistocene susceptibility records spanning the oligotrophie subtropical SouthAtlantic (SUSAS stack) were successfeJIly tuned to obliquity and precession despite their lowaverage sedimentation rates< 1cm/kyr. Multiple bandpass filtering and evolutionary spectral analysisreveal two m~jor base line shifts at arOlUld 0.95 and 0.6 Ma ,md various superstmctures (e.g., amplitudemodulations) of orbital pro:\.; response. Compared to analogous analyses of an adapted astronomicaltarget curve, converse residue patterns in the 41 kyr and 100 kyr bands indicate that Pleistocene iceage cycles were mainly triggered by obliquity before 1.25 Ma andfrom 1.05 to 0.7 Ma, while eccentricitymodulation ofprecession predominated between 1.25 and 1.05 Ma and during the last 0.6 Ma. In thesecond example, eight susceptibility records from the western equatorial Atlantic Ceara Rise (CEARISstack) were tuned to a lagged precession index signal. Subsequently, a high resolution core correlationscheme was established on basis of their coherent high-frequency « 15 kyr) signal patterns. Basichannonics and intermodulation frequencies of obliquity, and, predominately, precession resultingfrom nonlinear pro:\."y response were detected by sub-Milankovitch spectral and bispectral analyses.Twin susceptibility peaks corresponding to a tropical double precession cycle appear even in theunfiltered records. Millennial signal variations « 7 kyr) seem to coincide with Bond- and DansgaardOeschger cycles.
Introduction
Many marine rock magnetic records exhibit Milan
kovitch cyclicities - sometimes in striking agreementwith 8180, 8 13 C or %CaC03 profiles. This phe
nomenon is generally explained by climatic impact
on the fluxes ofmagnetically enriched terrigenous
and 'non-magnetic' biogenic sedimentm'y compo
nents. They are increasingly employed as a data
base for high-resolution care correlation, orbital age
modelling and paleoceanographic time series analy
sis ofmarine sediments. Unlike ice-shield mediated
climate proxies, magnetic records sustain their
periodicity deep into pre-Quaternary times (Park
et a1. 1993; Hilgen et a1. 1995; Shackleton and
Crowhurst 1997). Paleoclimatic signatures ofmag
netic susceptibility and various remanence parame
ters have been described and stratigraphically ex
ploited in paleoceanographic studies from all major oceans (e.g., Radhakrishnamurty et a1. 1968;Kent 1982; Robinson 1986; Bloemendal et a1. 1988;deMenocal et a1. 1991; Bickelt et a1. 1997).
From FISCHER G, WEFER G (cds), 1999, Use 01 Prox/es /n Paleoceanography. Ex:amples ji'om the SOllth Atlant/c. Springer- VerlagBerlin Heidelberg, pp 601-633.
12 Chapter2
The article on marine 'Environmental Magnetism' by Fl-ederichs et al. (this volume) outlines thephysical and sedimentological principles of rockmagnetic climate proxies. Lithological changes insediment series can be traced by a range of specific rock magnetic parameters which indicate concentrations of para-, ferri- and antiferromagneticminerals, discriminate among remanence-carryingminerals (various iron oxides, oxyhydroxides andsulfides), estimate average ferrimagnetic grainsizes, or quantiry additional magnetic propeliies suchas coercivity or anisotropy.
There is no universal rationale for the linkageofclimate and sedimentary magnetic mineral inventories, but detailed rock magnetic studies usua1lyprovide the means to identify prevailing orbital forcing mechanisms within a regional context. Pronounced climate dependence has been observed foreolian dust load, ice-rafted debris, ocean currenttransport and sea-level related shelf erosion - a1limportant sources or pathways ofmagnetic mineraldeposition. The likewise cyc1ic accumulation ofcalcareous and siliceous microfossils plays a complementary and often dominant role in generatingmagnetic signals by modulating the concentrationofthe magnetic mineral fraction. Magnetic signalpatterns can be obscured by too stable or too complex sedimentation conditions, discontinuous deposition (bottom current erosion, debris ±lows), intercalated high- or low-magnetic (tephra, fossil) layers, and, most severely, by reductive diagenesis ofprimary felTimagnetic minerals in sub- or anoxicmarine environments (Tarduno 1994; Tarduno andWilkison 1996; Frederichs et a1., this volume).
Isothermal magnetic parameters are determinedby measuring the magnetic moment of artificiallymagnetized, but otherwise untreated bulk sedimentsampIes. They represent averaged volume properties, are precise and weIl reproducible. Because ofthe ubiquity ofmagnetic iron minerals in nature andthe availability ofhighly sensitive magnetometers,virtually every lithology can be investigated andcharacterized by magnetic parameter sets. Theattainable speed and spatial resolution ofwhole-coremeasurements convinced a growing munber ofmarine geoscientists to integrate rock magneticmethods, especially magnetic susceptibility 10gging,into their methodical repertory.
This paper intends to explain and illustrate standard and advanced concepts of Quaternmoy cyclostratigraphy with emphasis on marine rock magnetic records. The three main sections are concerned with- stepwise refinement ofmultiple care correlationschemes, phase analysis, and orbital tuning,- evolutionary spectral analysis ofMilankovitch andsuper-Milankovitch signal variations covering longperiodic (> 100 kyr) amplitude modulations andtransitions ofbaslc climate cycles- higher-order spectral analysis ofsub-Milankovitch.sIgnal components relclted to climate variations ofshort duration « 18 kyr).
As necessary restriction and tribute to the mostpopular rock magnetic parameter in paleoceanography, our case studies will exc1usively focus onmagnetic susceptibility records.
Experimental Methods
Traditional single sampIe measurements in rockmagnetism are increasingly replaced by faster nondestructive techniques. High-resolution whole-core01' half-core logging of magnetic susceptibility K
(Robinson 1990) is now a routine procedure inmarine sediment studies. The lateral sensor characteristics ofthe widely used 'Bartington' susceptometer have half-widths of 50 to 25 mm for loopsensors ofvarious diameters and of 12 to 4 mm fardifferent spot sensors. These values also representconservative minimum estimates for the attainablespatial resolution.
Modern pass-through SQUID magnetometersequipped with additionalmagnetization coils measure isothermal and anhysteretic remanent magnetization at resolutions ranging from 80 mm for wholecore logging down to 30 mm for 'U-channel' (axialsubcore) measurements (Weeks et a1. 1993). A10 mm resolution far surface remanence measurements is obtained by a newly developed high-TcSQUID 'Rock Magnetic Micro-Scanner' (vonDobeneck et a1. 1996).
Spatial resolutions down to 1 mm at the expense of a laborious single sampIe preparation areachieved by measuring magnetic hysteresis loops.The concept of the 'Alternating Gradient ForceMagnetometer' (AGFM) allows to detennine high-
Using rock magnetic proxy records for orbital tuning and extended time series analyses 13
quality hysteresis data from miniature 10-20 mgsampIes (Flanders 1988; von Dobeneck 1996).
Typical data sampling densities for high-resolution rock magnetic core-logging are 0.5-2 measurements/cm, which is more than the effective spatialresolution ofmost sensors. This over-sampling results in smooth, well-defined and complete records- an important prerequisite for the performance ofsignal analysis techniques. In this paper we presentsusceptibility records acquired with a 'BartingtonF type' spot sensor at a 1 cm spacing.
Orbital Tuning Strategies
Milankovitch' s (1941) visionary idea ofsynchronizing geological records to orbital variations hasevolved into a widely accepted chronostratigraphicmethod (Imbrie and Imbrie 1980; Imbrie et al. 1984;Martinson et al. 1987). Applied to records of isotopic 01' lithologic characteristics, 'orbital tuning' canyield unmatched precision and consistency in dating Quaternary sediment series. Ideally it takes justthree steps to align a proxy record to its orbitalpacemaker:- establish a rough initial chronology and detectspectral characteristics ofMilankovitch cyclicity,- choose an astronomie target curve and shift it bypostulated 01' empirically determined phase lags,- filter the proxy signal in the appropriate frequencyband and match it with its model record.
In practice this approach bears numerous pitfalls. Rock magnetic records, like most proxy signals, depend more on regional sedimentation conditions than on global equilibria. They are not, in principle, safe bets for orbital tuning pUllJoses and theirpatterns can be anything from a 'SPECMAP (Imbrie et al. 1984) double' to 'random wiggles'. Cyclostratigraphic age models critically rely on theinitial and thus very consequential age model. Simple peak-to-peak correlation of proxy and targetrecords is highly susceptible to elToneous matches(Shackleton et al. 1995), which have their origin inordinary stratigraphic problems such as undetectedhiati or turbidites, condensed or extended sections,bioturbation, lithologic and diagenetic effects. Algorithms such as signal slotting (Thompson andClarke 1993) 01' complex demodulation (Pisias etal. 1990) have been used to find the statistically
most probable proxy-target-correlation. Yet a COl11
bination ofmagneto- and biostratigraphic, eventually radiometric ages should be thc most reliablebasis to start from. Specific regional settings andscientific objectives may require adaptation of orbital tuning schemes into flexible, case-wise adapted strategies.
Two contrasting examples illustrate the scopeof possiblc approaches at different interpretationlevels. They are based on two series ofFS METEOR gravity COlTS collected within a longterm Quaternary South Atlantic research project(SFB 261) at the University of Bremen. The coring sitcs are located in the subtropical and in thewestern tropical South Atlantic. The two study areas differ greatly in size as weIl as in prevailingsedimentation conditions and rates. The first exampIe outlines the possibilities and limits ofsignal correlation and orbital tuning in low accumulation environments. Covering nearly the full Quaternary,these susceptibility records are rewarding targetsfor evolutionary spectral analysis. The particularinterest ofthe second example lies in a much highersedimentation and signal detail opening the possibility ofanalyzing periodic sub-Milankovitch signalcomponents. We will primmily focus at the methoclical aspects ofthese magnetic time series analyses.
Subtropicctl South Atlantic
The Subtropical ~outh Atlantic Susceptibility(SUSAS) Stack stretches over the total width ofthe pelagic South Atlantic (Fig. 1) at around25-35°S, connects its fourmajor basins (Argentine,Brazil, Angola and Cape Basin) and crosses threesubmarine ridges (Rio Grande Rise, Mid Atlanticand Walvis Ridge). Via two deep water passagesin the western South Atlantic, Vema and HunterChannel, carbonate undersaturated Antarctic Bottom Water flows northward into the Brazil Basin,from where only a minor part reaches the AngolaBasin through the Romanche Fracture Zone. 1'0day, the slightly carbonate supersaturated NorthAtlantic Deep Water (NADW) flows southward atdepths below 2000 m and controls the bottom watel' conditions in the Angola Basin. Therefore, thecarbonate lysocline is much deeper in the AngolaBasin (4700 to 4900 m) than in the Argentine, Bra-
14 Chapter 2
20 0 E0°
35°8
f-l---L.L--'--'--'--L...L-L-'-L-l-.J.-'--'-'---'---'---'-'-''-'-L..L-L..L-~~~.L..L--'--'--'--'--'--L...L-L-'-L~~~I~I~I~I~I~I~I~I~I~1L.L1 I j I I j I L ...L...i.
[m] M '<tMMd> ..o~ ~
1,000 ~ gg~~ ~~-r-
o-1,000
-2,000
-3,000
-4,000
-5,000
-6,000
Fig. 1. GeoB core locations defining the SUSAS transect. The depth profile follows the white dashed line in themap and is plotted against longitude to depict the cores' affinities to the four major pelagic basins and their deepwater bodies. It is not suitab1e to identify the positions and depths of deep water passages.
zil, and Cape Basins (4000 m). This west-eastasymmetry of carbonate preservation diminishesduring glaeial times, when NADW produetion isredueed and the lysoeline rises to about 3800 m inall four basins (Biekert and Wefer 1996).
Due to oligotrophie eonditions along the profilethe average sedimentation rates over the last1.5 Ma range between 0.4 and 1.2 em/kyr (Fig. 5).Suseeptibility eare means range from 18.5 to470x 10-(, SI. They inerease with depth and showa drastie west to east decline originating from amueh higher terrigenous influx to the western partsofthe subtropical South Atlantie (Lisitzin 1996). AllSUSAS logs exhibit distinet periodie oseillationsprimarily in the 100 kyr and 41 kyr band refleeting
carbonate aeeumulation variations, with eoherent10ng-tern1 signal shifts superposed.
Oxygen isotope reeords exist für four SUSASeores, GeoB 1034-3, 1035-4, 1211-3 (Biekeri 1992;Biekert and Wefer 1996) and 1309-2 (W. I-laIe,pers. eomm.). Their stratigraphie interpretability islimited by very low sedimentation rates, partieularlyin the deeper, obliquity-dominated (41 kyr eyeles;e.g., Ruddiman et al. 1986) seetions. Magnetostratigraphie dating at 5-10 em spaeing (Fig. 2) provided three 01' four reliable tie points in ten oftwelveeores. By attributing these ages to the respeetivemagnetie suseeptibility logs, some speetraleharaeteristies of Milankoviteh eyelieity beeomeapparent (Figs. 3a, 4a) suggesting elimatie foreing.
Using rock magnctic proxy records for orbital tuning and extended time series analyses 15
ChRM Inclination n2821·1 2820-2 1311-1 1309·2 3814-6 3813-3 3812-1 1729-3 1034-3 1211-3
o 00( 0 m
{
!L.
o 0~ 0 m
ro .C'~ 'e::
cuCf) (5Q) 0-OJ«mCf)'-Q)>Q)
0:::
• I
0.78
1.95
0.991.071.19 CM
cuEcu>.::J
cu:2:
1.77
Fig. 2. Combined magnetostratigraphies (ChRM: Characteristic Remanent Magnetization) ofSUSAS cores comprising the Brunhes/Matuyama boundary and the Jaramillo event (Jar). The Cobb Mountain (CM) event cannot beclearly identified at all sites. The Olduvai (Old) event is not reached in any of the cores. Reversal ages are givenaccording to Shackleton et a1. (1990). Data ofcores GeoB 1034-3 ane! 1211-3 are from Thießen (1993).
Lomb-Scargle Fourier transfom1 (LSFT; Lomb1976; Scargle 1982, 1989) embedded in the SPECTRUM program (Schulz and Stattegger 1997) wasused throughout all spectral analyses unless indieated, as it can be direetly applied to unevenlyspaeed times series. The widely applied BlaekmanTukey method of spectral analysis (Blaekman andTukey 1958) employs Fourier Transform of thetruncated and tapered autoeovariance function andrequires evenly spaced time series. The necessaryinterpolation procedure not only fails at age gaps,but inevitably underestimates high-frequeney eomponents and thereby 'reddens' the spectrum (Horowitz 1974). Moreover, interp01ated data are statistically dependent (Mudelsee and Stattegger 1994)eomplicating the assessment ofthe signifieance ofcomputed speetra.
Another important concem is spectralleakage,a mathematieal consequence ofthe finite length oftime series. Due to this effect dominant orbitalcycles generate spectral ripp1es at higher frequeneies, which may form misleading interference patterns. Tapering records with suitable window func-
tions such as a Welch or Hanning window redueesthis leakage effect at the expense of attainableresolution (HaITis 1978). Individual compromisesare found by eomparing results from different tapers. A seeond 'spurious peak problem' arises fromrandom fluctuations related to stochastic proeesses. A relatively simple eounterstrategy is'Welch's overlapped segment average' (WaSA)proeedure (Welch 1967). By dividing a proxy reeordinto overlapping segments and averaging their rawspectra, 'noise' peaks are gradually suppressed, butagain resolution decreases.
Here a combination strategy of tapering andWaSA is applied. Following signal analytica1 nomenelature, speetra by Fourier transform of rawdata are called 'periodograms', tapering yields'modified periodograms' and segment averaging'averaged periodograms' . Solely 'autospeeh'a' givea complete spech'al representation oftime-dependent processes. They cannot be determined on basis offinite diserete time series, but only estimatedwithin increasing1y nalTOW eonfidenee limits by theabove defined spectra.
16 Chapter 2
All initial spectra display significant maxima ataround 40 kyr as core GeoB 3814-6 (Fig. 4a), butshow little or no indications 01' a precessional signal component. As a spectrum represents an average over the integral record, it would be overlyoptimistic to inter that a11 obliquity-related oscillations can be extracted at this stage by filtering therecord in the respective frequency band. Instead,all susceptibility records were mutually correlatedto identify missing 01' expanded intervals. This comparison enables to estimate the extent o1'some shortgaps and two laminated diatom ooze layers (Fig. 5),which had obviously been deposited in very shorttime. Core to COlT correlation revealed that simply
removing these sections 1'rom the sequence resultsin complete records.
A pattern matching to standard 8 180 records.e.g. from equatorial Pacific ODr Site 677 (Shackleton et al. 1990) was feasible over most signalsections, but remained ambiguous between oxygenisotope stages 16 and 13 between about 0.65 and0.5 Ma. Contll1uously high simdarities, includinglong-term f1uctuations 01' about 500 kyr duration,exist between all SUSAS susceptibility signals andthe benthic 8 13C record at Pacific ODP Site 806(Bickert et al. 1998), based on an orbita]]y tuned8 180 stratigraphy (Berger et al. 1994). The compIete and undisturbed susceptibility record of core
Polarity
ISuscepti-bility K i
~~~~.__. .__ "" ~--.J
1.5141.31.21.10.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9Age [Mal
0.1
K
filtered15-47 kYrh,-""T"T'"",,,,--"T'""r"1---.---.-.+--r-,,,-.---r-r-ro"""TT"-.--Tlrr."'--rrT-rT"",,,-r-r-,-,'-,Ho-rr-rrT"""ri"""T""TO""""rT"'-rrT--'r--rr-r'r"""ri""T'T-rl
o
TargetCurve
Fig. 3. Four step refinement ofthe age model for core GeoB 3814-6. (a) Susceptibility record dated by magnetostratigraphy (3 tie points). (b) Signal pattern conelation to the tuned 0 \lC record ofODP Site 806 (44 tic points). Incritical intervals other SUSAS records were inspccted for supplementary inf0l111ation (c) Susceptibility filtered (3547 kyr) in the obliquity range (dashed line) and hmed (solid line) to an astronomical obliquity signal shifted by-4.5 kyr (73 tie points). (d) Susceptibility filtered (15-47 kyr) and tuned to an astronomical (obliquity and precession) target curve (95 tie points).
Using rock magnetic proxy records for orbital tuning and extended time series analyses 17
required. 8 180 phase lags have been determmcd onbasis ofradiometric ages (Hays et a!. 1976: 9 kyrtor obliquity, 3 kyr for precession) and assumptionson the coupling of insolation and ice mass (Imbrieet a!. 1984: 7.9 (7.4-8.2) kyr for obliquity, 5.0 (4.85.1) kyr and 4.2 (4.1-4.3) kyr for precession). Taestimate orbital phase lags for regional proxy parameters such as susceptibility it is therefore sufficient to know thc phase lags relative to a Öl80curve [rom the same COlT.
Here we use a revised 8 180 time scale of coreGeoB 1211-3 (Bickert 1992) based on correlationto thc SPECMAP stack (Imbrie ct a!. 1984). Crossspectral analysis of8 180 and K yields a coherenceofO.99 and 0.96 for the 100 kyr and 41 kyr cyclesand reasonable phase angles of -41 0 and -30 0
. Thisimplies that susceptibility leads8 180 by 3.4 kyr inthe obliquity band and therefore lags the obliquityforcing function by 7.9 kyr-3.4 kyr=4.5 kyr. Thetemporal resolution ofthe 8 180 reeord was tao Jowto determine the precessional phase lag with anacceptable error margin. 'rhe subsequent orbitaltuning process was performed in two stages:
At first the obliquity-related signal componentwas extracted applying a 1Sl order butterworthbandpass (35-47 kyr) filter in forward and reversedirection. This 'zero-phase-filtering' avoids phasedistortion of the resulting sequence (Oppenheimand Schafer 1989) and was thereforc used in allinstances. It effectively doubles the filter order. Alow filter order results in a short filter length andproduces little reverberation. Its soft transition between stopband and passband limits the risk ofsignalloss in record sections, which are unduly compressed 01' stretched (and hence shifted in frequency) by imperfect initial age models.
Each maximum and minimum in the filtered signal was assigned the age (+4.5 kyr) of its postulated equivalent in the ash'onomical record (Bergerand Louh'e 1991), stmiing from the most uncriticalsections near magnetostratigraphic 01' other reliabletie points. As illustrated in Fig. 3c, relative changesin the conelation age model are small, typically lessthan a half-cycle of obliquity. In the frequencydomain (Fig. 4c) one notes just a small rise ofpeakamplitudes.
The second tuning step proceeds from the observation that faintly visible precessional peaks in
40
40
50
~~~""iIl'-rl""""'r"M".,..,..,.""':'~30
60
>- >- >- >- >-CL CL >- >- >- CL CL >-.q r-- CL CL CL CL r-- .q CL
0 N <.D .q ::;; m 0'0 N m.q ~ m CD N N NU U u :D :D :D <J.)
~<J.)
u u u 0:: 0::w w w 0 0 0 Cl.
60
50
40
30
@ 60
<J.)0)
«-'(':'t?.8
(i) <J.)c
(ü 0)cu
Ü ::2:(J)'-(1J <J.)(lJ 0)c «2 c
0
"-- ~-- ~
~ CiU
N[;?b
E '.0
2 <J.)
TI 0)
«(lJ
0)0.. c
(J) 'cQj
::JI-
3: :20 ::J
0.. .sr:D
"0 0(lJ
cuE <J.)
0)
~ «0)
W c'c::Jl-
m0)
CoI-
0
GeoB 3814-6 (Fig. 3b) cOITelates particu1arly welland he1ped to substantiate the other eleven matches(Fig. 5). The improved spectral definition ofMilankovitch frequencies owing to this correlationprocedure is obvious (Fig. 4b).
The most widely used orbital tuning procedureassumes a phase lock between climate and metronome record (Martinson et a!. 1987). In order toassess the absolute lag of any proxy record withrespect to astronomical obliquity and precession, anindependent and precise absolute chronology IS
~~~~"!""NL,..,..~1-rT-1"rT--,Jl-300.01 0.02 0.03 004 0.05 0.06
Frequency [1/kyr]
Fig. 4a-d. Estimated spech'a (Hanning taper, WOSA: 2segments) of the GeoB 3814-6 susceptibility signal onlinear (solid b1ack) and logarithmic (line) scales at eachage model stage ofFig.3a-d. The cross is re1ated to thelogarithmic scale and delineates 6 dB bandwidth and90 % confidence interval of all estimated spech'a. Thedotted lines depict orbital periods ofeccentricity (Ecc),obliquity (ObI) and precession (Pre).
18
:MM,t4
Chapter 2
CIM MatuyamaI,! r! I!! I! I, I t,! I!!! I,! ! I I! i!!! I! I!! I! I!! I, I!! I !!, I,!!! I!!! I L..l._....l....-~ __.LlJ......L-LLJ
0.2 0.3 0.4 0.5 0.6 0.7 0.8Age [Ma]
0.9 1.0 1.1 1.2 13 1.4
1(+S'l
K-IK-SJ
1.5 16
Fig. 5. Individually tuned SUSAS recards and resulting stack (arithmetic mean with standard deviation band). Tocompensate for different mean amplitudes, each recard was standardized by subtracting its mean and dividing bythe standard deviation s. Dots mark identified paleomagnetic reversals (Jar: Jaramillo; CM: Cobb Mountain). Theyare generally located deeper in the cares than the corresponding reference ages of Shackleton et a1. (1990), depicted by dotted lines. The resulting average lock-in depth is l2±9 C111. There is an obvious east-west trend (topto bottom) of increasing signal amplitude, but less detailed signal features. The gray vertical bar marks a zone ofpeculiar lithologies in six cores related to a conspicuous increase of averaged sedimentation rate (bottom, individual sedimentation rate records were previously n0l111alized to the overall 0.5-0 Ma average ofO.75 cm/kyr).
Using rock magnetic proxy records for orbital tuning and extended time selies analyses 19
the logarithmic spectra gradually sharpened at bothrefinements ofthe age mode!. To benefit from thisprecessional signal component, a target curve ofn0l111alized obliquity and precession index (mixingratio of2:3 chosen by visual evaluation) and 4.5 kyrlag was calculated, comprising not just more, butalso more prominent signal features than either single orbital parameter. The obliquity-tuned primarysusceptibility record was filtered using a widebandpass (15-47 kyr) to include obliquity and precession cycles. As many extrema as possible werematched (Fig. 3d), admittedly at the resolution limitofthese records. The resulting final age models areagainjust slight modifications ofthe previous models and possibly not even more precise 111 absoluteages as the precession time lag is undetermined.Nevertheless, this higher-frequency tuning leads toa better mutual signal correlation of the SUSASrecords and therefore to an improved conservationof signal details in the SUSAS stack (Fig. 5). Because of the very different amplitude ranges, thestack was determined as an arithmetic mean ofnormalized (subtraction ofcore mean and divisionby standard deviation), interpolated (Lit=2 kyr)records.
Bioturbation should be responsible for the comparatively low spectral power in the precessionband. Its effect can be assessed from the fact thatseetions ofhigher sedimentation rate (e.g. interval0.6-0.2 Ma) clearly coincide better with precessionpatterns. A precessional cycle at a typical sedimentation rate ofr=0.5 cm/kyr corresponds to roughly10 cm core depth, twice the minimum estimate forthe pelagic mixed layer thickness L (Peng et a!.1977). Mathematical models of vertical mixing(Berger and Heath 1968; Dalfes et a!. 1984) permit to quantify the frequency selective amplitudedamping. Under the stated sedimentation conditionsthe 'gain factor' G at an averaged precession frequency of[= 1/21 kyc l is
Less intense damping of the obliquity relatedpaleosignal (G=0.30) and nearly negligible damping of 100 kyr (G=O.72) and 400 kyr (G=0.98)cycles account for a relative loss of spectral inten-
sity in the precession band. Bioturbation also influences phase relations to some extent (Dalfes eta!. 1984) and can bias orbital tuning ages.
Fig. 5 summarizes the combined resll1ts 01'magnetic age modelling for the complete transecLA 1'ew question marks remain in core sections,where individual signal features seem to be missing01' incomplete (e.g., GeoB 1311-1 at 0.42Ma,GeoB 2820-2 at 0.8 Ma). The SUSAS stack is freeof such local effects and therefore representativefor the whole Quaternary oligotrophie deep SouthAtlantic Ocean.
Cearä Rise
The Ceara Rise .§.usceptiblity (CEAR1S) Stackcomprises two short tTansects on the northern flankof the Ceara Rise, a submarine elevation 700 kmnorth-east offthe Amazon Delta (Fig. 6). A steady,but highly variable deposition ofterrigenous partieIes strongly modulates the susceptibility signa!.Several pathways are involved. The sea-Ievel fallin glacial periods causes erosion ofthe inner continental shelf, channeling ofthe Amazon dischargedirectly into the Amazon Canyon (Milliman et a!.1975; Damuth 1977) and intensification ofmassflow events on the slopes of the Amazon Fan(Maslin and Mikkelsen 1997). These mechanismsenhance the suspended particle load at differentwater depths reaching the Ceara Rise either bysurface transport via the retrofleetion ofthe NorthBrazil Current into the North Equatorial Countercurrent (Johns et a!. 1990) or by intensified nepheloid transport via the southeastward directedNADW flow (Kumar and Embley 1977; Francoisand Bacon 1991). Increased eolian input from Africa in glacial periods (Sarnthein et a!. 1981;Matthewson et a!. 1995) and possibly enhancedsediment yield of the Amazon in glacials are further factors which contribute to the susceptibilityvariations.
This rather complex scenmio ofcompeting sediment fluxes might be suspected ofproducing complex composite susceptibility signals. Yet allCEARIS records are perfectly suitable for orbitaltuning and exhibit very consistent patterns not onlyin the Milankovitch band, but also down to periodsof2-5 kyr.
20
52°W 50 0 W 48°W
Chapter 2
46°W 44°W 42°W 40 0 W
6°N
5°N
4°N
3°N
2°N
1°N
0°
1°S
2°S
Fig. 6. Simplifieel hyelrography ofthe Ceani Rise region anel GeoB core locations inclueleel in the CEARIS stack.The retrofleetion ofthe NOlih Brazil Current (NBC) into the NOlih-Equatorial Countercurrent (NECC) is activateelfrom June to January (Muller-Karger et al. 1988). The North Atlantic Deep Water (NADW) flows e1irectly over theCeara Rise, while the Antarctic Bottom Water (AABW) fully circumvents this bathymetric high.
Although no CEARIS eore was believed toreach the Matuyama chron, a11 were submitted tosystematic paleomagnetic analysis (Bleil andvon Dobeneck, this volume). Several reversedintervals, some with multiple rebounds, were deteeted in eaeh polarity record. Dating these features, whichjust vaguely coineide with previouslyreported Brunhes polarity events (e.g., Nowaczyket a1. 1994) seems diffieult.
Oxygen isotope stratigraphies are available fortwo of eight CEARIS eores (Mulitza 1994; Rühlemann et a1. 1996). These ehronologies (Fig. 7)were based on a SPECMAP (Imbrie et a1. 1984)eorrelation and represent eonvineing initial agemodels. They were adopted for the two respeetivesuseeptibility reeords (Fig. 7) and transferred to theother six cores (Fig 10a) by graphie eorrelation.
The pronouneed similarity of8 180 and suseeptibility records and their obvious agreement withSPECMAP testify to the elimatie dependenee ofboth parameters during late Quaternary glaeial/interglaeial eycles. It is interesting to note that thesynehrony of oxygen isotope and susceptibilityvariations is disturbed by varying peak lags andincoherent sections (Fig. 7). These effects are toolarge to be simply explained by the higher spatialresolution of the magnetie (1 cm) versus the isotopie (5 cm) record. Potential eauses are- proxy- and eycle-speeific gain and lag va1ues resulting from different signal forn1ation meehanisms(Martinson et a1. 1987),- hatmonie and intermodulation frequencies generated by (eoupled) nonlinear response to orbitalforeing (e.g. saw-toothed eurves),
Using rock magnetic proxy records for orbital tuning and extended time series analyses 21
GeoB 1515-1
8180 [%0] K [10-6 SI]
-2 -1.5 -1 -0.5 0 50 100150200250u,,-c'J..u.~.LUJ..u.JW-.LLli-- 0
GeoB 1523-1
8180 [%0] K[10-6 SI]
-2 -15 -1 -0.5 0 50 100150200250
SPECMAP
6180 [%0]
-3 -2 -1 0 1 2
o 1.1
2.23 1
50 334.2
100 53254
5.56.2
150 - 643
65 0m 66 (l).Y 7.1 -0'Q)' 200 72 s:Ol 7.3 4~«
74 2,7.5
2505
300
6
350111 7
400
Fig. 7. Initial age model for CEARIS cores. Globigerinoides sacculij'er 8180 records 01' cores GeoB 1523-1 (Mulitza1994) and GeoB 1515-1 (Rühlemann et al. 1996) were graphically correlated to SPECMAP (Imbrie et al. 1984) substages (dotted lines). Their ages provide 27 (19) tie points for the respective susceptibility records. Note inconsistent lead (stages 4.2,5.4,6.2,8.4) and lag (stage 5.1) OfK to 8180.
- the influence of extraneous (non-Milankovitch)forcing sif,'11als.
All possibilities are realistic and have impOliantimplications for the tuning sh-ategy, pmiicularly forchoosing the most suitable target record.
The spectral representations of Fig. 8 revealsome information in this respect. Botho l80-basedspech-a exhibit well-defined and similar, albeit notidentical orbital characteristics. The susceptibilityspectrum has considerably more precessionalpower and some additional peaks. Reasons for thediffering distribution ofspech-al power will be givenlater. In the context of orbital tuning, it is sufficientto state that magnetic susceptibility chiefly followsprecession and retains a rather stable phase lag(Fig. 9a, c). A shifted astronomical precessionrecord is therefore the most evident tuning target.
SPECMAP determines the phase lag of ourinitial age model. According to cross-spectral analyses of both proxy parameters and the precessionindex (PI), the lags at the three main precessionfrequencies amount to 3-8 kyr foro '80 and 0-5 kyrfor K (Fig. 9c). Since there are more than threeorbital precession frequencies (the peliods 23.7 kyr,22.4 kyr and 19 kyr are only the three 1m-gest amplitude terms in the h'igonometric expansion ofastronomical precession; Berger and Loutre 1991),the applicable lag should be represented by aweighted average combining these phase angles,period lengths and spectral power. The time-domainequivalent of cross-spectral density, the crosscovariance function c ,supplies this weighted av-
xy
erage lag. As both compared records are band-lim-ited (precession by nature, susceptibility by band-
22 Chapter 2
Fig. 8. Spectra (0-377 ka, Welch taper) ofSPECMAP andGeoB 1523-1 8 180 and K records according to the correlation age model ofFig. 7 on linear (solid black) and logarithmic (line) scales. The cross is related to the logarithmic scale and delineates 6 dB bandwidth and 90 % confidence interval ofal! estimated spectra. The dotted linesmark basic orbital periods. Note the various influenceof precession forcing in each spectrum and additionalpeaks in the K record. The lower spectral resolution incomparison to Fig. 4 is a mathematical consequence ofthe much shorter time series.
pass filtering), the horizontal shift of the centralmaximum of the cross-covariance function(Fig. 9b, d) corresponds to the requested lag(Fig. 9c). According to this analysis, the susceptibility record ofcore GeoB 1523-llags precessionalforcing by 1.77 kyr and leads 8 180 by 3.90 kyrwhich adds up to a plausible total delay of 5.67 kyrbetween PI and 8 180.
Starting out from the 8 180 based correlationage (Fig. 10a), precession tuning is a straightforward, but typica11y iterative procedure (Fig. lOb).Each step starts with filtering the time series in an
extended precession band (15-28 kyr to include erroneously compressed or stretched signal sections),proceeds with identification and correlation ofclearly developed maxima and minima to the shiftedprecession signal and ends with releasing the preceding and adopting the improved age model. Theseage modifications averaged 1.3-2.2 kyr in the firstand 0.4-0.9 kyr in the sec(md tuning step.
While the initial pattern cOlTelation scheme wasessentia11y based on matching individual peaks, theorbital tuning procedure relies on thc maxima andminima of a filtered signal, to which, due to thenecessary filter length, entire record seetions numerically contribute. Consequential1y, a tuncd agemodel relies more on the continuous evolution thanon singularities of a paleoccanographic record. Onthe other hand, nalTOW peaks and steep slopes havea strong impact on filter-extTacted signals and hencealso on their dating. The origin ofthese signal com·poncnts is apriori uncertain and provides no ties toexternal time sca1es unless convincing links toknown high-frequency climate variations such asHeinrich events (Grousset et a1. 1993; Robinson eta1. 1995; Chi and Mienert 1996), Bond (Bond et a1.1993) or Daansgard-Oeschger cycles (Dansgaardet a1. 1993; Moros et a1. 1997) can be established.
Spike and slope features, which are we11 reproducible within a core co11ection, can neverthelesshelp to improve the internal coherence ofcombinedage models. Exh-acted and emphasized by highpassfilteling in the sub-Milankovitch band (here <15 kyr)the high-frequency patterns of the CEARIS corecollection was graphica11y conelated at an averagespacing on kyr between tie points (Fig. 1Oc). Eachtie point was set to a common mean age (CMA)assuming that relative delay times in the study areaare negligible. The graphic cone1ation ofhigh-frequency signal components results in mean ageshifts ofO.8-1.5 kyr. From a statistical viewpoint,this procedure corrects for orbital tuning errorsrelated to core-specific features and thereforeimproves the precision ofeach age model. Becausea11 conelated cores were previous1y dated by thesame proxy and principles, the cumulative age shiftequals zero for the total core co11ection and, forstatistical reasons, also for each core.
At this refinement stage, it is justified to demand,how precise this tuned age model might be. A major
20
10
o
40
30
(j)
0::
>->--'" -'"t'-'<;f"
MNN N(j) (j)
0::0::
""""''''l'''!''"'I''"I'''I''"'1'-:r'-H.J,-,-"'I''"I''''l''\L....,...,~'-rl- -1 0 S"20 '§.,
s:10 :3
o'o g>
~CD
.....,..~..,..,."!--10CL
50 ~
~~""!"l-20
0.01 0.02 0.03 0.04 0.05 0.06Frequency [1/kyr]
);, );,);, );, );,-'" -'"
'<;f" t'- -'" -'" -'"0 N (0 '<;f" ~
'<;f" ~ rn LI) '<;f"0 0 0 Li .D0 0 0W W W 0 0
o
Using rock magnetic proxy records for orbital tuning and extended time series analyses 23
€I..••·.'~." ':':" <':' :.'X.·':':\.. " ...>:..... :,.'...:. ' .... ~.. :.. ~>. ::•. ,: •• '.:~:" ...:,,' •• '. :C>:.:':. '.. J.""".':':'. '; ',.,/1\.,. ').'~~.'.~·:>:-I'\§J' •..••• ". • • •. . ••••••. ••••. k- \J
/ ..'+"" -
;L ~I\/V~-'I
~1:iili!:jl·jl:{-T~ij":lli'il
o 20 40 60 80 100 120 140 160 180 200 220 240 260 280 300 320 340 360 380Age [ka]
·50 o 50 100 150 200 250 300 350 400
~ ~ ~ \~ Signal Lag [kyr]
~ ~ ~ ------------C'J 0J ~ __
g:. g: CL 0180 ---,. ~;",.<,":rl\ • I 5.67kyr .l.LL...L.L....t:'...L.l.•....LL......L~
• '·I·g·,;rI·} \. TI '\. JI·I·l
• 1
-300 -250 -200 -150 -100-350
.·h
-400
o
Lag[kyr]
10
-10 .L...... _10bJ .... .. KI_1:~~:~r77kY'
3.9 kyr
1,77 kyr
5.67 kyr
0,03 0.035 0.04 0.045 0,05 0.055 0,06Frequency [1/kyr]
-10 -8 -6 -4 -2 0 2 4 6 8 10Signal Lag [kyr]
Fig. 9. Quantification ofthe 0'80 and K signallags in relation to precession forcing (example: GeoB 1523-1). (a)Based on the SPECMAP correlation age model (Fig. 7),0 '80 and K were zero-phase filtered in the precession band(3rd order butterworth 15-28 kyr) and compared to the precession index (PI) ofBerger and Loutre (1991). All signalsexhibit similar amplitude modulation and fairly stable phase relations. (b) Comparison of cross-covariance functions ofPI and precession-filtered proxies. The centralmaximum (shown magnified in (d)) indicates the best matchofdelayed metronome record and proxy response. Estimated lag values are 1.77 kyr for susceptibility and 5.67 kyrfor 0'80. (c) The phase spectra of PI and unfiltered proxy records (derived from cross-spectral analysis) were converted to 'lag spectra' to show their frequency-dependence and good coincidence with lag values obtained from(b). Enor bars represent 80% confidence intervals, corresponding auto-spectrum amplitudes are shaded gray. (d)Close-up at cross-covariance functions of (b).
error source is age-depth-nonlinearity (Schiffelbeinand Dorman 1986). The applied (linear) age interpolation does not consider fluctuations ofthe sedimentation rate between tie points. The relative agediscrepancies between CEARIS cores due to thiseffect should on average be less than 1 kyr, as thematched signal features are densely spaced andtheir high resemblance implies regionally uniformsedimentation conditions. A quasi-continuous pat-
tern correlation (Martinson et a1. 1982) may further reduce this margin. The synchronism with theexternal forcing signal is potentially more affectedby interpolation elTors. Its precision can be roughlyestimated from cross spectral analyses and is in theorder of 2 kyr. The absolute enor margins of orbital tuning are even larger, but difficult to quantify. They result from uncertain and unstable phaserelations (Martinson et a1. 1987), vertical mixing by
24 Chapter 2
E:.ciliQ)ü({)
:::J(J)
Ü
~COl<U 1521-1::2:
1523-1
PI
"0 1515-1~
~li=
~ 1517-1-'"coN
1519-1J:,~
"0C<U.0 1521-1c0'(ij 1522-2(()Q)ü~ 1523-1
0..
"0Q)
~ 1515-1li=
'C' 1516-2>.-'"L{)
1517-1~
Y...-"0 1519-1c<U.0
1520-2.c.B.:;
1521-10-'"c~ 1522-2~..6 1523-1:::J
(J)
rso
r1SOI150
EW 250~ , ' " ",'--'"a:ü ••.•..•..•...•..•....••.•...••......•.•.••..• \. / ..•.•.••. /< / .«~a1::.,',,",,'., " ',. , ',',,,'.. ,,/ ... (. ,.,.. ,... ,... ,." ,',.,•,
o 20 40 60 80 100 120 140 160 180 200 220 240 260 280 300 320 340 360 380Age [ka]
Fig.l0. Tlu'ee step refinement ofage models for CEARlS eores. (a) Graphie eorrelation to 8180 age model of eareGeoB 1523-1 (48 tie points). (b) Initial (dashed) and subsequent stage (solid) of synehronizing all bandpass filteredsuseeptibility signals to the astronomieal preeession index (lag 1.77 kyr, 33 tie points). (e) Graphie eorrelation ofsub-Milankoviteh band « 15 kyr) features to their eonm10n mean age (90 tie points). (d) CEARIS stack.
Using rock magnetic proxy records for orbital tuning and extended time series analyses 25
bioturbation (Berger and Heath 1968) and an increasing error ofastronomical calculations with age(Berger and Loutre 1992).
Specijic Aspects ofTuning Rock Magnetic
Records
Orbital tuning ofproxy records implies the assumption that 21 parameter is almost exclusively responding to primary orbital forcing and that the responsefunction is not misleading the correlation ofproxyand target signals. Saturation effects, competinganticorrelated orbital response mechanisms andsecondary overprinting will bias the tuning approach and eventually make it pointless. It may bepossible to master such problems with specificallydesigned target curves (e.g., Berger ami Jansen1994(1), but this strategy requires 21 deep understanding 01'2111 processes involved. Obviously thereis 21 need for simpler criteria defining whichsediments are appropriate for orbital tuning. A fewgeneral rules 1'01' rock magnetic age modelling invarious depositional environments can be outlined.
01' 2111 rock magnetic records those from subor anoxic sediments are the most precarious as theyare usually affected by magnetite reduction. Thesecomi example in the paper by Frederichs ct 211.
(this volume) illustrates, how multi-parametricmagnetic methods can be used to reconstruct 21
susceptibility record which has been diageneticallyoverprinted during temporary suboxic conditions.
An abrupt coarsening 01' fining ofthe magneticmineral fraction as indicated by magnetogranulometric proxies such as M IM (ibid.) reveals
ar Ir
changes in the sedimentation modus, that probablyalso mask paleoceanographic inf01mations 01'otherproxy parameters. Whatever may cause this disturbance - erosion, winnowing, reductive diagenesis,slumping 01' intercalated turbidite 01' tephra layers the respective sections should be carefully studiedand eventually dissected. Tbe remaining record canstill be reassembled and tuned, if 21 pattel11 correlation with undisturbed records permits 21 convincing quantification 01' resulting age gaps.
We found the simultaneous orbital tuning of520 regionally related cores much more conclusivethan 21 treatment of individual records, because itgives 21 better impression 01' pattel11 variability in
problematic sections and c1ues 1'01' conelation uncertainties. Parallel core processing is suppen'tcd bythe speed 01' experimental data acquisition in envinmmental magnetism.
Partial 01' total carbonate dissolution does notaffect the stability ofthe magnetic mineral fractJon.However, without carbonate or opal sedimentation(pelaglc c1ay) the most important mechamsm ofsusceptibility modulation, the mutual dilution ofterrigenous and biogenic sediment fractions (Robinson 1990; Diester-Haass 1991; Mienert and Chi1995) is missing. Unless orbital variations inducecompositional changes within the lithogenic fraction, concentration-dependent magnetic parameterswill merely reflect consolidation \vith depth. In thisas well as other cases ofpurely terrigenous sedimentation, it is advisable to base orbital tuning uponmagnetomineralogic (e.g., hematite-magnetite ratio) 01' -granulornetric parameters (e.g., SOH 01'
M IM). These records often carry 21 climatic sig-;}l Ir
nal related to varying somce regions 01' transport(oscillating eolian 01' glaciomarine fluxes, bottomcurrent activity).
In oceanic regions with very low, inegular 01'
high sedimentation rates, rock magnetic records deviate more or less severely from standard pattel11s(as do other types 01' records). A promising agemodelling strategy is to start out with establishingsusceptibility-based core-to-core conelation frameworks. When 2111 inconsistencies in this multiplecorrelation scheme are resolved, the most convincing and complete record is dated by the best available method and may then serve as initial age model1'01' subsequent tuning ofthe entire collection.
In principle, magnetostratigraphy provides idealinitial age models 1'01' orbital tuning of rock magnetic parameters. Paleo- and rock magnetic investigations can share the same sampIe set which isadvantageous for cross-validation. At least onecore reaching Matuyama age within 21 set 01'
magnetically paralleled cores is needed to assign 21
rough age frame to 2111 others. Although numerousShmi inverse polarity events ofthe Brunhes Chron« 0.78 Ma) have been detected (Nowaczyk et 211.
1994) their chronostratigraphic use is severely hampered by lithology-dependent lock-in effects (Bleiland von Dobeneck, this volume).
26 Chapter 2
Extended Time Series Analyses
Having tuned a rock magnetic record to basic astronomical cycles in the Milankovitch range givesaccess to time series analyses into the adjaccntfrequency bands. These are essential to investigatethe temporal evolution and physical principles 01'proxy response to orbital forcing. Suitable methodsto analyze super- (> 100 kyr) and sub-Milankovitch« 18 kyr) signal variations with and without thevie'Vvpoint 01' orbital forcing are presented in thefollowing.
Evolutionary spectral analysis requires long timeseries and is therefore demonstrated for the SUSASstack covering the last 1.5 m.y. The high temporalresolution ofthe CEARIS stack makes it a suitableexercise for spectral analyses in the sub-Milankovitch frequency range.
Super-lvfilankovitch Signal Variations 0/theSUSAS Stack
Linear orbital forcing through insolation variationscan only accOlll1t 1'01' proxy oscillations in the precession and obliquity band (Imbrie et al. 1992). The96 kyr, 127 kyr and 404 kyr eccentricity cyclesmodulate the precession amplitude, but producenegligible insolation changes in their own frequencyrange (Fig. 12a). Yet, the SUSAS stack like otherPleistocene climate records exhibits strong 100 kyrvariance (Fig. 12c).
Many attempts have been made to explain this, 100 kyr cycle problem', frequently invoking nonlinear processes channeling energy into the 100 kyrband (1'01' a summary see Imbrie et al., 1993). Inrejecting the assumption that variations in insolation alone are responsible for the observed climatic
changes, MuHer and MacDonald (1995) explain the100 kyr cycle by inclination changes ofthe Earth' sorbital plane relative to the sym111etry plane ofthesolar system. Liu (1992; 1995) concludcs tha1 1'requency variations o1'the obliquity cycle can give riseto strang 100 kyr forcing 01' climate. Like variousothers, Raymo (1997) fa vors the idea that eccen··trici1y modulation ofthe precession index amplitudegenerates quasi -periodic 100 kyr ice age eyc1esThe climate changes from iee shield growth todecay, when summer insolation exeeeds a certainthreshold value. A 10ng-term reduction 01' the a1mospheric pC0
2level (Saltzman and Verbitsky
1993), eventually caused by tectonic processes(Raymo et al. 1988; Raymo ami Ruddiman 1992),is thought to have induced a s1eady rise of the insolation 1hreshold tor 1arge-sca1e melting in Quaternary times. This may exp1ain, why late Plioeene joearly Pleistocene paleoclimate records exhibitmainly obliquity and precession related variance(e.g., Raymo et al. 1989; Bloemendal anddeMenocal 1989; deMenocal 1995), while latePleistocene climate f1uctuations, in spite ofsimilarinsolation variations, are dominated by 100 kyrcycles (e.g., Hays ct al. 1976; Ruddiman et al.1989). The exact timing ofthis mid-Pleistocene climate transition and the question whether it was agradual Cmid-Pleistocene evolution', Ruddiman etal. 1989) or a sudden change Cmid-Pleistocene revolution', Berger et al. 1994) have been subject 01'many studies (Shackleton and Opdyke 1976; Pisiasand Moore 1981; Prell 1982; Ruddiman et al. 1989;Berger and Jansen 1994b; Mudelsee andStattegger 1997) and attempts to model or characterize the transition with different statistical techniques (e.g., DeBlonde and Peltier 1991; Park andMaasch 1993; Mudelsee and Schulz 1997).
Fig. 11. Temporal analysis ofvarying frequency content ofthe SUSAS stack. From bottom to top: Lowpass filtering of the normalized susceptibility stack reveals three distinct periods with different base levels and transitionscentered at around 0.95 and 0.6 Ma. Extracting the basic Milankovitch cycles by bandpass filtering depicts frequency shifts within the eccentricity band and fading of the obliquity related variance in early Pleistocene.Precessional amplitudes are generally low due to bioturbation damping, particularly in periods of lower sedimentation rates. The envelopes were derived by Hilbert transformation (Bendat and PiersoI1986). For the evolutionaryspectral analysis (see text) 500 kyr seetions ofthe highpass filtered SUSAS signal (purpie) progressing in 10 kyrsteps are tapered and transformed into the frequency domain. Related to the center of each section, the resultingspectra are color-coded and form a so-called 'spectrogram'. The colored horizontal bars represent the limits usedin the bandpass analyses below.
Using rock magnetic proxy records for orbital tuning and extended time series analyses 27
""""
1.6 1.71.2 1.3 1.4 1.51.0 1.10.8 0.9Age [Mal
0.4 0.5 0.6 0.70.0 0.1 0.2 0.3
0---1---""""""'"
i\ n A n1 1\ A 1\ 1\ 1/\
0.5 \JI !VI 1\ I~\ MI0:· llAi· \vv \~ V\0.5
0~_.Y1A
g -0.5
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.70.07--j--'-'-...L.L-l--'--'--'-'-~_
0.00
1
0.02
}~ 0.04>.CJcQ)
:::J 0.030-
~LL
0.05
0.06
:0~~ 0.5
~ 0~ -0.5~
~i.L
g:0~Q)CJlFl:::J
Cf)
goZ
28 Chapter 2
0.07 -00(1)...,
::::!. 2':0 .....c.. DltJl
0.06 ~~
19.00.05
.-.::' 22.4>- 23.7
..l<::-- 0.04.--
~ 29c<I>
5- 0.03<I>.....
LL41
0.0254
0.01 96127
0.000.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4
0.07 -00(1)...,
:::::!. sr.0 .....c.. DltJl
0.06~
~~
19.00.05
.-.::'22.4
~ 23.72. 0.04
>-<.> 29c<I>
5- 0.03<I>.....
LL41
0.0254
0.01 96127
0.000.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4
Age [Ma]
Fig. 12. Evolutionary spectral analysis of(a) insolation and (b) adapted ETP signal (see text for explanations).
Using rock magnetic proxy records for orbital tuning and extended time series analyses 29
0.07 -00CD ...,...,0-o' ;::;:c..~(J)
0.06~.=.
19.00.05
'i::'22.4
~ 23.7;:: 0.04
>-<.l 29cQ)
5- 0.03Q)....
u..41
0.0254
0.01 96127
0.000.3 004 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 104
0.07 -00CD...,:::l.2:0 .....c..~(J)
0.06~
~.=.
19.00.05
'i::'2204
~ 23.7;:: 0.04
>-<.l 29cQ)
5- 0.03Q)....u..
41
0.0254
0.01 96127
0.000.3 004 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4
Age [Mal
Fig. 12. Evolutionary spectral analysis of(c) SUSAS stack and(d) SUSAS-ETP residue (see text for explana-tions).
30 Chapter2
These controversies of 100 kyr cyclicity andPleistocene climate transitions essentially addressthe issue, whether long-term paleoceanographicvariations are related to basic Milankovitch cyclesand their superstructures (e.g. envelopes), or ratherto extraneous (forced by non-Milankovitch mechanisms) or free oscillations. A simple, frequentlyapplied method to analyze proxy signals in this context is multiple bandpass filtering. In the lower halfofFig. 11 five signal components ofparticular interest are extracted from the SUSAS record.
The prominent long-term trend obtained fromlowpass filtering (1 sI order Butterworth >350 kyr)is a remarkable feature not found in oxygen isotope signals and therefore not directly related toglobal ice volume. It is mirrored by the sedimentation rate stack (Fig. 5) and also by CaC0
3concen
tration (P. Müller, unpublished data), hence by carbonate accumulation. Resembling more a doublebase line shift than a cycle, this signal marks twotransitions:- a first from 1.0 to 0.9 Ma corresponds to the ageofthe mid-Pleistocene (r)evolution,- a second from 0.65 to 0.5 Ma culminates in thelarge-scale sedimentation event outlined in Fig. 5.
Their duration may have been shorter, since lowpass filtering inevitably broadens step functions.Thethree age intervals delimited by these two transitions show different spectral characteristics in theMilankovitch bands. The precession-related signal(16-26 kyr) is poorly documented due to bioturbation damping with the exception of sections ofhigher sedimentation rates and enhanced forcing.Apart from superimposed amplitude fluctuations,the influence ofobliquity (35-47 kyr) on the SUSASstack diminishes continuously with a major decrease near 1.2 Ma. The analysis in the 100 kyrrange was subdivided into two bands from 85 kyrto 110 kyr and from 110 kyr to 135 kyr to distinguishthe 96 kyr and 127 kyr eccentricity cycles. Theseperiods clearly dominate the signal in late Pleistocene, but also give important contributions to oldersections. As the amplitude variations in the 41 kyrand 100 kyr bands resemble in great detail the patterns found in 8180 records of Pacific and Atlantic sites (Park and Maasch 1993), we conclude thatthe observed features are of a global nature andclosely related to changes in ice volume.
Evolutionary spectral analysis is an excellentmethod to visualize the variability ofcyclic climatecharacteristics in arecord through time (e.g.,Pestiaux and Berger 1984; Bloemendal anddeMenocal1989; Mwenifumbo andBlangy 1991).Spectra generated within a moving window aredepicted in three-dimensional 'spectrograms' combining frequency and time domain (top ofFig. 11).These calculations were made using the'SPECGRAM' algorithm embedded in the'MATLAB Signal Processing Toolbox'. A 500 kyrframe was advanced at 10 kyr steps and, in orderto minimize cut-offeffects, tapered with a Hanningwindow, thereby focusing to the central section ateach step. Long-term trends (periods >350 kyr)were previously removed by highpass filtering(purple curve) to avoid disturbance by these unresolved signal components. The resulting data matrix was normalized to an average spectral densityof 1by dividing all values by the total matrix mean.This procedure merely scales all spectrograms uniformly and enables their comparison, but does notalter relative variations in the time or frequency domalll.
In Figs. l2a-d spectrograms of forcing andresponding variables are related. The July midmonth insolation signal at 65°N (Berger and Loutre1991) and a normalized ETP curve are shown asreference. ETP curves are artificial, but often employed target records composed by calibrating andadding the time series ofEccentricity, Tilt (obliquity) and Precession (Imbrie et al. 1984). Here, thecumulative spectral intensity in each band of theETP spectrogram was calibrated to equal that ofthe SUSAS spectrogram. Being based on the sameorbital variations, both reference signals exhibitidentical amplitude modulation patterns, but verydifferent spectral power in each band. In the eccentricity band this difference amounts to severalorders ofmagnitude - an expression ofthe 100 kyrcycle problem. Characteristic Milankovitch superstructures as the cyclic 400 kyr modulation of the100 kyr, 54 kyr and 23 kyr cycles and the synchronous decay of the 100 kyr and 19 kyr cycles areclearly illustrated in Fig. 12b. The faint 29 kyr cycle is a subordinate component ofthe obliquity signal and shows similarmodulation as the main periods near 41 kyr.
Using rock magnetic proxy records for orbital tuning and extended time series analyses 31
The numerieal differenee between the speetrograms ofthe SUSAS record and the ETP curve isdisplayed in Fig. 12d. For the two precessiona1bands at 19 kyr and 23 kyr, the assumption of aconstant proxy response appears justified, as bothrecords largely compensate. In contrast, the residues in the 41 kyr band show a variable responseto obliquity forcing. Two marked spectra1 intensitydeclines at 1.2-1.1 and 0.7-0.6 Ma (Fig. 12e) cannot be exp1ained as a result of steady response toorbital forcing (Fig. 12b) and suggest that the potential of obliquity to drive the observed carbonatedissolution eycles decreased stepwise during PIeistocene times. This finding is in accordance withanalyses made for 8 180 records (e.g., Ruddimanet al. 1989; Joyce et al. 1990) and implies that theclimatic system was less sensitive to obliquity forcing during the late Pleistoeene. The coincidence oftarget and proxy records in the 100 kyr band iseven lower, as indicated by larger residues andsteeper gradients. While the ETP model looses100 kyr power (Fig. 12b), the SUSAS stack, likemost climate records, documents an intensificationof 100 kyr oscillations in late Pleistoeene.
This trend is interrupted by an early Pleistoeene100 kyr maximum appearing between 1.2 and1.1 Ma. This 'premature 100 kyr ineident', alsoobserved in 8 180 records (e.g., Mudelsee and Stattegger 1997), merges with a phase of strong 70 kyreyelieity (Fig. 12c) which has also been found inother paleoeeanographic reeords (e.g., Ruddimanet al. 1989; Robinson 1990; Bassinot et al. 1994;Bolton et al. 1995). Two low-frequeney positivemaxima eentered at 0.95 and 0.6 Ma in the SUSASand the residual spectrogram (Fig. 12c, d) resultfrom remainders ofthe earlier diseussed base linetransitions of the long-term susceptibility trend(Fig. 11 bottom).
The most striking result ofthe residual spectrogram is the reciprocity ofspectral intensities in the41 and 100 kyr bands (dashed outlines in Fig. 12d).In relation to the stationary ETP model, the SUSAS100 kyr cyclicity develops over-proportiona11y inseetions with reduced response to obliquity (blueoutline) and retreats, where the response to obliquity is strong (yellow outline). These findings imply an exchange of spectral energy between obliquity and eccentrieity in overall aeeordanee with
insolation threshold models (e.g., Saltzman andVerbitsky 1993; Raymo 1997). Intensified northernhemisphere summer insolation eapable oftriggering major deglaeiations henee should result ti'om aninterferenee of obliquity maxima with preceSSlOnindex minima. At times, when the insolation threshold is relatively low, obliquity maxima alone willtrigger the withdrawal ofeontinental iee shields and41 kyr eyelieity prevails. When the insolationthreshold is higher, the required peak insolation 15
only reaehed by optimum interaetion ofobliquity andpreeession. As eeeentrieity modulates the preeession amplitude, suffieiently large preeessional peaksoccur only during 100 kyr eeeentrieity maximaeausing the observed 100 kyr elimate eycles.
Raymo (1997) assumes a linear, pCOlcontrol
led rise of the insolation threshold over the past1.2 m.y. and obtained a single transition from 41 kyrto 100 kyr predominanee at about 0.7 Ma. Our evolutionary speetTal analyses (Figs. 12e, d) imply, thatthis tumover was predated by a 100 kyr episodearound 1.2 Ma.
Sub-Milankovitch Signal Variation5' ofthe
CEARIS Stack
A most interesting feature of a11 CEARIS reeordsare eonsistent high-frequeney variations in the subMilankoviteh to millennial range. As shown inFig. 1Oe, the < 15 kyr signal eomponents extraetedby high-pass filtering ean be matehed in a11 detail.Again, speetral analysis is a promising strategy todetect the environmental faetors eausing this unusua11y regular pattem. Stoehastie proeesses wouldyield smeared 'eolored noise' speetra with poorlydefined maxima, while periodie signals are represented by mueh sharper peaks appearing at eharaeteristie frequeneies - under the premise that arecord is undisturbed, accurately dated, and covering enough time to provide ample spectral resolution. Because these eonditions are never fully metin practiee, elaborate statistiea1 teehniques havebeen deve10ped to distinguish harmonie and randomsignal oseillations (e.g., Thomson 1982; for appIieations see Yiou et al. 1991; Nobes et al. 1991;Cortijo et al. 1995).
Sinee all fundamental Milankoviteh periods exeeed 18 kyr (e.g., Berger and Loutre 1992;
32 Chapter 2
rn,m'..-:r-:N:N:+, +'t.....:v:C'0:N
~~~E!~:
,",,'
'N:~...-, :
r..... : +;"<t"n.r..... :0.i:NC0:N"
N~:~
:t:~::t
o
Period g g 0 0 0 lD N 0 co <D lD '7 (0 N ~ 0
[kyr] _lD",1~~I~_<D'-,I..-'"--,>7--+-,(0,,,,1P+~~~''c-~LL,N+I_I~I---',._~-,I--!-+--,-I-,-L-+I ~...JI--!~I,--~_Iy-~-+-(J)'-I +1~-'--,L-l'f_+.---+-,-L-,-',--!~~~'~
:-:;r:~ fj::~: :
!.....:rt,:...,.:m: "'<(0:"1~: : 0!;[::I:l:::'+: m: ~:+:~
~. ~: ~:~:~::s :~: c: c' c
.c (f)o (j)
'5 :-eo 0
-,,<::"0C 0cu oe:,- (j)
20..
(j)
~ -20Cf)
l-::~v~~ ~::j (9
j ~:J' "0.. .
Cf) -80
~ -100o
0.. -60"0(j)
cu -80Et; -100'w
-120
-140o
-20
-40
-60"""rrrrtTJ'rn+"'fn""'H-rr<+ri1-n-.rrrtnirrn'>b'ncion'i,'rrl~Ti-rnci-rrr~""'ci-n'r~rrior,+rn+rn'rr+n;,..rt,,:,..,.,;,,;,.lJ..i,.rt-rI',.;'f,-,rln'rrn-~
o 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 009 01 011 0.12 0.13 0.14 015 016Frequency [1/kyr]
Fig. 13. Reference spectra (Blackman-Harris taper, WOSA: 3 segments) offundamental orbital frequencies andtheir low-order (:::;3) combinations based on astronomical time series (2500-0 ka; Berger and Loutre 1991). (Sub-)Milankovitch periodicities (vertical dorted lines) are classified in groups (gray bars) of common order (index numbers) and origin (intermodulations (Int) and harmonics (Har». Crosses depict 6dB bandwidths and 90 % confidence intervals. (a) A simple nonlinear transform of (b) the 65°N July insolation signal shows spectral characteristics in common with (b), (e) obliquity and (d) precession signal, even in the sub-Milankovitch 8-18 kyrrange. Otherhigher-order cycles of (a) coincide with (e) squared precession and (f) coupled obliquity and precession.
Using rock magnetic proxy records for orbital tuning and extended time series analyses 33
co~
Q)
(IJu
(f)
uE:S'e:(IJDJo-l
:!:::N
b
(IJ....tlQ)Q.
(f)....Q)
:s:o0--0Q)
roE.~w
20 • •• •
101c((w"ioJm ... CD .
: :::. : : :::: :: :::: ::: :: :: ::"1""1""1"';"("':1":'1':""1"'''1';':1'':'i""'i",ii"i'I:"il"i'I""I"';I,""I""I""I'"
0.01 002 0.03 0.04 005 0.06 0.07 0.08 0.09 0.1Frequency [1/kyr]
roJe
: :: :
/M
Fig. 14. Extended spectral analysis ofCEARIS stack. References are (a) spectrull1 of Fig. 13a, (b) raw and (e) averaged periodograll1 ofsignal section spanning the CEARIS age range (377-6 ka). Ana10gous1y, (c) and (f) depict theraw and averaged periodograll1 ofCEARIS stack. In (d) the eight individual CEARIS periodograll1s are stacked. Theinitially narrow standard deviation band widens towards 11igher frequencies. The coherency spectrull1 (g) ofreference (e) and proxy signal (f) yields high squared coherencies not on1y far the basic Milankovitch periods but alsofar spectralll1axima near 30 kyr, 15 kyr, 13 kyr, 11.5 kyr, 10.5 kyr, and 9 kyr. COll1bined with the overall coincidence ofspectra1 patterns this identifies these CEARIS sub-Mi1ankovitch peaks as the hypothesized cOll1bination tones.
34 Chapter 2
Schwarzacher 1993) and known solar activity cycles Ce .g. sunspot, Haie ami Gleissberg eyeles) frequently deteeted in varved sediments have perIOds<200 years (Glenn and Kelts 1991), whieh penodieities may actua11y be expeeted in the speetralrange from 1 to 18 kyr? Besides fi'ee, quasi-penodie elimate t1uetuations, the most Iikely candidatcsare multiples ami eombination tones ofMilankovitehfrequeneies originating from nonlinear transformations in the response ehain 01' insolation, elimate,proxy physics and sedimentation.
'I-Iarmonies' result from nonlinear response toa single orbital eyeIe, while 'intermodulation frequeneies' arise from nonlinear eoupling 01' two ormore basIc orbital eyeles. The frequeneies ofthcseadditionalmodes are simply integer (k,l, 111, 11, ... )
eombinations ofthc fundamental foreing frcqueneiesf;,j~.f;.f~, '" (Le Treut and Ghil 1983):
f (k ,I, /11. 11, ) = kir +I/i +/IIj] +IIj4 +
k,l,/II,II E{ ... ,-I,O, 1,2, .. };
oreler = Ikl +1II +1/111 +1111+..The number ofpotential combination tones withinan investigateel frequeney band can be boosteel aelIib by admitting high order eombinations involvingnegative coefficients. Coneerning orbital foreing,only the lowest order eombinations within a givenfrequeney range are bound to represent substantial speetral power (e.g., Yiou et al. 1991). Allloworeler (:S: 3) eombination tones calculated from obliquity anel preeession and their realization and signifieance in the spectra ofvarious astronomical timeseries are shown in Fig. 13. The five most important eccentricity frequencies can be obtained as linear combinations ofthe basic precession frequencies (e.g., 11404 kyc l = 1122.4 kyc l
- 1123.7 kyc l;
Berger and Loutre 1992). It is therefore unnecessary to include them separately in the calculationofpotential orbital combination tones. I-Iowever, itis essential to distinguish not just two, but a11 threemajor precession frequencies.
Even pure obliquity and precession signals(Fig. 13c, d) and their transformation into latitudedependent insolation (Berger and Loutre 1994) involve (slightly) nonlinear process (Berger andPestiaux 1984) and display sub-Milankovitch peaksabove the noise floor of numerieal precision
(Fig. 13b). From the viewpoint ofphysieal ciimatemodels, non-linear response is rather the rulc thanthe exeeption. Not just iee-sheet dynamles CLeTreut and Ghil 1983), but also ü1ctors controllingmonsoon intensity (Short et al. 1991:. Crmvlcy d
al. 1992) favor the generation of orbital harmonics. Periods 01' 10-12 kyr, corresponding to a doubled preeesslOn cyele (simulated here by squaringthe precession index; Fig. 13e), appear in insolation (Fig. 13b: Berger and Loutre 1997) and proxyspeetra (Pestiaux et a1. 1988: Park et a1. 1993;Hagelberg et al. 1994), partieularly at low latitudes.The coupling ofobliquity ami precession (Fig. 131)generatcs charactenstie peaks in the 13-15 kyrband, whieh are also found in sediments (e,g ..Berger et al. 1991) and ice core reeords (Yiou etal. 1991). A simple non-linear mathematical transform oLm astronomical insolation signal (Bergeranel Loutre 1991) yields a nearly eomplcte speetrum 01' a11 theoretiea11y derived orbital eombinationtones (Fig. 13a) and may be used as unbiased reference to identify significant sub .. Milankovitchpeaks ofproxy spectra. (Fig. 14a). In order to provide equivalent frequency resolution with respectto the CEARIS spcctra, the underlying time serieswas reduced to the same time range (377 - Cl ka:Fig. 14b,e).
The spectral amplitudes 01' insolation (Fig. 13b)fade steeply in the sub-Milankovitch range, whilethose 01' its nonlinear transform (Fig.13a/14a) andthe CEARIS record (Fig. 14e) decrease much less.This discrepancy implies spectl'al power channelinginto the sub-Milankoviteh band. By visual eomparison ofthe synthetie spectra (Fig. 14b,e) with those01' the CEARIS stack (Fig. 14c,f) and the stack 01'the eight individual CEARIS speetra (Fig. 14d)many analogous peaks ean be deteeted. The mostprominent sub-Milankovitch peaks appear neal' 9,10.5, 11.5, 13 and 15 kyr. They eoineide with theearlier mentioned 100v-order eombination tones andharmonies ofobliquity and precession. A combinedevaluation ofmodel and proxy spectra and their coherence speetrum (Fig. 14g) indicates, that the coherence at these four sub-Milankovitch frequeneies is significant at an 80% level and thus nearlyas high as at the basic Milankoviteh frequeneies.This result is statistically valid, but is based upon asynthetic and somewhat arbitrary reference eurve.
Using rock magnetic proxy records for orbital tuning and extended time series analyses 35
Int 1123-1/41
Int 1/41+1123-1/19
Int 1119-1123=Ecc 1/109
Period 500 100 60 40 30 25 20 18 16 15 14 13 12 11 10 91//1 I" , I, , ! 1 1 I,!, ! 1 ! , 1 ! 1 ! 1 1 1 , 1 , 1 , 1 1 1 , ! ! I I[kyr] cn cn cn cn
~~ ~ ~-- -- -.
'+ ~
(')';; M~ (') cn M (') (') cn
~~~ ~ ~~
~~
~ ~
~o --~W '+ '+ '+ ""+ (') '+0,11 M~
d, (') cn d,~ ~ (') N~
~ ~ ~
~(')
~ ~~ ",,"N
~ -- ~ ~ ~ ~-. -. :;:::c::J N~ :c ~
ffi:E :E :E ~ ~~ :E ~ ~ :E ~ ~
0 E <:: <:: n. n. E E I <::
0.05
'C'g 0.04--.......
N.......~ 0.03cQ):::s0-Q) 0.02....u..
0.01
0.00
I 'o
, , I
0.01I '
0.02I ' I " I"" I " I'
O.ro O.M O.~ QOO QWFrequency /1 [1/kyr]
I ' I ' ,0.08 0.09
, I
0.1, I
0.11
Fig. 15. Bispectral analysis of CEARIS stack (see text for explanation).
A more direct approach is bispectral analysis(e.g., Rao and Gabr 1984; Nikias and Raghuveer1987). This higher-order statistical method wasemployed by Muller and MacDonald (1997) toanalyze 100 kyr cyclicity and by Hagelberg et al.(1991) for detection of sub-Milankovitch climatevariations. In contrast to power and cross spectra,which regard individual or compare equal frequeneies, bispectra are capable ofdetecting non-linearcoupling ofdifferent frequencies. A phase-lock between ahypothetical combination frequencyJ; andits assumed generating frequenciesJ; and}; is condition for a peak in the mirror-symmetric, two-dimensional diagram. Noise components and random oscillations are efficiently suppressed.
The bispectrum ofthe CEARIS stack (Fig. 15)displays a considerable number of peaks, mostlyatthe intersections ofgrid lines representing a simplified set of basic Milankovitch frequencies andtheir low-order combinations. A combination frequencyJ; can result from different pairs (fJ:). Forevident mathematical reasons these realizations arepositioned along diagonallines. Besides basicobliquity and precession peaks resulting from combinations oftheir own intennodulation frequencies,varlous low-order hannonics (as 2/23 kyr- I ) and intennodulations (as (1/41+1/23) kyrl
) are detected.The four m~or sub-Milankovitch peaks ofthe coherence spectrum (Fig. l4g) are validated as phaselocked combination tones of obliquity and preces-
36 Chapter2
o 20 40 60 80 100 120 140 160 180 200 220 240 260 280 300 320 340 360 380250 --!-'-T-'-~',-'-l-'-'--'-r'-'-'-'-.J,-L..L..L1+L-L.l---+-,---'--.L-'-+-~'-'-+--'--'---'--TL.Li--Y-..L.L.,+,c...L.L-"-T--'-'-L..y-'-'--'-'-..L.L.L.L.T'--'--'---'-f-..L.L.-',--'-'
200
150
100
::J500 U>o
§I450 g'~
400 320 40 60 80 100 120 140 160 180 200 220 240 260 280 300 320 340 360 380 ....!:l
Age [kaI
Fig. 16. (a) Susceptibility record ofGeoB 1523-1 (anows mark twin peaks conesponding to precession halIDonics).(b) idem, lowpass filtered (2nd order ButterwOlih) to exc1ude (> 18 kyr, dotted line) and inc1ude (> 7 kyr, solid line)sub-Milankovitch hmIDonics. (c) The residue ofboth filtered signals (7-18 kyr) frequently shows deep minima inphase with insolation maxima (dotted verticallines), followed by twin maxima (shaded as in b) before and afterinsolation minima. (d) The millennial band « 7 kyr) signal component appears to be aperiodic, although in someseetions, its envelope (shaded) resembles curve c (see also Fig. 16) (e) Precession index (solid line), shifted by-1.77 kyr (dashed line), and 15°N July insolation curve (shaded) by Berger and Louu'e (1991).
sion. Two further peak ehains near 65 kyr and30 kyr are identified as third-order intermodulations((1/41+1123-1/19) kyr J and(1I41-1I23+1I19) kyr 1
)
within the Milankoviteh band. Their important elimatie implieation will be discussed below.
In essence these results suggest that large partsof the observed sub-Milankoviteh signal components down to periods of at least 10 kyr can beexplained by non-linear c1imate response to orbitalvariations. Below 10 kyr 'subtone dispersion'comes into play (Fig. 14). Preeession and obliquityeyc1es are given as trigonometrie expansions ofseveral sinusoidal signals, most ofthem with verysimilar frequeneies (Berger and Loutre 1992;Sehwarzaeher 1993). The frequeneies oftheir seeond and third order harmonies are still c10se enoughfor their peaks to merge. With inereasing order, the
differenees beeome larger and eaeh intermodulation peak splits up into a peak ehain, eventually interfering with other peak ehains ... so that evenprominent and signifieant spectral peaks are nomore safely identified. Considering that for theCEARIS eores the estimated age uneertainty is inthe order of 2 kyr, there is no sense in searehingfor orbital effeets mueh belowperiods on kyr. Therelative increase of the noise eomponent withhigher frequeney, resulting from c1imatic and depositional random proeesses and stratigraphie elTor,is doeumented by the widening of the standarddeviation band in Fig. 14d.
In view of paleoeeanographie interpretation,signal phase and regularity are almost as importantas frequeney and amplitude. A band seleetive timedomain representation ofthe suseeptibility reeord
Using rock magnetic proxy records for orbital tuning and extended time series analyses 37
! I, ! f I ! ! ! ! I ! I ! ! I! '! I~~t ! ! ! ! I ! ! , I I ! , , ! j , !...J.....LJ.! I , I ! ! ! ! ! I I ! ! , ! , I LLLLJ...J-L1..L..L..LLLLLLLjI! ! I I!
50- Gi)(j)
<0
b,...- 100
'0QJt 150QJ
>c=~ 200
>,
'}0~
r---..vD~
~4=
~ 20
0'~
0ro
ce
0...er:l')
0 10 20 30 40 50 60Age [kaI
70 80 90 100 110
Fig. 17. Tentative conelation of 110-0 ka section of (a) susceptibility record of GeoB 1523-1, (b) idem, highpassfiltered (2ml order Butterworth < 7 kyr) to extract millennial scale variations, to (c) GRIP 8180 record on Dansgaardet al. (1993) age model. Labels mark Dansgaard-Oeschger cycles 1-24, YOlll1ger Dryas, and Heinrich events HI-H6.As considerable disagreement over details of the GRIP timescale persists (e.g., Hammer et al. 1997; .lohnsen et al.1997; Adkins et al. 1997), the conelation ages are not sufficiently reliable to evahiate the precision oforbital tuning.
of core GcoB 1523-1 is shown in Fig. 16 for theMilankovitch (> 18 kyr), sub-Mi1ankovitch range(18 to 7 kyr) and millennial range « 7 kyr). The7 kyr boundary is clearly a statistical and not apaleoclimatic limit.
The coherent orbital periodicities in the Milankovitch band (Figs.14b, c, 16b) suggest that globalglacial interglacial sea-level changes modulate theflux of(Fe-rich) Amazon sediments from the shelfregion into the adjacent pelagic realm. Only tropical insolation variations can explain the much higher23 kyr spectral power of K compared to 8 180(Fig. 8) and precessional susceptibility variationsprior to Northern hemisphere glaciation at theCeani Rise (HalTis et al. 1997). At low latitudesprecession large1y controls insolation (Fig. 16e,Berger and Pestiaux 1984) and modulates sea surface temperature, monsoon intensity and nutriclinedepth (e.g., McIntyre et al. 1989). Periods of elevated summer insolation result in increased car-
bonate accumulation (Rühlemann et al. 1996) andreduced Saharan dust deposition (Sarnthein ct al1981; Balsam et al. 1995) in the western equatorial Atlantic. Both effects lower the precessionalsusceptibility signal (Fig. 16c,e). In periods of reduced insolation, 'twin peaks' of susceptibility arefrequently observed - not just in the filtered(Fig. l6c), but also in the raw signal (anows inFig. 16a). They are therefore authentic paleoceanographic features and not artefacts of signal distortion by varying sedimentation rates (Martinsonet al. 1982; Schiffelbein and Dorman 1986). Detailed rock magnetic analyses (T. von Dobeneck,unpublished data) indicate that the earlier ofbothpeaks carries more Saharan, the second moreAmazonian characteristics. Together they reflectthe double precession cycle ofmonsoon intensity.
Further support to pC02
threshold models(' 100 kyr cyc1e problem'), discussed in the contextof SUSAS time series analyses, comes from the
38 Chapter 2
extraordinarily high bispectral intensities ofthe intermodulation frequencies (1/41 + 1/23-1/19) kyc l
and (1/41-1/23+1/19) kyc l. Both combination
tones express nonlinear interaction of obliquitywith the envelope ofthe composed precession signal, which has a frequency of(-1/23+1/19) kyc l
,
corresponding to an (averaged) eccentncity frequency of 1/109 kyc l
. This interaction provides thetrigger mechanism 10r the late Quaternary 100 kyrice age cycles.
Millennial climate cycles are believed to resultfrom free, self-sustained glaciomarine oscillationsas proposed by many coupled climate models (e.g.,Birchfield et a1. 1994; Paillard 1995). Large millennia1 temperature oscillations documented by icecore 8 180 studies from both hemispheres (Dansgaard et al. 1993; Bender et a1. 1994) have recent1yalso been 10und in various marine high-resolutionproxy records (Behl and Kennett 1996; Charles etal. 1996; Moros et a1. 1997; Adkins et a1. 1997),with some reservations also at the Cearä Rise(Curry and Oppo 1997). It is therefore not overJyambitious to tentatively cOlTe1ate the millennia1signal component ofGeoB 1523-1 (Fig. 16d) withthe GRIP 8 180 record (Fig. 17). Although bothpattems are not exactly minor images, the simi1arity is quite reasonable given the fact that a 2.8 mand a 2800 m record from two climate extremesare compared.
Conclusions
The SU SAS and CEARIS studies demonstratehow the methods ofcyclostratigraphy and environmental magnetism can be fruitfully combined.Other than most physical properties parameters,magnetic susceptibility depends linearly on theconcentrations of a few (iron-bearing) minerals,predominate1y (titano-) magnetite. In most marineenvironments its mean value and climatic modulation primarily mÜTors the ratio oftenigenous andbiogenic sediment accumulation. Susceptibilityvariations can be elosely related to carbonate dissolution cycles as in case ofthe SUSAS cores, butmayaiso primarily respond to climatic cyeles ofterrigenous supply as in the CEARlS example. Thebasic mechanisms of these two cycles are weIlstudied and largely understood. They are effects
of deep water formation, sea level change, aridityand wind intensity, all intimately linked to orbitalcycles. Although the phase relations 01'8 130 and K
differ regionally, the orbital response of susceptibility to climatlc change is o1'ten adequate forcyc10stratigraphy and orbital tuning. PiJot spectralstudies on selected cores yield applicab1e phase lagsand indicate best tuning targets.
Particularly in low-accumulation environmentsphase settll1gs are not as important as pattern reproducibility. Because susceptibility Jogging generates a complete low-noise image based on thesediment volume, subtle signal characteristics areresolved enabling pattem identification even in problematic sections. Aperiodic trends, shifts and amplitude fiuctuations of the susceptibility recorddocument modifications o1'the depositiona1 system.Extracted by lowpass filtering and evolutionm'yspectral analysis, these features may have moreinteresting climatic implications than a perfect orbital correlation scheme. In the SUSAS example,analogue long term trends 01' subtropical SouthAtlantic K and %CaCO, and global 8 l3C (but not(
180) records indicate marked changes at around0.95 and 0.6 Ma. Repeated converse shi1'ts 01'100 kyr and 41 kyr spectral power relative to anETP model substantiate pC0
2threshold concepts
(e.g., Sa]tzman and Verbitsky 1993; Raymo 1997)ofthe mid-P1eistocene transition.
In addition, most high-resolution susceptibilityrecords calTY important signal components in thesub-Milankovitch and millennial bands. In case ofthe eight CEARIS records these high-frequencyvariations are largely coherent and can be used toimprove the intemal precision of the combinedcyclostratigraphic age model. As shown by spectral and bispectral analysis, the signal modulationin the 7-18 kyr band is mainly due to periodic harmonics and combination tones of orbital frequencies such as the doubled precession cyele predictedfor tropical climate by models (Short et al. 1991).Similar spech'al interpretations in the < 7 kyr bandare inhibited by the complexity ofcombination tonesand uncertainty ofthe age model. A tentative correlation of millennial CEARIS variations toDansgaard-Oeschger cycles (Dansgaard et al.1993) yields quite convincing results within the limitations ofboth age models. Both categories ofhigh-
Using rock magnetic proxy records for orbital tuning and extended time series analyses 39
frequency variations open attractive perspectivesto improve chronostratigraphic precision beyondthe present state of the art.
Acknowledgments
We thank U. Bleil for many helpful disCllssions amiconstructive criticism. Reviews and many valuabJesuggestions by W.H. Berger and M. Schulz aregratefully acknowledged. T. Bickert kindly provided unpublished 8 11C data ofODP Site 806. Tbisstudy was funded by the Deutsche Forschungsgemeinschaft (Sonderforschungsbereich 261 atBremen University, contribution No. 219). F. S. wassupported by the Deutsche Forschungsgemeinschaft in the framewor1c ofGraduiertenkolleg 221.
Data are available under www.pangaea.de/Projects/SFB26l.
References
Aelkins JF, Boyle EA, Keigwin L, Cortijo E (1997) Variability of the North Atlantie thermohaline eireulation eluring the last interglaeial period. Nature390:154-156
Balsam WL, Otto-Bliesner BL, Deaton BC (1995) Modern and last glaeial maximum eolian sedimentationpatterns in the Atlantie Oeean interpreteel from seeliment iran oxide content. Paleoeeanography 10:493-507
Bassinot FC, Beaufort L, Vineent E, Labeyrie LD,Rostek F, Müller PJ, Quielelleur X, Laneelot Y(1994) Coarse fraetion fluetuations in pelagie carbonate sediments from the tropical Indian Oeean:a 1500-kyr reeorel of carbonate dissolution.Paleoceanography 9:579-600
Behl RJ, Kennett JP (1996) Briefinterstadial events inSanta Barbara basin, NE Paeifie, during the past 60kyr. Nature 379:243-246
Bendat JS, Piersol AG (1986) Random elata. Wiley, NewYork,pp 1-566
BeneleI' M, Sowers T, Diekson ML, Orehardo J,Grootes P, Mayewski PA, Meese DA (1994) Climatecorrelations between Greenland anel Antaretiea during the past 100,000 years. Nature 372:663-666
Berger A, Pestiaux P (1984) Accuracy and stability oftheQuaternary terrestrial insolation. In: Berger A,Imbrie J, Hays J, Kukla G, Saltzman B (eds)Milankovitch anel Climate Pali 1, Reidel, Dorelreeht,pp83-ll1
Berger A, Loutre MF (1991) Insolation values for the
climate of the last 10 million years. Quat Sei Rev10:297-317
Berger A, Meliee JL, Hinnov L (1991) A strategy for ti'equeney speetra ofQuaternary climate recorels. ClimDyn 5:227-240
Berger A Loutr'e MF (1992) Astronomieal solutions forpaleoelimate studies over the last 3 million years.Earth Planet Sei Lett 111 :369-382
Berger A Loutre MF (1994) Preeession, eceentricity,obliquity, insolation anel paleoclimates.ln: DuplessyJC, Spyrielakis MT (eds) Long-term climatie variations. Springer, Berlin, pp 107-151
Berger WH, Heath GR (1968) Vertieal mixing in pelagicseeliments. J Mar Res 26: 134-142
Berger WH, Yasuda MK, Biekert T, Wefer G, TakayamaT (1994) Quaternary time seale for the Ontong JavaPlateau: Milallkoviteh template for Oeean DrillingPragram Site 806. Geology 22:463-467
Berger WH, Jansen E (1994a) Fourier stratigraphy: spectral gain aeljustment of orbital iee mass models as anaid in elating late Neogene deep-sea seeliments.Paleoeeanography 9:693-703
Berger WH, Jansen E (1994b) Miel-Pleistoeene climateshift - the Nansen conneetion.ln: JoharUlessen OM,Muench RD, Overlanel JE (eels) The polar oceans andtheir role in shaping the global environment.Geophys Monogr Ser 85, AGU, Washington, pp 295311
Biekert T (1992) Rekonstruktion der spätquartären Boelenwasserzirkulation im östlichen Süelatlantik überstabile Isotope benthischer Foraminiferen. PhD thesis, Bel' FB Geo Univ Bremen 27,pp 1-205
Bickert T, Wefer G (1996) Late Quaternal'y eleep watercireulation in the South Atlantic: reeonstruetion fromcarbonate elissolution anel benthie stable isotopes.In: Wefer G, Berger WH, Siedler G, Webb DJ (eels)The South Atlantie: present anel past eireulation,Springer Verlag, Berlin, pp 599-620
Biekert T, Curry WB, Wefer G (1997) Late Plioeene toHolocene (2.6-0 Ma) western equatorial Atlantieeleep-water eireulation: inferenees from benthie stable isotopes. Proe ODP, Sei Results 154:239-254
Biekeri T, Berger WH, Wefer G (1998) TIle eleep westernequatorial Pacifie in Plio-Pleistoeene times: Resultsfrom Leg 130 (Ontong Java Plateau). Paleoceanography (subm)
Birehfield GE, Wang H, Rieh JJ (1994) Century/rnillenililllinternaI clinlate oseillations in an oeean-atrnospherecontinental iee sheetmoelel. J Geophys Res 99: 1245912470
Blaekman RB, Tukey JW (1958) The measurement ofpower speetr·a. Dover Publ Ine, New York, pp 1-190
40 Chapter 2
Bloemendal J, Lamb B, King J (1988) Paleoenvironmentalimplicalions ofrock-magnetic properties oflate Quaternary sediment cores from the Eastern EquatorialAtlanlic. Paleoceanography 3:61-87
Bloemendal J, deMenocal P (1989) Evidence for a changein the periodicity oftropical climate cycles at 2.4 Myrfrom whole-core magnetic susceptibility measuremenls. Nature 342:897-900
Bolton EW. Maasch KA, Lilly JM (1995) A wavelet analysis of Plio-Pleistocene climate indicalors: A newview of periodicity evolution. Geophys Res Letl22:2753-2756
Bond G, Broecker W, Johnsen S, McManus J, LabeyrieL, Jouzel J, Bonani G (1993) Correlations betweenclimate records from North Atlantic sediments andGreenland ice. Nature 365: 143-145
Charles CD, Lynch-Stieglitz J, Ninnemann US, FairbanksRG (1996) Climate connections between the hemisphere revealed by deep sea corehce core correlations. Earth Planet Sci Lett 142: 19-27
Chi J, Mienert J (1996) Linking physical property recordsof Quaternary sediments to Heinrich events. MarGeology 131:57-73
COliijo E, YiouP, Labeyrie L, Cremer M (1995) Sedimentary record ofrapid elimatic variability in the NorthAtlantic Ocean during the last glacial cycle.Paleoceanography 10:911-926
Crowley Tl, Kim KY, Mengel JG, Short DA (1992)Modeling 100,000 year climate fluctuations in prePleistocene time series. Science 255:705-707
Curry WB, 01'1'0 DW (1997) Synchronous, high-frequency oscillations in tropical sea surface temperatures and North Atlantic Deep Water production during the last glacial cyele. Paleoceanography 12:1-14
Dalfes HN, Schneider SH, Thompson SL (1984) Effectsofbioturbation on climatie spectra inferred from deepsea cores. In: Berger A, Imbrie J, Hays J, KuklaG, Saltzman B (eds) Milankovitch and Climate Pali 1,Reidel, Dordrecht, pp 481-492
Damuth JE (1977) Late Quaten1a1'y sedimentation in thewestern equatorial Atlantic, Geol Soc America BuH88:695-710
Dansgaard W, Johnsen SJ, Clausen HB, Dahl-Jensen D,Gundestrup NS, Hammer CU, Hvidberg CS,Steffensen JP, Sveinbjornsdotiir AE, Bond G (1993)Evidence for general instability ofpast climate froma 250-kyr ice-cOl'e record. Nature 364:218-220
DeBlonde G, Peltier WR (1991) A one-dimensional modelof continental ice volume fluctuations through thePleistocene: implications for the origin ofthe MidPleistocene climate transition. J Clim4:318-344
deMenocal P (1995) Plio-Pleistocene African climate.Science 270:53-59
deMenocal P, Bloemendal J, King .I (1991) A rockmagnetic record ofmonsoonal dusl deposition tothe arabü1l1 sea: evidence for a shirt in the modeof deposition at 2.4 Ma. Proc ODP. Sci Res\llts117:389-401
Diester-I-Iaass L (1991) Rhythmic carbonate contclltvariations in Neogene sediments above lhe oceaniclysocline. In: Einseie G, Ricken W. Seilacher A (eds.)Cycles and events in stratigraphy, Springer Verbg,Berlin, pp 94-109
Flanders PJ (1988) An allernating gradient force magnetometer,.1 Appl Phys 63:3940,3945
Franeois R, Baeon MP( I991) Variations in terrigenousinput into the deep Equatorial Atlantic during the past24,000 years, Science 25 1: 1473- I475
Glenn CR, Kelts K (1991) Sedimentary rhythms in lakedeposits. In: Einseie G, Ricken W, Seilacher A (eds.)Cycles and events in stratigraphy, Springer Verlag.Berlin, pp 188-22 I
Grousset FE, Labeyrie L, Sinko JA, Cremer M, Bond G,Duprat.l, Cortijo E, Huon S (1993) Patterns oficerafted detritus in the glacial North Atlantie (40-55°N).Paleoceanography 8: 175-192
Hagelberg TK, Pisias N (1991) Linear amI nonlinear couplings between orbital forcing and the marine BIROrecord during the late neogene. Paleoeeanography6:729-746
Hagelberg TK, Bond G, deMenocal P (1994) Milankovitch band foreing ofsub-Milankovitch climate variability during the Pleistocene, Paleoceanography9:545-558
Hammer CU, Andersen KK, Clausen HB, Dahl-JensenD, Hvidberg CS ,Iversen P (1997) The stratigraphicdating of the GRIP ice core, Special Report of theGeophysical Department, Niels Bohr Institute forAstronomy, Physics and Geophysics, University ofCopenhagen
Harris, FJ (1978) On the use ofwindows for harmonicanalysis with the Discrete Fourier Transform. ProcIEEE 66:5 1-83
HaITis SE, Mix AC, King T (1997) Biogenic andterrigenous sedimentation at Ceara Rise, westerntropical Atlantic, supports Pliocene-Pleistocenedeep-water linkage between hemispheres. Proc ODP,SciResults 154:331-345
Hays JD, Imbrie J, Shackleton NJ (1976) Variations inEarth's orbit: pace-maker of the ice ages. Science194:1121-1132
Hilgen FJ, Krijgsman W, Langereis CG, Lourens LJ,Santar'elli A, Zachar'iasse WJ (1995) Extending the
Using rock magnetic proxy records for orbital tuning and extended time series analyses 41
astronomical (polarity) time scale into the Mioccnc.Earth Planet Sci Lett 136:495-510
Horowitz LL (1974) The effects ofspline interpolationon power spectral elensity. IEEE Trans ASSP 22:2227
Imbrie J, Imbrie JZ (1980) Moeleling the climatic response to orbital variations, Science 207 :943-953
Imbrie J, Hays JD, Mal"!inson DG, Mclntyre A, Mix AC(1984) The orbital theory ofPleistocene climate: support from a reviseel chronology ofthe marine 8 180recorel. 111: Berger A, Imbrie J, Hays J, Kukla G, Saltzman B (eels) Milankovitch anel Climate Part 1, Reielel,Dorelrecht, pp 269-305
Imbrie J, Berger A, Boyle EA, Clcmens SC, Duffy A,Howarel WR, Kukla G, Kutzbaeh.J, Martinson DG,MeIntyre A, Mix AC, Molfino B, Morley JJ, PetersonLC, Pisias NG, Prell WL, Raymo ME, Shackleton N.J,Toggweiler JR (1992) On the strueture anel origin ofmajor glaciation cyc1es 1. Linear responses to Milankovitch forcing. Paleoceanography 7:701-738
Imbrie J, Berger A, Boyle EA, Clemens SC, Duffy A,Howarel WR, Kukla G, Kutzbach J, Martinson DG,McIntyre A, Mix AC, MolfillO B, Morley JJ, PetersonLC, Pisias NG, Prell WL, Raymo ME, Shackleton NI,Toggweiler .JR (1993) On the structure anel origin ofmajor glaeiation eyc1es 2. The 1OO,OOO-year eyc1e.Paleoeeanography 8:699-735
IolU1S WE, Schott FA, Zantopp RJ, Evans RH (1990) -D1eN Ol"!h Brazil Current retrofleetion: seasonal stmctureanel eelely variability, I Geophys Res 95 :221 03-22120
Iolmsen SJ, Clausen HB, Dansgaarel W, Gunelestrup NS,Hanmler CU, Anelersen U, Anelersen KK, HvielbergCS, Dahl-Jensen D, Steffensen IP, Shoji H,Sveinbjörnsel6ttir AE, White IWC, Jouzel I, FisherD (1997) The 8 180 recorel along the Greenlanel IceCore Project eleep iee eore anel the problem ofpossible Eemian elimatie instability. I Geophys Res102:26397-26410
Ioyee JE, Tjalsma LRC, Pmtzman IM (1990) High resolution planktie stable isotope reeorel anel speetralanalysis for the last 5.35 m.y.: Oeean Drilling ProgramSite 625, northeast gulfofMexieo. Paleoceanography5:507-529
Kumar N, Embley RW (1977) Evolution anel origin ofCeara Rise: an aseismie rise in the westem equatorial Atlantie, Geol Soe Ameliea BuIl88:683-694
Kent DV (1982) Apparent eorrelation ofpaleomagneticintensity anel c1imatic reeorels in eleep-sea seeliments.Nature 229:538-539
Le Treut H, Ghil M (1983) Orbital forcing, c1imatie interaetions, anel glaeiation eyc1es. I Geophys Res88:5167-5190
Lisitzin AP (1996) Oeeanie seelimentation: lithologyanclgeochemistry. AGU, Washington D.C.:400 pp
Liu HS (1992) Frequeney variations ofthe Earth 's obliquity anel the 1OO-kyr ice-age cyeles. Nature 358:397399
Liu HS (1995) A new view on the elriving mechanism ofMilankovitch glaciation cyc1es. Eal"!h Planet Sci Lett131: 17-26
Lomb NR (1976) Least-squares frequency analysis ufunequally spaeed elata. Astrophys Spaee Sci 39:447462
Martinson DG, Menke W, Stoffa P (1982) An inverseapproach to signal eorrelation. I Gcophys Res 87:4807-4818
Martinson DG, Pisias NG, Hays .JD, Imbrie J, Moore TC,Shackleton N.J (1987) Age elating anel the orbitaltheory ofthe iee ages: elevelopment of a high-resolution 0 to 300,000-year chronostratigraphy. Quat Res27:1-29
Maslin M, Mikkelsen N (1997) Amazon fan mass-transport eleposits anel unelerlying interglacial eleposits:age estimates and fan elynamics. Proc GDP, Sci Results 155:353-365
MatthewsonAP, Shimmielel GB, KroonD (1995) A 300kyr high-resolution arielity reeorel ofthe North African continent. Paleoceanography 10:677-692
Mclntyre A, Rueleliman WF, Karlin K, Mix AC (1989)Surface water response of the equatorial AtlanticOcean to orbital forcing. Paleoceanography 4: 19-55
Mienel"! J, Chi J (1995) Astronomical time-scale for physi-cal property recorels from QuatemalY seeliments ofthenorthem North Atlantic. Geol Rundschau 84:67-88
Milankoviteh M (1941) Kanon eier Erelbestrahlung unelseine Anwenelung auf elas Eiszeitenproblem. RoySerb Acael SpecPubl 133:1-633
Milliman JD, Slmmlerhayes CP, BalTetto HT (1975) Quaternary seelimentation on the Amazon continentalmargi.l1: a moelel, Geol Soc America Bu1l86:61 0-614
Moros M, EnellerR, LackschewitzKS, Wallrabe-AelamsHJ, Micnel"! J., Lemke W (1997) Physical propertiesof Reykjanes Rielge seeliments anel their linkage tohigh-resolution Greenlanel Ice Sheet Project 2 icecare elata. Paleoceanography 12:687-695
Muelelsee M, Schulz M (1997) Tbe Miel-Pleistocene climate transition: onset of 100 ka cyc1e lags ice volume builel-up by 280 ka. Earth Planet Sci Lett151:117-123
Muelelsee M, Stattegger, K (1994). Plio-lPleistocene climate mOelelling baseel on oxygen isotope time seriesfrom eleap-sea seeliment cores: The GrassbergerProcaccia algorithm anel chaotic c1imatic systems.Math GeoI26:799-815
42 Chapter2
Mudelsee M, Stattegger, K (1997) Exploring the structure ofthe mid-Pleistocene revolution with advanced111ethods of time-series analysis. Geol Rundschau86:499-511
Mulitza S (1994) Spätquartäre Variationen der oberflächennahen Hydrographie im westlichen äquatorialen Atlantik. Berichte, Fachbereich Geowissenschaften, Universität Bremen 57, pp 1-97
Muller RA, MacDonald GJ (1995) Glacial cycles andorbital inclination. Nature 377: 107-108
Muller RA, MacDonald GJ (1997) Simultaneous presence oforbital inclination and eccentricity in proxyclimate records from Ocean Drilling Program site 806.Geology 25:3-6
Muller-Karger F, McClain CR, Richardson PL (1988) '111edispersal ofthe Amazon's water. Nature 333:56-59
Mwenifumbo CJ, Blangy JP (1991) Short-term spectralanalysis of downhole logging measurements fromSite 704. Proc ODP, Sci Results 114:577-585
Nikias CL, Raghuveer MR (1987) Bispec1TUll1 estimation:a digital signal processing framework. Proc. IEEE75:869-891
Nobes DC, Blo0111er SF, Mienert J, Westall F (1991)Milankovitch cycles and non linear response in theQuaternary record in the Atlantic sec tor ofthe southern oceans. Proc ODP, Sci Results 114:551-576
Nowaczyk NR, Frederichs T, Eisenhauer A, Gard G (1994)Magnetostratigraphic data from late Quaternarysediments from the Yermak Plateau, Arctic Ocean:evidence for four geomagnetic polarity events withinthe last 170 ka ofthe Brnnhes Chron. Geophys J Int117:453-471
Oppenheim AV, Schafer RW (1989) DisCl'ete time signalprocessing. Englewood Cliffs, NJ, Prentice Hall, pp311-312
Park J, D'Hondt SL, King JW, Gibson C (1993) LateCretaceous precessional cycles in double time: awarm-Earth Milankovitch response. Science261:1431-1434
Park J, Maasch KA (1993) Plio-Pleistocene time evolution of the 100-kyr cycle in marine paleoclimaterecords. J Geophys Res 98:447-461
Paillard D (1995) The hierarchical struchue of glacialclimatic oscillations: interactions between ice-sheetdynamics and climate. Clim Dyn 11: 162-177
Peng TH, Broecker WS, Kipphut G, Shackleton N (1977)Benthic mixing in deep sea cores as determined by14C dating and its implications regarding climatestratigraphy and the fate of fossil fuel CO}.In:Andersen NR, Malahoff A (eds) The fate of fossilfuel CO
2in the oceans. Plenum Press, New York, pp
355-373
Pestiaux P, Berger A (1984) An optimal approach to thespectral characteristics ofdeep-sea c1imatic recordsIn: Berger A, Imbrie J, Hays J, Kukla Cl, Saltzman 13(eds) Milankovitch and Climate Part], Reidel..Dordrecht, pp 4] 7-445
Pestiaux PI, van der Mersch I, Berger A, Duplessy Je(1988) Paleoclimatic variability at frequencies ranging from 1 cycle per 10,000 years to 1 cycle per 1000years: evidence for nonlinear behaviour of the climate system. Clim Change 12:9-37
Pisias NG, Moore TC (1981) The evolution ofPleistoceneclimate: a time series approach. Emih Planet Sci Lett52:450-458
Pisias NG, Mix AC, Zahn R (1990) Nonlinear response inthe global climate system: Evidence üom benthicoxygen isotopic record in core RC 13··11 O.Paleoceanography 5:147-160
Prell WL (1982) Oxygen and carbon isotope stratigraphyfor the Quaternary ofHole 50213: evidence for twonlOdes ofisotopic variability. Initial Rep DSDP 68:455464
Radhakrishnamurty C, Likhite SD, Amin BS, SomayajuluBLK (1968) Magnetic susceptibility stratigraphy inocean sediment cores, Earth Planet Sci Lett 4:464468
Rao TS, Gabr M (1984) An introduction to bispectralanalysis and bilinear time-series models. Springer,NewYork
Raymo ME (1997) The timing of major climate transitions. Paleoceanography 12:577-585
Raymo ME, Ruddiman WF, Froehlich PN (1988) Influence of late Cenozoic mountain building on oceangeochemical cycles. Geology 16:649-653
Raymo ME, Ruddiman WF, Backman J, ClementBM,Martinson DG (1989) Late Pliocene variation innorthern hemisphere ice sheets and North Atlanticdeep water cireulation. Paleoceanography 4:413446
Raymo ME, Ruddiman WF (1992) Tectonic forcing oflate Cenozoic climate. NahIre 359: 117-122
Robinson SG (1986) The late Pleistocene paleomagneticrecord of North Atlantic deep-sea sediments revealed by mineral-magnetic measurements. PhysEarth Planet Inter 42:22-47
Robinson SG (1990) Applications for whole-core magnetic susceptibility measurements ofdeep-sea sediments. Proc ODP, Sci Results 115:737-771
Robinson SG, Maslin MA, McCave IN (1995) Magneticsusceptibility variations in Upper Pleistocene deepsea sediments ofthe NE Atlantic: implications for icerafting and paleocirculation at the last glacial maximum. Paleoceanography 10:221-250
Using rock magnetic proxy records for orbital tuning and extended time seIles analyses 43
Ruddiman WF, Raymo ME, McIntyre A (1986) Matuyama 41 ,OOO-year cycles: North Atlantic Ocean andnOlihern hemisphere ice sheets. Earth Planet Sci Lett80: 117-129
Ruddiman \VF, Raymo ME, Mmtinson DG, Clement 13M,Backman J (1989) Pleistocene evolution: narthernhemisphere ice sheets and North Atlantie acean.Paleoeeanography 4: 353-412
Rühlemann C, Frank M, HaIe W, Mangini A, Mulitza S,Müller PJ, Wefer G (1996) Late Quaternary produetivity changes in the western equatorial Atlantie:evidence from 2.'oTh-normalized carbonate and organic carbon accumulation. Mal' Geology 135: 127152
Saltzman13, Verbitsky MY (1993) Multiple instabilitiesand nlOdes of glacial rhythl11icity in the Plio-Pleistocene: a general theory of late Cenozoie climatiechange. Clil11 Dyn 9: 1-15
Sarnthein MG, Tetzlaff G, Koopmann 13, Wolter K,Pflaumann U (1981) Glacial and interglacial wind regimes over the eastern subtropical Atlantic andnOlih-west Africa. Nature 293: 193-196
Scargle JD (1982) Studies in astronol11ieal time seriesanalysis. II: statistical aspects of spectral analysisofunevenly spaced data. Astrophys J 263:835-853
Scargle JD (1989) Studies in astronomical time seriesanalysis. III: Fourier transforl11s, autocorrelationfunctions, and cross-correlation functions of unevenly spaced data. Astrophys J 343:847-887
Schiffelbein P, D0I111an LR (1986) Spectral effeets oftimedepth nonlinearities in deep sea sediment records: ademodulation technique for realigning time anddepth scales. J Geophys Res 91:3821-3835
Schulz M, Stattegger K (1997) SPECTRUM: spectralanalysis ofunevenly spaced paleoclimatic time series. Camp Geosci 23:929-945
Schwarzacher W (1993) Cyclostratigraphy and theMilankovitch theory. Elsevier, Amsterdam, 225 pp
Shackleton NJ, Opdyke ND (1976) Oxygen-isotope andpaleomagnetic stratigraphy ofpacific eore V28-239late Pliocene to latest Pleistocene. In: Cline RM,Hays JD (eds) Investigations of Later Quaternarypaleoceanography and paleoclimatology. Mem 145Geol Soc Amer, Washington, pp 449-464
Shackleton NJ, Berger A, Peltier WR (1990) An altemative astronomical calibration ofthe lower Pleistocenetimescale based on ODP Site 677. Trans Royal SocEdinburgh: Ealih Sci 81 :251-261
Shackleton NJ, Hagelberg TK, Crowhurst SJ (1995)Evaluating the success of astronomical tuning: pit-
falls ofusing coherence as a criterion far assessingpre-Pleistocene timeseales. Paleoceanography10:693-697
Shackleton NJ, Crowhurst S (1997) Sediment fluxesbased on an orbital tuned time scale 5 Ma to 14 Ma,Site 926. Proc ODP, Sci Results 154:69-82
ShartDA, Mengel JG, CrowleyTJ, Hyde WT, North GR(1991) Filtering ofMilankoviteh cycles by Earth 'sgeography. Quat Res 35: 157-173
Tarduno JA (1994) Temporal trends ofmagnetic disso··lution in the pelagic realm: gauging paleoproductivity? Earth Planet Sci Lett J23:39-48
Tarduno JA, Wi!kison SL (1996) Non-steady state magnetic mineral reduction, chemicalloek-in, and delayedremanence acquisition in pelagic sediments. EarthPlanet Sei Lett 144:315-326
Thiessen W (1993) Magnetische Eigenschaften desöstlichen Südatlantiks und ihre paläozeanographische Relevanz. PhD thesis, Bel' FB Geo Univ Bremen,41,pp 1-170
Thompson R, Clarke RM (1993) Quantitative marine sediment core matching using a modified sequence-slotting algorithm. In: Hailwood EA, Kidd RB (eds) Highresolution stratigraphy. Spec Publ Geol Soc London70, pp 39-49
Thomson, DJ (1982) Spectmm estimation and harmonicanalysis. hoc IEEE 70: 1055-1 096
von Dobeneck T (J 996) A systematic analysis ofnatural mineral assemblages based on modelling hysteresis loops with coercivity-related hyperbolie basisfunctions. Geophys .I Int 124:675-694
von Dobeneck T, Freelerichs Th, Fabian K, Blei! U (1996)Rock Magnetic Micro-Scanning mit HTSL-SQUIDGraeliometer, VDI-Technologiezentrum (eels) Supraleitung und Tieftemperaturtechnik 1996. VDI-VerlagDüsselelarf(Germany), pp 51-54
Weeks R, Laj C, Endignoux L, Fuller M, Roberts A,Mangamle R, Blancharel E, Goree W (1993) Improvements in long-core measurement techniques: applications in paleomagnetism anel paleoceanography.Geophys .I Int 114:651-662
Welch PD (1967) The use offast Fourier transfOllli forthe estimation ofpower spectra: a method based ontime averaging over short, modified perioelograms.IEEE Trans Audio, Electroacustics, AU-15 :70-73
Yiou P, Genthon C, Ghi! M, .Iouzel J, Le Treut H, Bamola.IM, Lorius C, Korotkevitch YN (1991) High-frequency paleovariability in climate anel C02 levelsfrom Vostock ice core recoreIs. J Geophys Res96:20365-20378
44
Mid-Pleistocene climate transition:initiation, interim state andterminal event (submittecl to Nature)
F. Schmieder, T. von Dobeneck & U. Blei!
Fachbereich Geowissenschaften, Universität Bremen,
P.O. Box 330 440, 28334 Bremen, Germany
The mid-Pleistocene transition (MPT) of the globalcIimate system l -", initiated by a shift towards muchlarger northern hemisphere ice shields at around920 ka5 ami ending with predominance of 100 kyr iceage cyclicity since about 640 kaG-8, is one of the fundamental enigmas in Quaternary cIimate evolution9• Ofthe diverse explanations proposed for the 100 kyr cycIe,models invoking a pCOrcontrolled insolation thresholdfor the meIting of large ice shields lO- 14 currently yieldthe best reproduction of 8 180 signal pattern, aIthoughinconsistencies remain between oxygen isotope stages 15ami 13. Climate proxy records not excIusively linked toglobal ice volume are necessary to advance understanding of the MPTI5. Here we present high-resolutionPleistocene magnetic susceptibility time series of twelvesediment cores from the subtropical South Atlanticessentially ret1ecting variations in carbonate accumulation. In addition to characteristics known from 8180records, they reveal three remarkable features intimately related to the MPT chronology, (1) aprematureoccurrence of a near-100 kyr cycle around 1150 ka, (2) aMPT interim state of reduced carbonate depositionbound by transitions from 1000 to 920 ka and 640 to500 ka, and (3) a terminal MPT event at around 530 kadocumented in various unusual lithologies.
The twehre pelagic gravity cores recovered with R.V.METEOR in the framework of a long-term paleoceanographic research program (SFB 261 at the University ofBremen) form an oceanwide transeet spanning the subtropical South Atlantie between 22°S and 34°S (Fig. 1).Located on the Danks of major submarine elevations (RioGrande Rise, Mid-Atlantic Ridge, Walvis Ridge) in theelepth zone alternately inDuenceel by North Atlantie DeepWater (NADW) and Circumpolar Deep Water (CDW), theseeliments provide detailed records of carbonate elissolutioneyeles related to changes in deep water ehemistry. Oligotrophic open ocean eonditions result in low average sedimentation rates of 0.5 to I cm/kyr at these latitudes. In suchenvironments, a elose inverse eorrelation of magnetic suseeptibility to carbonate content is typieally observedlG. Thedata were acquireel on split core halves using a Bartingtonspot sensor at a 1 cm spacing allowing a signal definitionand temporal resolution commonly not attained in 8180 or%CaC03 analyses.
The recoreIs were founel particularly suitable to refineinitial magnetostratigraphic age models by means of astronomical tuning l7 to orbital variations ealculateel by Bergerand Loutre l8 . A phase lag applied (4.5 kyr to obliquity) was
Figure 1 Loeations 01' Geoß eores defining the SUSAS transect.The depth profile follows the dashed line in the Illap and is plottedagainst longitllde to depict affinities to the major basins 01' theSOllth Atlantie anel theil' deep water boelies. It is not sllitable toielentify positions anci depths 01' deep water passages.
determined by cross spectral analysis of 8180 and magneticsusceptibility of one of the cores (GeoB 1211-1, 8180 datafrom Bickert & Wefer I9). The perfect overall eonelation ofall twelve inelivielually tuneel records (Fig. 2) leel us toelefine the SUSAS (subtropical §outh 6tlantic ~uscepti
bility) stack.A prominent feature of this stack and all inelividual rec
orels is a baseline shift to 40% lügher average suseeptibililies during the MPT interim state (920 to 640 ka). Confinedby transitional intervals lasting about 80 anel 140 kyr at itsonset and termination, this period exhibits the most distinctdissimilarities between individual core logs. The precedingand following elimate states show similar mean susceptibilities modulateel by elearly eleveloped ancI coherenteyelieities of 41 and 100 kyr, respeetively. An exception is a'premature' near 100 kyr cyele around 1150 ka whieh wasalso identified in 8180 reeords and has been interpreted byMudelsee and Stattegger7 as an 'unsueeessful attempt of theelimate system to attain a non-linear Late Pleistocene ieeage state'. The MPT suseeptibility shift is notably mirroredin lowered average sedimentation rates (Fig. 2).
Evolutionary speetral analysis of the SUSAS stackexhibits a continuous Pleistoeene eleeline of 41 kyr eyelicity, but a eomparatively abrupt intensification of 100 kyrcyelieity after about 640 ka 17. Oxygen isotope reeordsdoeument larger iee shields sinee the beginning of the MPTinterim state at 920 ka. Why did 100 kyr eyelieity not startuntil 640 ka and henee lag the initial iee volume inerease byapproximately 280 kyr8? The observed baseline shift in theSUSAS reeords preeisely fills this time lag and should shedlight on this question.
The cause for the overall rise in magnetie susceptibilityeluring the MPT interim state is revealeel by a temporalanalysis of west-east trends. Plotted against longituele, eoremean suseeptibilities for the postulateel three elimate states(Fig. 3 top) reDect the well-known west-east asymmetry of
Mid-Pleistocene climate transition: initiation, interim state and terminal event 45
I41 kyr StateTran·sition
900
100 kyr State
100o 1000 1100 1200 1300 1400 1500 1600I, ,li !.I, i!."I!! ,I ! • .LL..u..Ll..u..u.I'!I,I,!! 1,1 ,r·! ,I,!,.l" !:"'I.J_U..L..LLd~~
150 IEarly Pleisloce~e 100 kyr Cycle I I I100 ':UJ I
50 GeOB1211-311o "--'__ +_,---c-, -II-_+-__~ ----- WO !.084 m
~~~~d"~A;tl GeoB 1034-J~II'0- 0,90 : :0,63 __--'W",O 3772 m~ ~ I
~~lMJ' ... .\mAAi T '! ~~:;~o Shift i ~:::i'~::~::::O~~:;O GooB 10354 1~ I01,18 ~ " WO 44_53 m
~!!t~J~~~ C;;~B4;~;~3 "
(J)~ ~~ooo~ ~ C;;~~;:~~6 ~L.:..C'-=-------- i-- --+-- , +-~'_--'__ __=.:=__=__=_=_'_'_'___I Q2
~ 400t_Ac-' AI A_ i iL,.r~"" i ~c 200k =VVV'''V-=VV'Ur : T Hi~ GeoB 3812-1 ~
1;::jO:~~1~~~G~::::::n3~~ J~.'~ : :0,57 ::0,68 WO 4331 m'~ 750-" ::: : ~C .O'l 500· UJ~ 250 eoB 814-6
o WO 4340 m
250
100
(I)
O--'-'-'c.:....:...-------------t --F-'---------,!\----;-:-:.:..:..... -r;--- -,;-----;------'-'-.=.c"-"-""--IF-'-I cE~. ~
c:, GeoB 1311-1 ~
o 0,39 ':0,34 ~iO,~ WO 2901 m ~
::~t~"~.. , '" ~,~~,,~ "~ ",i<"G {\" 0:: I ; \j.::J"IJc\)\J·V~OB 2820-2o 0.49 ; 10,37 ',0,74 WO 3615 m
:1 ~ ", : 18 ::14
:16" : 20 22: : 30 3,6 4~ 4,8l ... . " :'IJw-2 ,:" " SUSAS Stack
t 0 :J~~I~~=,~~~~:, ",!i~"",""" "',i"",i,~!;,;~«~:~;~i~;i~~!;~,~:;"'~~:,"~o 100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500 1600
Age [ka]
500
Figure 2 Individually tuned SUSAS records (WO = Water Oepth) and resulting susceptibility (arithmetic mean with standard deviationband) and sedimentation rate stack. A diamagnetic susceptibility of -15,10.6 SI units reprcscnting pure carbonate sediment was chosen as axisminimum To compensate for different signal levels in the stacking, each record was normalised by subtracting its mean and dividing by itsstandard deviation, Labels at the SUSAS stack indicatc even oxygen isotope stages, Horizontal dotted lines mark baseline averages for pre-,syn- and post-MPT states exclllding the bounding transitions, Ouring the interim statc, all cores display increascd susceptibilities and reducedsedimentation rates, Thc timing of the three intervals approximatcly corresponds to thc Laplacc (1800-1200 ka), Croll (1200-600 ka) andMilankovitch chron (600-0 ka) dcfincd by BergeT and Wefer5 Numbers in each section denote mean sedimcntation rates; grey vertical bars apremature 100 kyr cycle centred near 1150 ka and a terminal MPT evcnt fcatllring unusual1ithologies at about 530 ka,
46 Chapter 3
[!1 kyr State (> 1000 ka) ]
40'W 30'W 20'W 10'W 0' 10'E
2820-22.0 ,
2821-1
1.0
0.110'E 40'W 30'W 20'W 10'W 0'
1.21.11.00.90.8 2820-2
0.74&
1034-3• 0.6 1309-22821-1 ••0.5 •Ü 3812-1
1311-1
• 04 4&1211-3 3814-6
0.310'E 40'W 30'W 20'W 10'W 0' 10'E
051211;3
4&0.3
1034-3
• 0.2
•1034-3Enhanced SusceptibilityConstant W·E Gradient
Constant MagnetiteAccumulationConstant W·E Gradient
3814-62821-1 • 4&
4& • 3813-31309-2
•2820-2 4&3812-1
50
302010
0.6 381~ 3~33
0.5 2~21-"3~_2
DA 3812-1• •2820-2 Ü1311-1
300
200
100
0-'--,----,---,--,--,---,1-
10'E 40'W 30'W 20'W 10'W 0'
0"
3813-3
3'12-1•
•3814-6
[ 100 kyr State (500·0 k!J] [Interim State (920· 64~
30'W 20'W 10'W 0' 10'E 40'W 30'W 20'W 10'W 0' 10'E-,-'----'----'-----'--'-----"'-ri800-,-'--".,---"-L-".,---="-".,--".,-'--,-,-".,-'-----,,.,-"'-riJOO-r----'------'-----''----"-----'--,800
500
40"W 30"W 20"W 10"W
10'E 40'W 30'W 20'W 10'W 0'2 T"------'------'------'r,--------"---r'T-'-r 1.2 ,...L--'-------'--'---'---J,
1.1 Reduced Sedimentation10 Collapsed W.E Gradient0.9
0.8
07
•1.0 2820-2
0.5
0.3
0.1 -'--,---,---,--,-------,--,-'-040'W 30'W 20'W 10'W
(J)
co~::::J
E::::Joo«2 0.2~cOlco
::2:
~.Y. 3.0 3814-6
E 2.0 2821-1 : 38~-3OJ • 1309-2
O. 3J,-----,----,----,----,---,-'-0.3-'-,--,---,----,--,--,-1-
10"E 40"W 30"W 20"W 10"W 0"Longitude
Figure 3 Combined spatial and temporal trend analyses of magnetic susceptibility, magnetite aeeumulation (ealculated by splittingsuseeptibility into a diamagnetie background and ferrimagnetic, grain-size independent magnetite signal) and sedimentation rate. MajoraSy111111etries ancl shifts are summarised by white regression lines, dark-grey shading delimits mean data range, light-grey shadingwestern and castern South Atlantic eore sets. Three eores were exeluded from the analysis: eores GeoB 1035-4 and 3801-6, as they donot reaeh beyond the 530 ka event, and core GeoB 131 I-I recovered from 290 I m water depth clearly above the CDW/NADWtransition zone and therefore showing a somewhat different evolution.
terrigenous partic1e flux and accumulation in the SouthAtlantic20 . While the gradients are almost identical throughout Pleistocene, a shift to lügher susceptibilities is evidentduring the MPT interim state. This relative increase ofmagnetic mineral concentration cannot be explained in terms oftemporal changes in magnetite accumulation (Fig. 3centre), as its west-east decline from 2.5 to 0.3 g/m2kyrremains fairly constant over time. A time-varying dilutionof the terrigenous fraction by non-magnetic carbonate musttherefore be responsible far the observed shift.
Mean sedimentation rates (Fig. 3 bottom) deduced fromour age models are mainly controlled by carbonate accumulation. Even the westernmost core GeoB 2821-1 near theSouth American continent consists to 70 ± 9% of CaC03
(P. Mueller, pers. conun.). Both for the 41 and 100 kyrc1imate states, sedimentation rates vary from about
0.5 cm/kyr in the western to around 0.9 cm/kyr in the eastern South Atlantic. During MPT interim state sedimentationrates on either side of the mid-Atlantic Ridge are restrictedto between 0.4 and 0.6 cm/kyr. The dec1ine primarilyaffects the eastern part and brings the west-east asynunetryto collapse.
In the working area cyc1ic variations of the sedimentCaC03 content at Milankovitch frequencies are mainly dueto orbitally driven lysocline shifts resulting from an interplay of NADW and more corrosive CDW19. In the AngolaBasin (cores GeoB 1034-3, 1035-4 and 1729-3) the glacialinterglacial contrast is particularly manifest as the surrounding bathymetric highs restrict the access of CDWduring interglacials. The same reasoning should apply for10ng-term changes in carbonate accumu1ation. We thereforeassume that the influence of southern-source deep water
Mid-Pleistocene climate transition: initiation, interim state and terminal event 47
eurrents and intense produetivity blooms. Results from thcJoint Global Oeean Flux Study (JGOFS), based on photographs and measurcments from satellites, aireraft, ships, andthe Space Shuttle ATLANTIS, emphasise the potentiallyimportant role of oceanie frontal zones for the rapidaecumulation of diatom biomass28 They indicate dramaticbiological responses to eirculation and mixing processesassoeiated with open-oeean frontal systems separating coldfrom warm waters and gave rise to the idea that thiek diatom layers may be deposited in such environments.
The diversified terminal MPT events at around 530 kadoeumented in South Atlantic Oeean sediments might weHbe related to other paleoelimatie 'puzzles' like unusuallyhigh and low 8 180 values during isotope stages 13.2 and13.3, respeetively, found in the equatorial Indian Ocean29 ,an anomalous sapropellayer in the Mediterranean Sea dated528-525 ka30 and highest l11agnetic susceptibilitics of thepast 2500 kyr at about 500 ka in Chinese loess deposits31 ,il11plying an extreme1y warm and humid clil11ate.
The South Atlantic plays a key role for the global ther1110haline circulation. We have reasons, therefore, to assumethat the MPT phenomena documented in the SUSAS coresrespond to major global paleoceanographic shifts andsubstantially extend constraints on mid-Pleistocene e1imatereconstructions.
Acknowledgements. We are grateful to W.H. Berger forhelpful suggestions. This study was funded by the DeutscheForsehungsgemeinsehaft (Sonderforschungsbereich 261 atBremen University, contribution No. XXX). F.S. was supported by the Deutsche Forschungsgemeinschaft in theti'amework of Graduiertenkolleg 221.
I. Pisias, N.G. & Moore, T.C. The evolution of Pleistocene climate: a time series approach. Earth Planet. Sei.Lett. 52,450-458 (1981).
2. Prell, W.L. Oxygen and earbon isotope stratigraphy forthe Quaternary 01' hole 502B: evidence for two modes ofisotopic variability. [nil. Rep. DSDP 68, 455-464 (1982).
3. Maaseh, K.A. Statistical detection of the mid-P1eistocene transition. Clim. Dyn. 2, 133-143 (1988).
4. Ruddiman, W.F., Raymo, M.E., Martinson, D.G.,C1ement, B.M. & Backman, J. Pleistocene evolution:northern hemisphere ice sheets and North At1anticocean. Paleoceanography 4,353-412 (1989).
5. Berger, W.H. & Wefer, G. Klimageschichte aus Tiefseesedimenten. Naturwissenschaften 79,541-550 CI 992).
6. Berger, W.H., Yasuda, M.K., Bickert, T., Wefer & G.,Takayal11a, T. Quaternary time scale for the Ontong JavaPlateau: Milankovitch template for Ocean DrillingProgram Site 806. Geology 22, 463-467 (1994).
7. Mude1see, M. & Stattegger, K. Exploring the struetureof the mid-Pleistocene revolution with advanced methods of time-series analysis. Geol. Rundschau 86, 499511 (1997).
8. Mude1see, M. & Schulz, M. The mid-Pleistocene climate transition: onset of 100 ka eycle lags ice vo1umebuild-up by 280 ka. Earth Planet. Sei. Lett. 151, 117123 (1997).
544 546 548 550 552 554 556528
306 307 308 309 310 311 312 313 314 315 316
Core Position [em]
Figure 4 Terminal MPT event documented in a sharp colourtransition at unbioturbated finely laminated layers in core GeoE3812-1.
conelusion is supported by a simultaneous decrease ofglacial and interglacial kaolinite/chlorite ratios in coreGeoB 2821-1"1, interpreted to be induced by ehloriteenriched deep southern waters. Consequently, NADWshould have been reduced during that interval. Indeed, Öl3Crecords from Atlantic and Pacific ODP sites indicatesignificantly weaker NADW between 900 and 400 kaiS, Inview of this global evidence we suggest, that the MPTshould not be regarded as a gradual transition from a'41 kyr world' to a '100 kyr world', but rather as a third,contrasting climate state.
A conspicuous finding in several of our cores is the coincident occurrence of unusual sediment facies at around530 ka e10se to the end of the terminal MPT transition.Granulometric analyses of cores GeoB 2820-2 ancl 2821-]show a sharp grain-size shift with an almost total loss of the> 63 pm fraction 22, possibly related to the mid-Brunhesdissolution cycle23 . In core GeoB 2820-2 this episode corresponds to a short hiatus. The continuous Late Pleistocenerecord of core GeoB 1035-4 is interrupted by a thick turbidite dating to the same age. Core GeoB 3812-1 displays asharp colour change at a de1icately laminated horizon(Fig.4). Most impressive are thick interca1ated laminateddiatom ooze layers in cores GeoB 3801-6 (124 cm) and3813-3 (38 cm). Omitting these intervals in the susceptibility time series results in continuous Milankovitch signalsequences. Hence, the laminations must have been deposited in a very short time and should ret1ect an extremelyshort-term climate variability. Very high sedimentation ratesin the overlying carbonate sediments may have contributedto preserve the siliceous sections in both cores. The almostmonospeeifie layers eonsist of the giant diatomEthmodiscus rex (Rattray) Wiseman & Hendey (C.B.Lange, pers. cOlmn.) and are entirely uncommon at theselatitudes. Thick ooze deposits of this diatom were mainly
. I . f I ld 24-26reported from equatona regIOns 0 t 1e wor oceans .Even there, their occurrenee is puzzling, because Ethmodiscus rex very rarely oeeurs in plankton sampIes. However,enigmatie deep populations have been observed in thePaeifie Oeean27 . Several hypothesis have been proposed toclarify the paradoxiea1 'Etlunodiscus rex prob1em'26,ineluding differential dissolution, foeusing by bottom
48 Chapter 3
9. Imbrie et al. On the structure and origin of major glaciation cycles: 2. The 100,000 year cycle. Paleoceanography 8, 699-735 (1993).
10. Saltzman, B. & Verbitsky, M.Y. Multiple instabilitiesand modes of glacial rhythmicity in the Plio-Pleistocene:a general theory of late Cenozoic climatic change. Clim.Dyn. 9,1-15 (1993).
11. Berger, W.B. & .Iansen, E. Mid-Pleistocene climate shift- the Nansen connection. In The Polar Oceans and TheirRole in Shaping the Global Environment (eds..Iohannessen, M., Muench, R.D. & Overland, 1.E.) 295311 (Geophys. Monogr. 85, AGU, Washington, 1994)
12. Raymo, M.E. The timing of major climate transitions.Paleoceanography 12, 577-585 (1997).
13. Paillard, D. The timing of Pleistocene glaciations from asimple multiple-state c1imate model. Nature 391, 378381 (1998).
14. Raymo, M.B, Glacial puzzles. Science 281, 1467-1468(1998).
15. Raymo, ME, Oppo, D.W. & Curry, W. The midPleistocene climate transition: a deep sea carbon isotopic perspective. Paleoceanography 12, 546-559 (1997).
16. Robinson, S.G. Applications for whole-core magneticsusceptibility measurements oi' deep-sea sediments: Leg115 results. Proc. ODP, Sei. Res. 115,737-771 (1990).
17. von Dobeneck, T. & Schmieder, F Using rock magneticproxy records for orbital tuning and extended time seriesanalyses into the super- and sub-Milankovitch bands. InUse of proxies in Paleoceanography: Examples from theSouth Atlantic (eels. Fischer, G. & Wefer, G.), 601-633(Springer Verlag, Berlin, 1999).
18. Berger, A & Loutre, M.F Insolation values for theclimate oi' the last 10 million years. Quat. Sci. Rev. 10,297-317 (1991).
19. Bickert, T. & Wefer, G. Late Quaternary deep watercirculation in the South Atlantic: reconstructions fromcarbonate dissolution and benthis stable isotopes. In TheSouth Atlantic: present emd past circulation (eds.Wefer, G., Berger, W.B., Siedler G. & Webb, DJ.),599-620 (Springer Verlag, Berlin, 1996).
20. Lisitzin, AP. Oceanic sedimentation: lithology andgeochemistry. AGU, Washington D.e., 400 pp (1996).
21. GingeIe, FX., Schmieder, F, von Dobeneck, T.,Petschick, R. & Rühlemann, e. Terrigenous t1ux in theRio Grande Rise area during the last 1500 ka: evidenceof deepwater advection 01' rapid response to continentalrainfall patterns? Paleoceanography, 14,84-95.
22. Breitzke, M. Elastische Wellenausbreitung in marinenSedimenten - neue Entwicklungen der UltraschallSedimentphysik unel Seelimentechographie. Berichte,Fachbereich Gemvissenschaften, Univ. Bremen, 104,298 pp (1997).
23. Adelseck, e.G. .Ir. Recent and late Pleistocene sediments from the eastern equatorial Pacific Ocean: Sedimentation and dissolution. Ph.D. thesis, 192 pp, Univ. ofCalif., San Diego (1977).
24. Mikkelsen, N. On the origin oi' Etlunodiscus ooze. Mal'.Micropaleontol. 2, 35-46 (1977).
25. Stabeil, B. Variations of diatom flux in the easternequatorial Atlantic during the last 400,000 ycars("METEOR" cores 13519 andI3521). Marine (;eol. 72,305-323 (1986).
26. Gm'dner, 1.y. & Burckle, L.B. Upper Pleistocene Ethmodiscus rex oozes from the eastern eqllatorial Atlantic.Micropaleontol. 21,236-242 (1975).
27. Villareal, T.A Abllndance of the giant e1iatom Ethrno·discus in the Southern Atlantic Ocean and the ccntralPacific gyre. DiatOlIJ Res. 8, 171-177 (1993) .
28. Yoelcr , .I.A., Ackleson, S.G., Barber, R.T., Flament. P.& Baieh, W.M. A line in the sea. Nature 371, 689-692(1994).
29. Bassinot, Fe., Labeyrie, L.D., Vincent, E" Quidelleur,x., Shackleton, NJ. & Lancelot, Y. The astronomicaltheory of c1imate anel the age of the Brunhes-Matuyamamagnetic reversal. Earth Planet. Sci. Lett. 126, 91- J08(1994).
30. Rossignol-Strick, M., Paterne, M., Bassinot, FC.,Emeis, K.-e. & DeLange, G..I. An llnusual miel-Pleistocene monsoon period over Africa and Asia. Nature 392,269-272 (1998).
31. Bloemendal, 1., Liu, X.M. & Rolph, T.e. Correlation ofthe magnetic sllsceptibility stratigraphy of Chinese loessand the marine oxygen isotope record: chronologicaland palaeoclimatic implications. Earth Planet. Sci. LeU.131, 37J-380 (1995).
Cycles, trends and events ofPleistocene sedimentation in the oligotrophiesubtropical South Atlantie Ocean(to be subrnitted to Paleoceanography)
Frank Schmieder, Tilo von Dobeneck and Ulrich Bleil
Fachbereich Geowissenschaften, Universität Bremen, 28334 Bremen, (Jcrmany
Abstract. Most investigations of the mid-Pleistocene climate transition (MPT) were based on thc iee
volume proxy 8180. This study develops new aspects of this enigmatic period of Quaternary climate
evolution from twelve subtropical South Atlantic susceptibility records reflecting carbonate dissolution
cycles and hence changes in deep water chemistry. Comparative evolutionary spectra ofthese rock magnetic
proxy records and orbital forcing imply reciprocal exchange ofspectral energy between the 41 and 100 kyr
bands. A MPT interim state ofreduced carbonate deposition is defined from a temporal analysis ofwest
east-asymmetry. Presumably resulting from enhanced influence of southern source waters, this interval
exactly fills the time lag between the first occurrence oflarger glacial ice shields (~920 ka) and the onset
of 100 kyr ice age cyclicity (~650 ka). It ends with an unusual, probably global scale paleoclimatie episode
at about 530 ka, which is documented in several uncommon lithologies and interpreted as terminal MPT
event.
Introduction
The South Atlantic Ocean plays a major role in the
framewark of global thermohaline circulation.
'North Atlantic heat piracy' (Berger and Wefer,
1996), the interglacial exchange of cold North
Atlantic Deep Water (NADW) for warm surface
waters from the south, is essential far today' s warm
climate in northern Europe. Switching on and off
the Atlantic Heat Conveyor in the global ice age
rhythm also modifies conditions in the deep ocean.
Glacial reduction ofNADW gives way for a thicker
layer ofLower Circumpolar Deep Water (LCDW)
in the Atlantic Ocean and lifts the NADW/LCDW
boundary. Beside contrasting in temperature,
salinity and nutrient concentration these two deep
water masses differ in corrosiveness with respect
to calcium carbonate and hence in the preservation
of calcareous shells. The fluctuating NADWI
LCDW lysocline leaves carbonate dissolution
cycle's in transition zone sediments as documen
tation of paleoclimatic history.
The three major submarine ridges dividing the
subtropical South Atlantic into four pe]agic basins
are by far the best locations to find recordings of
this history as they intersect the transition zone
during glacia] as well as interglacial times. Penna
nent oligotrophie conditions in this 'open ocean
desCl't' keep the sedimentation rates as ]ow as 0.5
1 cm/kyr. On one hand, this makes it rather difficult
to build precise age models. On the other hand, gra
vity coring ofthe top 10m ofthe sediment column
yields continuous, nearly undisturbed records ofthe
whole Quaternary. These sequences offer the oppor
tunity to study not only cyclic changes but also long
tellli trends of Pleistocene climate evolution,
Still enigmatic within this epoch is the mid
Pleistocene transition (MPT) of the global climate
system (e.g., Pisias and Moore, 1981; Prell, 1982;
Ruddiman et a1. , 1989). In the course ofthe MPT,
the response to orbitally driven changes in insolation
received on Earth changed fundamentally. While
late Pliocene to early Pleistocene paleoclimate re-
50 Chapter 4
30 0 S
45°S '-'--__-'--_--'-__-'----_-----'__----'-__-'-----_-----L__---'---_-----'__---'---__-'-----_----'-_
50 0 W
[m]
1,000
o-1,000
-2,000
-3,000
-4,000
-5,000
-6,000
Figure 1. GeoB core locations defining the SUSAS transecL The depth profile follows the white dashed line in the map amiis plotted against longitude to depict the cores' affinities to the four major pelagic basins and their deep water bodies. It is notsuitable to identify the positions and depths of deep water passages.
cords exhibit l11ainly obliquity and precession rela
ted variance (e.g., RaYl110 et al., 1989; Ruddiman
et al., 1989; Bloel11endal and de Menocal, 1989),
which can be explained as linear responses (e.g.,
Imbrie et al. , 1992), the primm'y rhythm oflate Pleis
tocene climatic change with aperiod near 100 kyr
calls for more difficult explanations (for a summary
of research and proposed models see Imbrie et al.,
1993). The exact timing ofthe MPT and the question
whether it was a gradual ('mid-Pleistocene evolu
tion', Ruddiman et al., 1989) 01' a sudden change
('mid-Pleistocene revolution', Berger et al., 1994)
were subject to many studies, which attempted to
model 01' characterize the transition with different
statistical techniques (e.g., Maasch, 1988; DeBlonde
and Peltier, 1991; Park and Maasch, 1993; Mudelsee
and Schulz, 1997; Mudelsee and Stattegger, 1997;
Clark and Pollard, 1998). With few exceptions (e.g.,
Raymo et al. , 1997) all ofthese investigations were
based on oxygen isotope records. In our view, cli
mate proxy records not exclusively linked to global
ice volmne are necessary to advance understanding
ofthe MPT. Records ofcarbonate dissolution cycles
in the deep South Atlantic should provide an deci
sive complementary perspective of this phenome-
non.
In many marine environments, variations in car
bonate accumulation can be precisely traced by rock
magnetic methods. Here we present twelve high
resolution magnetic susceptibility time series of
Cycles, trends and events of Pleistocene sedimentation 51
-----,-------~
I Cruise report I
Wefer et a1. (1988)-j
Wefer et a1. (1988)
Wefer et al. (1990)
Pätzold et al. (1993)
Pätzold et al. (1993)
8chulz et al. (1992)
Bleil et al. (1994)
Blcil et al. (1994)
Wefcr et al. (1996)
Wefer ct al. (1996)
'J I VVeI'er et al. (1996)
~~~_~lj199§1_...J
Core METEOR Water Position Core
expedition depth (m) Latitude Longitude length (m)
GeoB 1034-3 M6/6 3772 21 °44.1'8 05°25.3'E 10.65
GeoB 1035-4 M6/6 4453 21°35.2'8 05°01.7'E 10.61
GeoB 1211-3 M1211 4084 24°28.5'8 07°32.0'E 8.61
GeoB 1309-2 M15/2 3963 31 °40.0'8 28°40.0'W 9.48
GeoB 1311-1 M15/2 2901 31 °30.7'8 29°05.9'W 7.42
GeoB 1729-3 M20/2 4401 28°53.6'8 () 1000.I'E 7.45
IGeoB 2820-2 M29/2 3615 30°49.4'8 38c 26.4'W 7.67
GeoB 2821-1 M29/2 3941 30°27.1'8 38°48.9'W 8.19
GeoB 3801-6 M34/3 4546 29°30.7'8 08°18.3'W 9.37 IGeoB 3812-1 M34/3 4205 31 °36.9'8 19°45SW 5.32
IGeoB 3813-3 M34/3 4331 32°16.1'8 21°58.0'W 9.83
GeoB 3814-6 M34/3 4340 34 0 11.0'8 28°38.0'W 7.95----
Table 1. Water depths, positions anel core lengths of the GeoB gravity cores presented in this paper.
Pleistocene sediment cores from the subtropical
South At1antic.
Material and Methods
The cores were chosen from material recovered
during six expeditions with R.Y. METEOR in the
framework of a long-term paleoceanographic
research program (SFB 261 at the University of
Bremen). This core selection covers the time range
of the MPT and fon11s a transect across the entire
deep South Atlantic Ocean between 200 S and 35°S
(Figure 1). Table 1 summarizes positions, water
depths and core 1engths. With the exception ofGeoB
1311-1, which was recovered fi-om 2901m, aIl cores
are located within the broad transition zone between
NADW and LCDW during glacia1 as weIl as inter
glacial times.
The westenm10st cores GeoB 2820-2 and 2821-1
are located on the east side of the Vema Channe1,
the predominant of two LCDW deep water passa
ges. Cores GeoB 1309-2, 1311-1 and 3814-6 lay
on the eastern flank of the Rio Grande Rise in the
vicinity of the Hunter Channe1, GeoB 3812-1 and
3813-3 on the western slope of the Mid-Atlantic
Ridge. As the flow of corrosive LCDW into the
Angola basin is to some extent restricted by sur
rounding ridges, especiaIly during interglacials (e.g.,
Bickert and Wefer, 1996), sedimentation there is
less affected by dissolution and sedimentation rates
are somewhat higher. While GeoB 1034-3 and
1729-3 from the northern and 1211-1 from the
southern flank of the Wa1vis Ridge reach back to
1200 ka, GeoB 1035-4 and 3801-6 do notl'each the
Brunhes/Matuyama boundary. They were selected
for this study as they contribute detailed records of
the last about 500 kyr and additionally document a
paleoceanographic event discussed later.
The pe1agic carbonate sediments mainly consist
ofnannofossi1 ooze with varying amounts offorami
nifera. Five cores were disturbed by one 01' two,
GeoB 3801-6 by five thin turbidites. Susceptibility
signal correlations showed that simply removing
the turbidites from the sequence results in complete
time series, only in GeoB 3812-1 arecord section
was lost. Siliceous fossils are generally negligible
in the sediments, but two cores contain an enormous
diatom ooze layer. In GeoB 3813-3 this layer is
38 cm thick, in GeoB 3801-6 even 124 cm. As in
the case of the turbidites, removal of the diatom
layers results in complete, correlatable susceptibility
records. Obviously, these 1ayers have been deposi
ted in relatively short time.
Paleomagnetic ana1yses were performed at
5-10 cm sampling intervals. Natural remanent mag-
52 Chapter 4
200 200
150R2 =0.83
150- . Ci(f) I . ü5 ill"? <0 70 D-0 . ". b 0=::. . •-.1.. ... =. :J
ill
.~ 100 ,:,..- . .-2 100 ur. ():fj '. j5 0
0. ß 80 :J
ill ~Q) 0ü if)(j) :J
~:J 50 (f) 50(f)
90
0 0100
100 90 80 70 60 50 0 200 400 600 800 1000 1200Carbonate content [%] Age [ka]
Figure 2. Comparision of magnetic susceptibility and carbonate content (T.Bickert, 1992) far care GeoB 1211-3. Far tbecross-plot on the left siele susceptibility elata were reduced to the 5 cm sampling intervall of the carbonte data. The inversecorrelation is very close in all period ranges, long-term trends anel Milankovitch cycles are weil reproduced. Due to bigherresolution, high-frequent signal features are expressed in more detail in magnetic susceptibIity.
netization was measured using a three-axis cryo
genic magnetometer (Cryo genic Consultants
GM 400) and statically AF demagnetized in ten
steps up to 100 mT (2G demagnetizer). The direc
tion of the characteristic remanent magnetization
(ChRM) was calculated by averaging over at least
three successive demagnetization steps.
Magnetic volume susceptibility K was measured
on the archive halves ofthe cores at a l-cm spacing
using a Baliington Instruments M.S.2.F spot sensor.
The suceptometer was operated in the sensitive
range and each measurement corrected with a sepa
rate background reading.
In most marine sediments susceptibility prima
rily quantifies magnetite content (Thompson and
Oldfield, 1986), which is generally part ofthe terri
genous input, although it may occasionally be
supplemented by bacterial magnetite (Petersen et
al., 1986, Vali et al. , 1987). The susceptibility signal
of marine sediment series frequently reflects the
ratio ofbiogenic and lithogenic components (Robin
son, 1990), which may vary due to changes in terri
genous input, carbonate production and dissolution,
or a combination of these often climatically con
trolled mechanisms. Susceptibility records have
been regionally established as excellent paleocli
matic proxies (e.g., Mead et al., 1986; Bloemendal
et al. , 1988; Bloemendal and de Menocal, 1989;
Park et al. , 1993; Robinson et al., 1995; Chi and
Mienert, 1996).
Ifmagnetic susceptibility is interpreted as inverse
carbonate proxy, all potential exceptions like sand
and ash layers 01' sections affected by reductive
diagenesis must be excluded (e.g., Bloemendal et
al. , 1989; Frederichs et al., 1999a). Except for the
earlier mentioned turbidites, there is no indication
of such effects in any of the twelve cores. Their
magnetic susceptibility records reflect variations in
carbonate content as the comparison of magnetic
susceptibility and %CaC03 (Bickert, 1992) records
of GeoB 1211-3 show (Figure 2). The inverse cor
relation of susceptibility and carbonate content is
very close in all period ranges. Long-term trends,
Milankovitch cycles and high-frequent signal fea
tures are weIl reproduced, although comparison of
the latter is restricted due to the much lower resolu
tion ofthe %CaC03 record (5 cm spacing). Only in
isotope stage 14 at about 550 ka considerable differ
ences appeal'. The extrapolated end-member values
for 100% and 0% carbonate are -5.6.10-6 SI units
Cycles, trends and events of Pleistocene sedimentation 53
and 464.1.10-6 SI units, respectively. The suscepti
bility of pure diamagnetic calcite, -15.10-6 SI units
(Thompson and Oldfield, 1986), is the theoretical
absolute minimum and logical origin ofthe suscep
tibility axis.
As mentioned above, the three factors goveming
the carbonate content of marine sediments are pro
ductivity fluctuations ofcalcareous organisms, dilu
tion by telTigenous material, and calcium carbonate
dissolution. Volat et al. (1980) analyzed Pleistocene
sediments in the Pacific, Indian, and Atlantic Ocean
and concluded that dissolution is the most impoliant
factor influencing carbonate content in all three
oceans. John50n et al. (1977) found carbonate cycles
in late Pleistocene sediments recovered in the Vema
Charmel and on the lower flanks ofthe Rio Grande
Rise at water depths between 2900 and 4000 m.
They evaluated the extent to which these cycles may
be dissolution controlled by comparing carbonate
content data and a semi-quantitative measure of
dissolution. Finding a strong cOlTelation between
low carbonate content and a high dissolution index
they interpret the variations in carbonate content as
dissolution cycles.
While in the South Atlantic carbonate dilution
by varying telTigenous sedimentation is important
near coasts and in regions of enhanced fluvial and
eolian sedimentation, it diminishes with distance
from the continents (Schmidt et al., submitted). In
spite of the fact, that telTigenous input is much
greater in the westem than in the eastem South
Atlantic (Lisitzin, 1996), even our westemmost core
GeoB 2821-1 consists primarily (70±9 %) of
CaC03 (P. Müller, unpublished data). Varying
terrigenous input certainly effects the susceptibility
signals to some extent, but carbonate dissolution
by alternating influence of NADW and more
cOlTosive LCDW is believed to be the main cause
of signal variance in the records presented here.
Chronostratigraphy
Magnetostratigraphies of GeoB 1034-3 and
1211-3 were established by Thiessen (1993) and
for all other cores by von Dobeneck and Schmieder
(1999). Except for GeoB 1035-4 and 3801-6 all
cores reach the Brunhes/Matuyama boundary
(780 ka) and the Jaramillo Event (990-1070 ka). In
three cases the short Cobb Mountain Event
(1190 ka) was detected despite low sedimentation
rates. Oxygen isotope records exist for
GeoB 1034-3,1035-4,1211-3 (Bickert, 1992;
BickertandWefer, 1996)and l309-2(W. Hale,per
sonal communication). Their stratigraphic interpre
tability is limited by very low sedimentation rates,
particularly in the deeper, obliquity-dominated sec
tions. By interpolating between the magnetostrati
graphic tie points, spectral characteristics ofMilan
kovitch cyclicity become apparent in the suscepti
bility logs. These initial spectra display significant
maxima at around 100 and 40 kyr, but show little
01' no indications ofa precessional signal component
(von Dobeneck and Schmieder, 1999).
A pattem matching to standard 8180 records, e.g.
from equatorial Pacific ODP Site 677 (Shackleton
et al. , 1990) was feasible over most signal sections,
but remained ambiguous between oxygen isotope
stages 16 and 13 between about 650 and 500 ka.
Continuously high similarities, including long-term
fluctuations of about 500 kyr duration, exist bet
ween all susceptibility signals and the benthic 813C
record at Pacific ODP site 806 (T. Bickeli, unpub
lished data) based on an orbitally tuned 8 180 strati
graphy (Berger et al., 1994). The complete and un
disturbed susceptibility record ofcore GeoB 3814-6
(Figure 3b) cOlTelates particularly weIl and helped
to substantiate the other eleven matches.
The most widely used orbital tuning procedure
assumes a phase lock between climate and metro
nome record (Martinson et al., 1987). In order to
assess the absolute lag of any proxy record with
54 Chapter 4
100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500Age [ka]
!~-1~,--~-+'------~+--+-----,~----'---+--'--'--+-----'-'--+-+-c---+---'-+--+-+---'- ' ',~_._._J
, ~~i_',_,_\ /_ !_i ! i! f !_~~__ \ I
ill:~1V1°1\MJ~~~\f\;Wr\
Polarity
813CODP 806(inverted)
Suscepti
bility K
Obliquity(inverted)
K
filtered15-47 kYr H'---r'-r-;+I--'r'+-'r-+'r+-r-'-r'+'++-.-'+-T'-'i-+'r+-r--T+'i-Y-',-'+'-r-+-r-r+,-,---r-'r+-r+i-'-r+'-+--1
o
TargetCurve
K
filtered
35-47 kyr 'Y-7-H+-~~+-~+-~~~+-~+-;-+-'----'c-'~--'-+c~',--+--!--,-c,-+,,-',-'--'--',-+-+T-~
Figllre 3. Four step refinement ofthe age model for core GeoB 3814-6. (a) Susceptibility record dated by magnetostratigraphy(3 tie points). (b) Signal pattern corre1ation to the tuned 81'C record ofODP Site 806 (T. Bickert, unpubl. data, 44 tie points).In critical intervals other SUSAS records were inspected for supplementary information (c) Susceptibility filtered (35-47 kyr)in the obliquity range (dashed line) and tuned (solid line) to an astronomical obliquity signal shifted by -4.5 kyr (73 tie points).(d) Susceptibility filtered (15-47 kyr) and tuned to an astronomical (obliquity and precession) target curve (95 tie points).
respect to astronomical obliquity and precession,
an independent and precise absolute chronology is
required. 8180 phase lags have been determined on
basis ofradiometric ages (Hays et al., 1976: 9 kyr
for obliquity, 3 kyr for precession) and assumptions
on the coupling of insolation and ice mass (Imbrie
et al. , 1984: 7.9 (7.4-8.2) kyr for obliquity, 5.0 (4.8
5.1) kyr and 4.2 (4.1-4.3) kyr for precession). To
estimate orbital phase lags for regional proxy para
meters such as susceptibility it is therefore sufficient
to know the phase lags relative to a 8180 curve from
the same core.
Here we use a revised 8 180 time scale of core
GeoB 1211-3 (Bickeli, 1992) based on a correlation
to the SPECMAP stack (Imbrie et al., 1984). Cross
spectral analysis of 8 180 and K yields a coherence
of 0.99 and 0.96 for the 100 kyr and 41 kyr cycles
and reasonable phase angles of -41 ° and -30°. This
implies that susceptibility leads 8180 by 3.4 kyr in
the obliquity band and therefore lags the obliquity
forcing function by 7.9 kyr-3.4 kyr=4.5 kyr and is
in good agreement with Imbrie et al. 's (1993) values
calculated for early proxy responses in the southern
hemisphere. The temporal resolution of the 8 180
record was too low to determine the precessional
phase lags within acceptable elTor margins.
The subsequent orbital tuning process was per
fonned in two stages. At first the obliquity-related
signal component was extracted applying a 1si order
butterworth bandpass (35-47 kyr) filter in forward
and reverse direction. Each maximum and minimum
in the filtered signal was assigned the age (+4.5 kyr)
Cyc1es, trends and events ofPleistocene sedimentation 55
Age [ka]
0.53
: : GeoB: : 2821-1
: : OA6
: : 0.74
:037 :
OA7: ,
:034: :
:051:
500 1000 1500
0,56
'0:8 : :GeoBiv:H .: 2820-2
0.49
: :GeoB: : 1311-1 -039
500 1000 1500
~.__••_ "',!""f--:: : •GeoB 1-
:1:0 :3813-3
0.70"':
: :GeoB: :3801-6
: : GeoB: : 1729-3
:OAO :
: :0.70
:0.47: :
: :089
. ,:GeoB'L :: 3812-1
TIH:~:
:0.37: ,: :050
500 1000 1500 0
500 1000 1500 0
104
0.69
5
5
5
5
5
:057: . I. r:068"-. t10 ---f-,-,-,-,--,i-,--,-.---c-,-,-..---.t- ---f-,-,-,-,--,i--r-rT':!'-.-n+
o-+--'-.-L.L--'--'+-.L-'-L.L...J-'---'--'-q: . GeoB. ·1034-3
~
~10 --h--.-r..,.--,'-rr-r-r-+-r-+cr-rl-- -t--,--,-,---,----r-r-r-.-r--r-r-..-rl
o-l:--'---'--'--'----S-'-7'----'---++-"--'--'-"---j- -k--'---'--'---,--*-,-,-,---,---+~---'--'--L-'---I-
10 --h--r-r--,-,i';-o.,--,-,o--r-r-,---,--'-+-- --h-,-,---,-,i..,..,--,---,-,---,..--rl-
o-t--'-'-J....L...J;;-lW---'-lL.L.L.L.L-'-f-- --k-'--'-L...L...J;--'-'-'----'--"--'---'---'--'--L.j--
10 --h--r-r--rl!e,-L,---.++-r-,---,--'-+-- --h--r-r--rl!e,-L,---.++-r-..--rl-
o---k-'-'-L--'-fi'-Y---'-';-+-'--'---'---'-f-- ---k-'-'-L--'-fi'-Y---'-';-+-'---'--'--L.j--
10 --h-,-,---,-,i..,..,--r-r-,---,-,---,,-f- --h--r-r..,.--,'-,-,--r-r-,---,-,---,rl-
o
E~ 10 --h-,-,-"""'';-'---r-r;-T--r--T-.--rl-- --h-'-T"""''''''-'---':--i-r''0+",5+-7rl--
ö.. 0 -k--'-.-L.L-'--If.!,.L-'-4--!-J--'--'-"---j- -k:'--'-L--'--4-.l,-J--'-4--!-J--'--'-"---j(])
o
Figure 4. Age depth relation ofall SUSAS cores. Sedimentation rates are generally fairly constant. Dashed lines mark theMPT interim state, bounded by two transitions. During thisinterval sedimentation decreased. The grey vertical bar denotesthe time ofa terminal MPT event indicated by unusuallithologies in several SUSAS cores (for abreviations see Figure 5).
of its postulated eguivalent in the astronomical
record (Berger and Loutre, 1991), starting from the
most uncritical sections near magnetostratigraphic
01' other reliable tie points. As illustrated in Fig
ure 3c, changes in the correlation age model are
small, typically less than a half-cyc1e of obliguity.
The second tuning step procceds from the observa
tion that faintly visible precessional peaks in the
spectra gradually sharpened at both refinements of
the age model (von Dobeneck and Schmieder,
1999). To benefit from this precessional signal com
ponent, a target curve of nOl1nalized obliguity and
precession index (mixing ratio of 2:3 chosen by
visual evaluation) and 4.5 kyr lag was calculated,
comprising notjust more, but also more prominent
signal features than either single orbital parameter.
The obliquity-tuned primary susceptibility record
was filtered using a wide bandpass (15-47 kyr) to
inc1ude obJiguity and precession cycles. As many
extrema as possible were l11atched (Figure 3d),
admittedly at the resolution limit of these records.
The resulting final age models are again j ust slight
moelifications ofthe previous models anel possibly
not even more precise in absolute ages as the pre
cession time lag is undetermineel. Nevertheless, this
higher-fiequency tuning leads to a bettel' mutual sig
nal cOlTelation ofthe twelve records.
Results and Discussion
Figures 4 and 5 sUl11l11arize the combined results of
magnetic age modeJing for the complete transect.
Resulting sedimentation rates are fairly constant
over time. A few question marks remain in core
sections, where individual signal features seem to
be missing 01' incomplete (e.g., GeoB 1311-1 at
420 ka, Ge oB 2820-2 at 800 ka). The high
confOlmity of all records led us to cOlnpute the
SUSAS (Subtropical.§outh Atlantic .§usceptibility)
stack, which is free of local effects anel therefore
representative for the whole Quatemary oligotrophic
56 Chapter 4
100 kyr State Interim State :!t~;~ 41 kyr State 1'---------------+--_f---- -1-_-l-________ _ J
o 100 200 300 400 500 600' 700 800 900 1000 1100 1200 1300 1400 1500 1600I, ! ! ! I ! ! ! ! I ! ! I , I ! ! ! , I ! ! , ! I I ! I ! I ! I ! ( I , I , ! I I I I , j ! ! ! ! ! ! ! ! ! I ! ! ! I I , ! ! i I ! ! ! , I , ! ! , I ! ! ! , I , ! ! ! I ! ! ! , I ! ; ! , ! , ! ! ! I ! , ! , I ! ! ! \ I ! I I I 11 ! I I I ' I ! I I I ! ! t I ! LU1J...uLLu.1..LLJ..l..1i.L.Lul~+-
::U~~1r1;;vvv~~~"OOkY~' ~~~ci:1~~i~""11~~"~~,!;:~ ~~B3i~;~31111
. .' , -----i·~i
75~ L Fine Laminated Layers Ii\j I;~ . . . .. 0 Laminated Diatom oo~ GeoB 1035-4 I:S: i
o· 1.18 .' •••_._'_.__ WO_445;3_JTL~ I
i~~ G"B1729"8i~!'0.47 :0.89 WO 4401 m
UJIGeoB 3801-6 ~
..........,- WO 4546 m ~
"""TH'~ GeoB 3812·' J-------~c___:__--c___:__----~.c-----------:---:0"--'-.5=---:0......,.,..______...________...__----WO 4205 m <;(
:E
"tAJGeOB 3813.3 :!WO 4331 m
500.
250 GeoB 1309-2 CI>
o---e:.:.:....:.-----------+-~-----i-~~---_______,Ä------:.--=-='---------=_______,_____,._--_______,-------'=----"--"'-"--"-___Ir___i ~200 ~
t::100 ~
(!J--c.:-=--=--------------:---c----;-:--:.:..:.-----------I\--------i-..::...:.-=------.--------,r-------------'--'-""----'=-"'---'--'-'--'-----1.!2
Cl::
21o·f iV-V·,' .,.f-'.. _."--d\",j;- -I,,·,· I··"'-"· .+VI,'.. j
-1
-2 ~ ""'.: : •
to:~~:,~"""""" "" "" """ I' ,C,. I" "",h1~" 'I '" '" ""1",, ,11,!~ C!:~~\,~i:,~~~~~~;~~,~~,~~"~,\~~,~ ,o 100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500 1600
Age [ka]
Cycles, trends and events of Pleistocene sedimentation 57
deep South At1antic Ocean. Because of different
amplitude ranges, the stack was determined as an
arithmetic mean ofnormalized (subtraction ofcore
mean and division by standard deviation), interpo
lated (Llt = 2 kyr) records.
Some important features of the SUSAS series
discussed in the following are visible by closer
visual inspection ofFigure 4:
• an east-west trend of increasing average mag
netic susceptibilities,
• changing signal cyclicities during the Pleistocene
with particularly uniform pattems during the
early (41 kyr state) and late Pleistocene (100 kyr
state),
• a shift towards enhanced magnetic susceptibili
ties and reduced sedimentation rates during the
intermediate time interval,
• the premature occurrence ofa near-1 00 kyr cycle
at about 1150 ka,
• a paleoceanographic event at about 530 ka
documented in unusual litho10gies in several
SUSAS cares.
Adapted statistical analysis must be applied to gain
insight into these phenomena and their interrela
tions.
Figure 5. (opposite) Individually tuned SUSAS records(WD = Water Depth) and resulting susceptibility (arithmeticmean with standard deviation band) and sedimentation ratestack. A diamagnetic susceptibility of -15 x10-6 SI units representing pure carbonate sediment was chosen as axis minimum. To compensate for different signal levels in the stacking, each record was normalized by subtracting its mean anddividing by its standard deviation. Labels at the SUSAS stackindicate even oxygen isotope stages. Horizontal dotted linesmark baseline averages for pre-, syn- and post-MPT statesexcluding the bounding transitions. During the interim state,all cores display increased susceptibilities and reduced sedimentation rates. The timing of the three intervals approximately corresponds to the Laplace (1800-1200 ka), eroll(1200-600 ka) and Milankovitch chron (600-0 ka) defined byBerger and Wefer (1992). Numbers in each section denotemean sedimentation rates; grey vertical bars apremature100 kyr cycle centred near 1150 ka and a telminal MPT eventfeaturing unusuallithologies at about 530 ka.
Cycles
Evolutionary spectral analysis is an excellent
method to visualize the variability ofcyclic climate
characteristics in arecord through time, either
presented as a sectionalized analysis (e.g., Pestiaux
and Berger, 1984; Bloemendal and deMenocal,
1989; Mwenifumbo and Blangy, 1991; Tiedemann
et al., 1994; Berger and Jansen, 1994) ar in the form
of'spectrograms' (e.g., Joyce et al., 1990; Yiou et
al. , 1991; Birchfield and Ghil, 1993; Glützner et
al., 1997; Harris et al., 1997; Paillard, 1998). These
three-dimensiona1 diagrams result from spectra
generated within a moving window and thus com
bine frequency and time domain. The SPECGRAM
algarithm embedded in the MATLAB Signal Pro
cessing Toolbox was used to calcu1ate the spectro
grams in Figure 6. For all time series analyzed a
500 kyr frame was advanced at 10 kyr steps and, in
order to minimize cut-off effects, tapered with a
Hmming window, thereby focusing to the central
section at each step. Long-term trends (periods
>350 kyr) were previously removed by highpass fil
tering to minimize disturbance by this signal compo
nent. The resulting data matrix was normalized to
an average spectra1 density of 1 by dividing all
values by the total matrix mean. This procedure
merely scales all spectrograms uniformly and
enables their comparison, but does not alter relative
variations in the time or frequency domain. As a
consequence ofthe window length used no spectra
cou1d be calculated for the first and the last 250 kyr
ofthe record. In order to avoid 10ss ofinformation
on the early Pleistocene the SUSAS stack was exten
ded beyond the age of 1530 ka by the normalized
signal of GeoB 1311-1 which reaches back to
1730 ka (Figure 4).
In the resulting spectrogram (Figure 6c) the
precession re1ated signal is poorly documented due
to bioturbation damping with the exception of
sections ofrelative1y higher sedimentation rates and
58
0.07
0.06
0.05
~ 0.04
>uc~~ 0.03
lL
0.02
0.01
0.00
Chapter 4
19.0
22.423.7
29
41
54
95124
300 400 500 600 700 800 900 1000 1100 1200 1300 1400 300 400 500 600 700 800 900 1000 1100 1200 1300 1400
0.07
0.06
0.05
};:0 0.04
>uc~~ 0.03
lL
0.02
0.01
0.00300 400 500 600 700 800 900 1000 1100 1200 1300 1400 300 400 500 600 700 800 900 1000 1100 1200 1300 1400
0.07
0.06
0.05
]<-'=. 0.04
>uc
~ 0.03~lL
0.02
0.01
0.00300 400 500 600 700 800 900 1000 1100 1200 1300 1400
Age[l<aJ300 400 500 600 700 800 900 1000 1100 1200 1300 1400
Age[l<a]
Figure 6. Evolutionary spectral analysis of driving and responding climate variables. The spectrograms were calculated byadvancing a tapered (Hanning window) 500 kyr frame at 10 kyr steps. All records were previously highpass filtered « 350 kyr)to minimize disturbance by low-frequency signal components.(a) Mid-month insolation at 65°N for July (Berger and Loutre,1991), (b) ETP curve, calculated from eccentricity, obliquity and precession by calibrating the cumulative spectral intensity ineach band to equal that of(c), the SUSAS spectrogram. (d) For a comparison offorcing and response the numerical differenceof (c) and (b) was calculated. In general, spectrograms of core GeoB 2821-1(e) and GeoB 3814-6 (f) displaya good overallcorrelation to each other and the SUSAS stack (c), slight deviations occur during the MPT interim state.
Cycles, trends and events ofPleistocene sedimentation 59
enhanced forcing (e.g., prior to 900 ka). The in
fluence ofobliquity on the SUSAS stack diminishes
continuously with a major decrease near 1200 ka
and a second, somewhat less enhanced reduction at
about 700 ka. Near 100 kyr periods clearly dominate
the signal in late Pleistocene since 650 ka, but also
give important contributions to older sections, par
ticularly expressed between 1250 and 1050 ka. In
this time range strong spectral components between
the obliquity and the eccentricity band emerge. At
about 1150 ka they fuse with near 100 kyr cyclicity
into a broad maximum. This merging is a result of
the compromise between time and fiequency resolu
tion. A conventional sectionalized spectral analysis
with a Ionger time window (1530-920 ka) separates
a component with aperiod ofabout 70 kyr (Figure 7,
top). Anomalous spectral peaks near 70 kyr have
been repOlied from Pleistocene paleoclimate records
ofthe North Atlantic (e.g., Ruddiman et a1., 1989),
the Indian (e.g., Robinson, 1990), and the Pacific
Ocean (e.g., Bassinot et a1., 1994a), but few attempts
have been made to explain their origin. Considering
the simultaneous presence of strong variance in the
100 and the 41 kyr band in the SUSAS record
(Figure 5, top and 7c), a nonlinear interference of
these two frequencies (1/41-1/100=1/69), pre
viously discussed by Robinson (1990) and Muller
and Mac Donald (1997), is a plausible explanation.
A simple nonlinear model of sea-level change
produces a spectral peak near 72 kyr beside moving
energy from the orbital forcing bands into the
approximately 100 kyr band (Berger et a1., 1996).
Similar to our results Mudelsee and Stattegger
(1997) identified high 100 and 41 kyr amplitudes
in the benthic oxygen isotope records ofODP sites
607 and 659 in the time period around 1200 ka.
They hypothesize that this interval of generally
stronger climate fluctuations was 'a first but
unsuccessful "attempt" of the climate system to
attain a nonlinear "late Pleistocene ice ages" state'.
Our spectrogram analysis strongly supports this
assumption.
The evolution ofthe 100 and 41 kyr cycIe docu
mented in the SUSAS spectrogram matches in great
detail the results of statistical analysis of benthic
oxygen isotopes of DSDP site 607 am] ODP sites
659 and 677 (Park and Maasch, 1993; Mudelsee
and Stattegger, 1997): strongest Pleistocene 41 kyr
amplitudes between 1400-1200 ka and nearly con
stant values since about 1100 ka as weIl as increas
ing 100 kyr cycles at about 650 ka and enhanced
values between 1200 and 1000 ka. Despite some
deviations the impressive correlation ofour analysis
and the calculations made for benthic 8 180 time
series from the midlatitude North Atlantic
(DSDP 607, ODP 659), and the equatorial Pacific
(ODP 677) suggests aglobaI character ofthe change
fr0111 41 to 100 kyr cyclicity. Evidently the cyclic
variations ofthe SUSAS stack are closely linked to
global ice volume via changes in deep water chemis
try.
How is the observed evolution of the cyclicity
related to the driving orbital fluctuations? In Fig
ure 6a-d spectrograms of forcing and responding
variables are compared. The mid-summer insolation
signal at 65oN (Figure 6a, Berger and Loutre, 1991)
and a norn1alized ETP curve (Figure 6b) are shown
as reference. ETP curves are artificial, but often
employed target records composed by calibrating
and adding the time series of Eccentricity, Tilt
(obliquity) and Precession (Imbrie et a1. , 1984).
Here, the cumulative spectral intensity in each band
ofthe ETP spectrogram was calibrated to equal that
of the SUSAS spectrogram. Being based on the
same orbital variations, both reference signals exhi
bit identical amplitude modulation patterns, but very
different spectral power in each band. In the eccen
tricity band this difference amounts to several orders
of magnitude - an expression of the 100 kyr cycle
problem (Imbrie et a1. , 1993).
60 Chapter 4
Figure 7. Spectral analysis ofthe SUSAS stack far the threePleistocene states defined in Figure 5.
taneous decrease in obliquity although the gradient
is steeper as in Figures 6a, b. The reduction near
700 to 600 ka and the subsequently low amplitudes
during the entire late Quaternary however do 110t
mirror the evolution ofthe forcing signal. Quite the
reverse, obliquity amplitudes increase again during
the late Pleistocene (Figure 6a, b) and thus praduce
negative residues (Figure 6d). Obviously the ability
of obliquity to drive the observed carbonate disso
lution cycles decreased during Pleistocene times and
remained low since the onset of 100 kyr cyclicity
at about 650 ka. This finding is in accordance with
the suggestion ofRuddiman et al. (1989) and Joyce
et al. (1990) that the climatic system was less sensi
tive to obliquity forcing during the late Pleistocene
and as well with the results of Imbrie (1992) who
found, that response in the obliquity band was fairly
constant over the past half-million years.
The coincidence of target and proxy records in
the 100 kyr band is even lower, as indicated by
larger residues and steeper gradients. While the ETP
modellooses 100 kyr power in the late Pleistocene
(Figure 6b), the SUSAS stack, like most climate
records, documents an intensification resulting in a
positive residue. The preceding residue maximum
originates from the early Pleistocene 100 kyr cycle
near 1150 ka. Two low-frequency positive maxima
centered at 950 and 600 ka in the SUSAS and the
residual spectrogram (Figure 6c, d) result from re
mainders of the earlier mentioned base line transi
tions ofthe long-term susceptibility trend (Figure 5).
The most striking result of the residual spectro
gram is the reciprocity of spectral intensities in the
41 and 100 kyr bands (dashed outlines in Figure 6d).
In relation to the ETP model, the SUSAS 100 kyr
cyclicity develops over-proportionally in sections
with reduced response to obliquity (blue outline)
and retreats, where the response to obliquity is
strong (yellow outline). These findings imply an
exchange of spectral energy between the obliquity
>:, >:, >:,>:, -" -" -"-" r- '<T 0
:; C') N 0)
N N
I I
I I
>, ~ ~
-" >, >,'<T -" -"N LO '<T~ 0) LO
I I I70 kyr
I I ~ I
0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07
Frequency [1/kyr]
20
o
10
-10
-10
-30
-20
co2-;:;.
20';nc -10Q)
u
~ 10Ü -20Q)CL
Cf) 0-30
The numerical difference between the spectro
grams of the SUSAS stack and the ETP curve is
displayed in Figure 6d. With restrietions due to the
overalliow variability in the 19 and 23 kyr bands,
the assumption ofa constant proxy response appears
justified for precession, as both records 1argely
compensate throughout the whole record. This
hypothesis can also be drawn from simple visual
comparison ofFigures 6a and c. The reduced 23 kyr
maximum centered at 600 ka and enhanced values
in the 19 kyr band near 400 ka may result from the
influence ofdifferent sedimentation rates (Figure 4)
on the definition of these high-frequency cycles.
41 kyr cyclicity is less affected by bioturbation
damping (von Dobeneck and Schmieder, 1999).
Yet, residues in the obliquity band show a more
variable, non-stationary response to forcing
(Figure 6d). The spectra1 intensity decline at 1200
to 11 00 ka (Figure 6c) may result from the simul-
Cycles, trends and events of Pleistocene sedimentation GI
and eccentricity band. A possible explanation is
provided by climate models invoking achanging
insolation threshold (e.g., Saltzman and Verbitsky,
1993; Raymo, 1997; Paillard, 1998). Intensified
northern hemisphere summer insolation capable of
triggering major deglaciations hence should result
fl:om an interference ofobliquity maxima with pre
cession index minima. At times, when the insolation
threshold is relatively low, obliquity maxima alone
will trigger the withdrawal ofcontinental ice shields
and 41 kyr cyclicity prevails. When the insolation
threshold is higher, the required peak insolation is
only reached by optimum interaction of obliquity
and precession. As the precession amplitude is
modulated by eccentricity, sufficiently large pre
cessional peaks occur only during 100 kyr eccentIi
city maxima suppressing obliquity and causing the
observed 100 kyr climate cycles. While a linearly
increasing threshold (e.g., Raymo, 1997; Paillard,
1998; Raymo, 1998) results in a single transition
from 41 to 100 kyr predominance, a slightly fluc
tuating threshold should also be capable to explain
the 100 kyr excursion near 1150 ka.
The good overall correlation of all records dis
cussed in the time domain is also visible in the
synchronous time-frequency view of individual
spectrograms. Slight deviations in the spectrograms
of cores GeoB 2821-1 and 3814-6 possibly relate
to local effects (Figure 6e, f) and are concentrated
in the MPT interim state, where both records exhibit
higher variance in the obliquity band than the stack.
In GeoB 3814-6 this response includes a strong
54 kyr component, which is also visible in
GeoB 1211-3 (not shown) and in the spectrum of
the corresponding SUSAS stack section (Figure 7,
center). In both cores this subordinate obliquity
cycle amplitude reaches values comparable to the
41 kyr cycle although the latter, summarizing four
major terms, contributes six times as much to
astronomical obliquity variation (e.g., Berger and
Loutre, 1992). The mid-Pleistocene increase in
response to obliquity is even more puzzling as it
precisely coincides with a time of reduced ampli
tudes in the forcing function (Figure Ga, b). The
three diverging cores were recovered from southern
ridge slopes 01' deep water passages (Figure 1) and
are thus more intensely influenced by southern
source waters. The influence ofobliquity-dominated
high latitude climate forcing was possibly intensi
fied during the MPT interim state.
The evolutionary spectral analysis ofthe SUSAS
stack documents a continuous Pleistocene decline
of41 kyr cyclicity, but a comparatively abrupt inten
sification of 100 kyr cyclicity after about 650 ka.
Oxygen isotope records indicate larger ice shields
since the beginning of the MPT interim state at
920 ka (e.g., Berger and Jansen, 1994). Why did
100 kyr cyclicity lag the initial ice volume increase
by approximately 280 kyr (Mudelsee and Schulz,
1997)? The observed baseline shift in the SUSAS
records (Figure 5) precisely fills this time lag and
ShOll1d shed light on this question.
Trends
A prominent feature of the SUSAS stack and all
individual records summarized in Figure 5 is a base
line shift towards 40% higher average susceptibili
ties during the MPT interim state (920 to 640 ka).
Confined by transitional intervals lasting about 80
and 140 kyr at its onset and termination, this period
exhibits the most distinct dissimilarities between
individual core logs. The preceding and following
climate states show similar mean susceptibilities
modulated by clearly developed and coherent 41
and 100 kyr cycles, respectively. An exception is
the 'premature' near-100 kyr cycle at around
1150 ka discussed above. The MPT susceptibility
shift is notably milTOl'ed in lowered average sedi
mentation rates (Figure 5, bottom).
62 Chapter 4
~~yr S-;-t~ [500-' 0 kaJlI
Interim State (920 • 640!ffiJJ [ 41 kyr State (> 1000 ka)-Ii
_~~_~_~ __J
- 40"W 30"W 20"W 10"W 0' 10"E 40"W 30"W 20"W 10'W 0" 10"E 40"W 30"W 20"W 10"W 0" WECf) 800 800 800<D 2821- 1 3814-6 3813_3 3814-66 500- 500 • I. 500 2821-1 • 500-,-
3813-3 2~0-21309-2 • • 1309-2
.~ 300 3'12-1 300 300 • • • 3003812-'12820-2 3813-3
:n 200 • 200 1729-3 200 200"ß. •(lJ
100 100 • 100u(fJ • 1211-3::J 1211-3Cf) 50 50 50u o -~ 30 wO', J30 • 30c 20
• 20_ Enhanced Susceptibility 1034-3 - 20OJ 10 1034-3 10 10 10(\l Constant W·E Gradient:2:
0 -, 0 0
'L' 40"W 30"W 20"W 10"W 0" 10"E 40"W 30"W 20"W 10"W 0" 10"E 40"W 30"W 20"W 10"W 0' 10"E>. I I I I ._L- I I I.Y 3_0 3814-8 ~30 3814-8 3.0
N
• 3813-32821-1 • •E 20 2821-1 • • • • 3813-3
~ • 2.01309-2 1309-2
C •C 3801-8 2820-2 •0 10 • 1_0~
3812-1 1211-3 3812-1
1729-3 • 1729-3::J •E •0.5 0 0_5::J 1035-4U 0U 1311-1« 0_3 0_3-
21034-3 Constant Magnetite 1034-302L • 0.2 •~ Accumulation
c Constant W·E GradientOJ(\l 01 . 01 0.1
:2:-~l -,
40"W 30"W 20'W 10"W 0" 10"E 40'W 30"W 20"W 10"W 0" 10"E 40"W 30"W 20"W 10"W 0" 10"E1_2 1.2 I I I 1_2 1.21_1 3801-8 1035-4 1.1 Reduced Sedimentation 1.1 1034-3 1_1
'L' •>. 1_0 1211-3 10Collapsed W·E Gradient
10 • 1_0~ 0_9 • 0.9 0_9E 0_8
1034-308 08Q 2820-2
(lJ 0-7 0.7 0.7 •ro 1034-3• •0::: 0.6 2821-1 3814-8 • 0_6 3814-8 3813-3 0.6 1309-2c • • • 2821-1 •0 3812-1
2821-1 • •~
0_5 • - 0_5 -0_5 •2820-2 • 1309-2 • 0 3812-1
C 1729-3 1311-1(lJ OA 0 OA 3812-1 • OA •E '1311-1 • • 1211-3 3814-8
iJ 2820-2 0(lJ 1311-1
Cf) 0_3 0.3 0.3 0.340 0 W 30 0 W 20 0 W 10 0 W 0° 10 0 E 40 0 W 30 0 W 20 0 W 10 0 W 0° 100 E 40 0 W 30 0 W 20 0 W 100 W 0° 10 0 E
Longitude
Figure 8. Combined spatial and temporal trend analyses ofmagnetic susceptibility, magnetite accumulation (caIculated bysplitting susceptibility into a diamagnetic background and ferrimagnetic, grain-size independent magnetite signal) and sedimentation rate_ Major asymmetries and shifts are sunu11arized by white regression Iines, dark-gray shading delimits mean datarange, light-gray shading western and eastern South Atlantic care sets. Three cores were excIuded from the analysis: coresGeoB 1035-4 and 3801-6, as they do not reach beyond the 530 ka event, and core GeoB 1311-1 recovered from 290 1 m waterdepth cIearly above the LCDW/NADW transition zone and therefore showing a somewhat different evolution.
The cause for the overall rise in magnetic suscep
tibility during the MPT interim state is revealed by
a temporal analysis of west-east trends. Plotted
against longitude, core mean susceptibilities far the
postulated three climate states (Figure 8 top) reflect
the well-known west-east asymmetry ofterrigenous
particle flux and accumulation in the South Atlantic
(e.g., Balsam and McCoy, 1987; Lisitzin, 1996).
While the gradients are almost identical throughout
Pleistocene, a shift to higher susceptibilities is evi
dent during the MPT interim state. This relative in
crease ofmagnetic mineral concentration cannot be
explained in terms oftemporal changes in magnetite
accumulation (Figure 8 center), as its west-east de
cline from 2.5 to 0.3 g/m2kyr remains fairly constant
over time. A time-varying dilution ofthe terrigenous
Cycles, trends and events of Pleistocene sedimentation 63
fraction by non-magnetic carbonate must therefore
be responsible for the observed shift.
Mean sedimentation rates (Figure 8 bottom)
deduced from our age models are mainly controlled
by carbonate accumulation. Both for the 41 and
100 kyr climate states, sedimentation rates vary
from about 0.5 cm/kyr in the western to around
0.9 cm/kyr in the eastern South Atlantic. During
MPT interim state sedimentation rates on either side
ofthe mid-Atlantic Ridge are restricted to between
0.4 and 0.6 cm/kyr. The decline primarily aifects
the eastern part and brings the west-east asymmetry
to collapse.
In the working area cyclic variations ofthe sedi
ment CaC03 content at Milankovitch frequencies
are mainly due to orbitally driven lysocline shifts
resulting from an interplay of NADW and more
corrosive LCDW (Bickert and Wefer, 1996). In the
Angola Basin (cores GeoB 1034-3, 1035-4,1729-3
and 3801-6) the glacial-interglacial contrast is
particularly manifest as the surrounding bathymetric
highs restrict the access of LCDW during inter
glacials. The same reasoning should apply for long
tern1 changes in carbonate accumulation. We there
fore assume that the influence of southern-source
deep water was greatly enhanced during the MPT
interim state. This conclusion is supported by a
simultaneous decrease of glacial and interglacial
kaolinite/chlorite ratios in core GeoB 2821-1
(Gingele et a1., 1999), interpreted to be induced by
chlorite-enriched deep southern waters. Conse
quently, NADW should have been reduced during
that interval. Indeed, 013C records from Atlantic and
Pacific ODP sites indicate significantly weaker
NADW between 900 and 400 ka (Raymo et a1. ,
1997). In view of this global evidence we suggest,
that the MPT should not be regarded as a gradual
transition frOl11 a '41 kyr world' to a '100 kyr
world', but rather as a third, contrasting climate
state.
Terminal MPT event
In several of the SUSAS COl'es unusual sediment
facies occur coincidentally at around 530 ka, dose
to the end ofthe terminal MPT transition. Granu!o
metric analyses of cores GeoR 2820-2 and 2821-1
show a sharp grain-size shift with an almost total
loss ofthe> 63 flm fraction (Breitzke, 1(97), possi
bly related to the mid-Brunhes dissolution cyc1e
(e.g, Adelseck, 1977). In core GeoB 2820-2 this
episode corresponds to a short hiatus. The continu
ous late Pleistocene record of core GeoB 1035-4 is
interrupted by a thick turbidite dating to the same
age. Core GeoB 3812-1 displays a sharp color
change at a delicately laminated horizon (Figure 9).
Most impressive are thick intercalated laminated
diatom ooze layers in cores GeoB 3801-6 (124 cm)
and 3813-3 (38 cm). As discussed above, the
laminations must have been deposited in a very Sh011
time and should reflect an extremely short-term
climate variability. Very high sedimentation rates
in the overlying carbonate sediments may have
contributed to preserve the siliceous sections in both
cores. The almost monospecific layers consist of
the giant diatom Ethmodiscus rex (Rattray) Wise
man and Hendey (C.B. Lange, pers. comm.), the
largest solitary (not chain-fol1ning) diatom known,
reaching diameters of2-3 mm (Wiseman and Hen
dey, 1953; Round et a1., 1990). They are entirely
uncommon at these latitudes as Ethmodiscus rex is
found in plankton primarily in equatorial regions
in a range of temperature frOl11 19° to 29.5°C
(Lisitzin, 1996). Recently an unusual diatom ooze
layer was detected in core GeoB 5112-4 recovered
on R.V. METEOR cruise M41/3 at 23°49,5'S
16°15,5' W from a water depth of3842 m (Pätzold
et a1. , 1999). C01Telating the magnetic susceptibility
record of this core to the SUSAS stack proves that
the diatoms were deposited at the same time as in
the two SUSAS cores (Frederichs et a1. , 1999b).
Thick ooze deposits of this diatom were mainly
64 Chapter 4
Age [kar528 530 532 534 536 538 540 542 544 546 548 550 552 554 556
306 307 308 309 310 311 312 313 314 315 316 317 318 319Core Position [ern]
Figure 9. Terminal MPT event documented in a sharp colourtransition at unbioturbated finely laminated layers in coreGeoB 3812-1.
reported from equatorial regions ofthe world oeeans
(e.g., Gardner and Burekle, 1975; Mikkelsen, 1977;
StabeIl, 1986; Lisitzin, 1996). Even there, their
oeeurrenee is puzzling, beeause EthnlOdiscus rex
very rarely oceurs in plankton sampIes (not more
than 0.5 eells/m3; Lisitzin (1996)). However, enig
matie deep populations have been observed in the
Paeific Oeean (Villareal, 1993). Several hypothesis
have been proposed to clarify the paradoxieal
'Ethlnodiscus rex problem' (Gardner and Burekle,
1975), including differential dissolution, foeusing
by bottom eUlTents and intense productivity blooms.
Results from the Joint Global Oeean Flux Study
(JGOFS), emphasize the potentially important role
of oeeanie frontal zones for the rapid aeeumulation
of diatom biomass (Yoder et a1., 1994). They indi
eate dramatic biological responses to circulation and
mixing processes assoeiated with open-ocean fron
tal systems separating cold from wann waters and
gave rise to the idea that thick diatom layers may
be deposited in such environments.
Pieking up this suggestion and taking into account
the OCCUlTence of the tel111inal event at the end of
the MPT interim state of strongly reduced NADW
flux we believe it to result from elemental changes
in the global ocean circulation system. Possibly an
oceanic frontal zone shifted into the subtropical
South Atlantic for a short time and thus initiated
the massive diatom growth. The diversified terminal
MPT events at around 530 ka documented in South
Atlantie Oeean sediments might weIl be related to
other paleoc1imatic 'puzzles'. Recently, Rossignol
Strick et a1. (1998) found an anomalous sapropel
layer in the MeditelTanean Sea dated 528-525 ka
and interpreted it to result from a 'massive odd 1110n
soon'. Previously Bassinot et a1. (1 994b) reportecl
extremely high and low 0180 values during isotope
stages 13.2 and 13.3, respectively, in a giant piston
core from the equatoriallndian Ocean. Analogously
to the sapropel layer the unusually depleted value
at stage 13.3 ean be interpreted as a regional effeet
resulting from low surface-water salinity due to
heavy monsoonal fluvial discharge (Rossignol
Strick et a1. , 1998). But insolation eonditions would
not predict heavy monsoon rainfall over Afriea and
Asia during that time. However, also a terrestrial
climate proxy indieates an extraordinary warm
climate. Chinese loess sequences and interbedded
paleosols are known to reeord paleoc1imate, as loess
was deposited in a cold, dry climate, while soils
formed during wet and mild episodes. Paleosol S5,
which can be related to oxygen isotope stages 15
13, is by far the thickest and most weathered 0 f the
whole sequence (An and Wei, 1980). At its top, near
stage 13, the highest magnetic susceptibilities of
the past 2500 kyr (Kukla et a1., 1988; Heller et a1.,
1991; B10emenda1 et a1. , 1995) imp1y an extremely
warm and humid climate in Asia, although oxygen
isotope records identify this stage as one ofthe least
expressed interglacials during the late Pleistocene.
Similarly as the MPT interim state, the terminal
event can not be explained by astronomieal forcing
nor is it documented in global iee-volume records.
Conclusions
The rock magnetic view at sediment sequences from
the oligotrophie subtropical South Atlantic uneovers
several new and decisive aspects of the mid-
Cyc1es, trends and events of P1eistocene sedimentation 65
P1eistocene climate transition (MPT). The ana1yses
are based on high-resolution magnetic age models
which were built despite very 10w sedimentation
rates for aseries oftwe1ve sediment cores by orbital
tuning of their magnetic susceptibility records.
Located on the submarine ridges in water depths
affected alternate1y by NADW and LCDW these
sediment sequences have recorded ocean history
over the past 1500 kyr in a region very important in
the framework ofthe global thermohaline circu1a
tion. The good overall cOlTe1ation of all records let
us to compute the SUSAS stack.
Cyclic patterns, which partly were used for the
magnetic cyc10stratigraphy, essentially reflect or
bitally forced changes in deep water chemistry. In
the course of the MPT they document the well
known change from variations with aperiod of
41 kyr to the 1ate P1eistocene 100 kyr ice age
rhythm. Compared to the driving functions in an
evo1utionm'y spectra1 analysis an a1ternating ex
change of spectra1 energy between the 41 and
100 kyr bands is implied. In accordance with the
suggestions ofRuddiman et al. (1989) and Joyce et
al. (1990) residues between the SUSAS stack and
an adapted ETP record characterize the c1imate
system as less sensitive to obliquity changes during
the 1ate P1eistocene. According to our analysis the
shift towards strong1y reduced response to obliquity
forcing takes place at about 650 ka and is synchro
nous with the onset of 100 kyr cyc1icity in our data
and as weIl in global ice volume (e.g., Ruddiman et
al., 1989; Mude1see and Schulz, 1997). This result
substantiates thresho1d models, which mimic the
occurrence ofnear-1 00 kyr cyc1es in the late Quater
nary by introducing a decreasing atmospheric pC02
level (e.g., Raymo, 1997; Paillard, 1998). The
residual analysis also documents reduced response
in the 41 kyr band coinciding with the premature
occurrence ofa near-1 00 kyr cyc1e at about 1150 ka.
Being previously reported from severa18 180 records
(e.g., Mude1see and Stattegger, 1997), this incident
involves global ice volume.
In addition to features which mimic oxygen iso
topes, the rock magnetic proxy unveils a MPT in
terim state of reduced carbonate deposition obvi
ous1y not directly linked to changes in global icc
vo1ume. Bounded by transitions lasting fi'om 1000
to 920 ka and from 640 to 500 ka this interval pre
cise1y fills the time lag between the first occurrence
of 1arger glacia1 ice shie1ds and the onset of near
100 kyr cyc1icity (e.g., Mudelsee and Schulz, 1997).
Our trend ana1yses implies that the outstanding fea
ture ofthis episode is enhanced influence of south
ern source deep waters. Support for this hypothesis
comes from ana1yses of 813C records imp1ying a
consistent reduced influence ofNADW during this
time interval (Raymo et al., 1997).
At the end of the MPT interim state a terminal
event is documented in severa1 SUSAS cores. It is
believed to be due to strong changes in the ocean
circulation system linked to the end of the MPT
interim state and the beginning of the late Pleisto
cene 100 kyr state. A global character of this event
is suggested as it occurs coincidentally with other
unusual c1imate excursions which point towards
extremely warm and humid c1imates in Asia and
Africa (e.g., Kukla et al. , 1988; Bassinot et al. ,
1994b; Rossignol-Strick et al., 1998).
Acknowledgements
We thank T. Bickert for kindly providing CaC03data
ofGeoB 1211-1 and the unpublished 813C record of
ODP Site 806. This study was funded by the Deutsche
Forschungsgemeinschaft (Sonderforschungsbereich
261 at Bremen University, contributionNo. XXX). F.S.
was supported by the Deutsche Forschungsgemein
schaft in the framework ofGraduiertenkolleg 221.
66 Chapter 4
Referellces
Adelseck, e.G. Jr., Reeent ,md late Pleistoeene sediments from
the castern equatorial Paeifie Oeean: sedimentation and
dissolution, Ph.D. thesis, Univ. of Calif., San Diego,
192 pp, 1977.
An, Z.S. and L.Y. Wei, The fifth layer paleosol in the Lishi
Loess and their paleoelimatie signifieanee, Acta Pcdolo
gica Siuim, 17. 1-10, 1980.
Balsam, W.L. ami F. W. I\1eCoy, Jl'., Atlantie sediments:
glaeial/interglaeial eomparisons, Palcoceauogmph.1', 2.
531-542,1987.
Bassinot, F.e., L. Beaufort L, E. Vineent, L.D. Labeyrie. F.
Rostek, P.J. Miiller, X. Quidelleur and Y. Laneelot, Coarse
fraetion fluetuations in pelagie carbonate sediments [rom
the tropieal Indian Oeean: a I 500-kyr reeord ofearbonate
dissolution, Palcoccauogm/Jhy, 9, 579-600, 1994a.
Bassinot, F.e., L.D. Labeyrie, E. Vineent, X. Quiclelleur, N.J.
Shaekleton ami Y. Laneelot. The astronomieal theory of
elimate ancl the age of the Brunhes-Matuyama magnetie
reversa!, Earth Plane!. Sci. Lct!., 126, 91-108, 1994b.
Berger, A. ami M.F. Loutre, Insolation va lues for the elimate
ofthe last 10 million years, QIIa! Sei Rc\', 10,297-317,
1991.
Berger, A. ancl M.F. Loutre, Astronomieal solutions for
paleoclimate stuclies over the last 3 million years, Earth
Plauct. Sei. Lett., 111,369-382,1992.
Berger, W.H. and G. Weftr, Klimageschichte aus Tiefsee
sedimenten, NatllrJvissensc/wjtcn, 79, 541-550, 1992.
Berger, W.H. and E. Jansen, Mid-Pleistoeene elimate shift
the Nansen eonneetion, in The Polar Oceans aud their
Rolc in Shapiug thc Glohal Envil"Onlllcut, Geophys.
Monogr., 84, 295-311, AGU, Washington, 1994.
Berger, W.H., M.K. Yasuda, T. Biekert, G. Weter and T.
Takayama, Quaternary timc seale for the Ontong Java
plateau: Milankoviteh tcmplate for Ocean Drilling Pro
gram Site 806, Gcology, 22, 463-467, 1994.
Berger, W.H. ancl G. Weier, Expeditions into the past: pa
leoceanographic studies in the South Atlantic, in Thc South
Atlantic: Present a/1(l Past Circulation edited by G. Wefer,
W.H. Berger, G. Siedler ancl D.J. Webb, Springer Verlag,
Berlin, pp 363-410, 1996.
Berger, W.H., T. Bickert, M.K. Yasuda ancl G. Wefer,
Reeonstruetion of atmospherie CO 2 from iee-eore data and
the deep-sea reeord ofOntong Java plateau: the Milanko
viteh ehron, Gco!. Ruudschau, 85,466-495, 1996.
Biekert, '1'., Rekonstruktion der spätquartärel1 Bodenwasser
zirkulation im östlichen Südatlal1tik über stabile Isotope
bentl1iseher Foraminiferen, Ph.D. thesis, Bel'. FB Geow.
Univ. Bremcn, 27, 205 pp, 1992.
Bickert, T. and G. Wefer, Late Quaternat·y deep water
cireulatiol1 in the South Atlantic: reeol1struction from
carbonate dissolution and bel1thic stable isotopes, in The
South Atlantie: Presentaud Past Cireu/ariou edited by G.
Wefer, W.H. Berger, G. Siedler and D.J. Webb, Springer
Verlag, Berlin, pp 599-620, 1996.
Birchfield, G.E. and M. Ghil. Climate evolution in the Plioeene
and Pleistocene from marine-sediment reeords ami sinlllla
tions: interna I variability versus orbItal foreing, J G('()
pln's. Res, 08,10385-10399,1993.
Bleil, U. and eruise partieipants, Report amI prcliminary results
of METEOR eruise 29/2, MonteVIdeo Rio de J,lnciro,
15.7.1994· 8.8. I 994. Ba. FIJ Gc01\' Um1' Bremcn.59.
153 pp, 1994.
Bloemencla!, .1., 13. Lamb ancl .I. King. Paleoenvironmcnt:i1
implications 01' rock-magnetic properties 01' Latc QLwter
nary sediment COlTS from the eastern equatorial Atlantic.
Pal('()eclluography, 3, 61-87,1988.
Bloemendal, J. and P. deMenoeal, Evidence for a change in
the periodicity oftropieal climate cycles at 2.4 Myr from
whole-eore magnetie suseeptibility measuremel1ts, Nature
342, 897-900, 1989.
Bloemenclal, .1., .l.W. King, L. Tauxe ami J.-P. Valet, Rock
magnetic stratigraphy 01' leg 108 sites 658, 659, MI, amI
665, eastern tropieal Atlantic, Pme" ODP, Sei. Reslllrs,
108,415-428, 1989.
Bloemendal, J., X.M. Lill and T.e. Rolph, CorrelatioJ1 of the
magnetic sllseeptibility stratigraphy 01' Chinese loess and
the marinc oxygen isotope record: chronological ami
palaeoclimatic implications, Earth Planct. Sci. Lct!, 131,
371-380,1995.
Breitzke, M. Elastische Wellenausbreitung in marinen Sedi
menten - neue Entwicklungen der Ultraschall Sediment
physik lind Sedimentechographie. Bel'. FB Geo\\'. Uni1'.
Bremen, 104,298 pp, 1997.
Chi, .I. ancl J. Mienert, Linking physieal property records 01'
Quaternary sediments to Heinrich evenls, /l1ar. Geolog1'.
131,57-73,1996.
Clark, P.U. and D. Pollard, Origin of the middle Pleistocene
transition by ice sheet erosion ofregolith, Paleoceanogra
phy, 13, 1-9, 1998.
DeBlonde, G. ancl W.R. Peltier, A one-dimensional model 01'
continental ice volume fluctuations through the Pleisto
cene: implications for the origin of the Mid-Pleistoeene
climate transition, J Clim., 4, 3 I 8-344, 199 I.
Frederiehs, 'rh., U. Bleil, K. Däumler, T. von Dobeneek ancl
A. Schmidt, The magnetic vievv on the marine paleoen
vironment: parameters, techniques and potentials of rock
magnetic studies as a key to plaeoclimatie and paleoceno
graphie ehanges, in Use 01' Proxies in Paleoecunogra
phy: Examplesji'om the SOllth Atlantie edited by Fischer,
G. and G. Wefer, Springer-Verlag Berlin Heidelberg, in
press, 1999a.
Frederichs, Th., K. Fabian and J. Funk, Physical properties
studies, in Report und prelimin([/)' re.l'lllt.l' 01' METEOR
cmise M4/ /3, Vitöria - Salvador de Buhiu, /8.4. /998 -
Cyclcs, trends and events of Pleistocene sedimentation 67
! 5.5.! 998, Ber. F/3 Gco\\'. Ulli,'. Bremell, in press, 1999b.
Gardner, ,LV. and LH. Burekle, tIpper Pleistoeene Etll
lI/odiscus rex oozes ti'om the castern equatorial Atlantie,
Micropo!eolltol. 21, 236-242, 1975.
GingeIe, F.X., F. Schmieder, T von Dobeneek, R. Petsehiek
and C. Rühlemann, Terrigenous t1ux in the Rio Grande
Rise area during the last 1500 ka: evidenee of deepwater
adveetion or rapid response to eontinental rainfall
patterns?, Paleoceallogmpln', in press, 1999.
Grlitzner, l, F.C Bassinot and J. Mienert, High-resolution
eompressional-wave veloeity measurements in Pleistoeene
sediments oethe Ceara Rise (western equatorial Atlantie):
implieations for orbital driven sedimentary eycles, Pmc.
GDP, Sci Rcsults, 154, 135-149, 1997.
Hays, lD., l Il11brie and NJ. Shaekleton, Variations in Earth's
orbit: pace-makel' ofthe iee ages, Sciellce, 194, 1121-1132,
1976.
HaITis, S.E., AC. Mix and T King, Biogenie and terrigenous
sedimentation at Ceara Rise, western tropieal Atlantie,
supports P1ioeene-Pleistoeene deep-water linkage between
hemispheres. Pl'Oc. GDP. Sci. Results, 154,331-345, ]997.
Heller, F., X Liu, T. Liu and T. Xu, Magnetie suseeptibility
01' loess in China, Earlll Plalle!. Sei Let!, 1OJ, 301-3] 0,
1991.
Imbrie, l, Mcasuring the gain ofthe climate system 's response
to Milankoviteh foreing in the preeession and obliquity
bands, in LOllg-tel'lll clill/atie variations -- data amlmode/
/ing, Springer, Berlin, pp 403-410, 1992.
Imbrie, l, lD. Hays, D.G. Martinson, A. McIntyre, A.C
Mix, The orbital theory of Pleistoeene elimate: support
üom a revised chronology of the marine 0180 reeord, in
Milankovitcll (md Clill/ate Part!, edited by A. Berger, l
Imbrie, l Hays, G. Kukla, B. Saltzman, Reidel, Dordreeht,
pp 269-305, 1984.
Imbrie, l, A. Berger, E.A. Boyle, S.c. Clemens, A Duffy,
W.R. Howard, G. Kukla, l Kutzbaeh, D.G. Martinson,
A. McIntyre, A.C Mix, B. Molfino, II Morley, LC
Peterson, N.G. Pisias, W.L. Prell, M.E. Raymo, N.J.
Shaekleton and lR. Toggweiler, On the strueture and
origin of major glaeiation eycles 1. Linear responses to
Milankoviteh foreing, Paleoceanograplly, 7, 701-738,
1992.
Imbrie, l, ABerger, E.A. Boyle, S.C Clemens, A Duffy,
W.R. Howard, G, Kukla, l Kutzbaeh, D.G. Martinson,
A. Mclntyre, AC Mix, B. Molfino, J.J. Morley, L.C
Peterson, N.G. Pisias, W.L. Prell, M.E. Raymo, N.J.
Shaekleton and lR. Toggweiler, On the strueture and
origin of major glaeiation eycles 2. The 100,000-year
eycle, Paleoceanogmplzy, 8, 699-735, 1993.
Johnson, D.A., M. Ledbetter and L. Burekle, Vema Channel
paleo-oeeanography: Pleistoeene dissolution eycles and
episodie bottom water t1ow, Mal'. Geology, 23, 1-33, 1977.
Joyee, lE., L.R.C Tja1sma, lM. Prutzman, High resolution
planktie stable isotope reeord and speetral analysis for the
last 5.35 m.y.: Oeean Drilling Progral11 Site 625,Ilortheast
gulf ofMexieo, Pa!eoccallogl'llplly, 5, 507-529, 19()0.
Kllkla, G., F. Heller, X.M. Liu, T.C Xu, T.S. Liu alld Z.S.
An, Pleistoeene elimates in China da ted by magnetie
suseeptibility, Geolog)', 16, 81] -814, 1988.
Lisitzin, A.P., Oeeanie sedimentation: Jithology and gco
ehemistry, AGU, Washington D.C, 400 pp, 19%.
Maaseh, K.A., Statistieal deteetion 01' the micl-P!elstoeene
transition, C/im. D1'II., 2, 133-143, 1988.
Martinson, D.G., N.G. Pisias, J.D. Hays, J. Imbric, r.c.Moore, N..!. Shaekleton, Age dating and the orbital theory
of the iee ages: development 01' a high-resolution 0 (0
300,000-year ehronostratigraphy, QUa!. Rcs .. 27, 1-29,
1987.
Mead, G.A. and L Tallxe, Oligoeene paleoeeanography of
the South Atlantie: paleoelimatie implieations ofsediment
aeeumulation rates and magnetie suseeptibility measurc
ments, PalCOCC(//lOgmpllv, I, 273-284, 1986.
Mikkelsen, N., On the origin of Ethrnodiscus OOl:e, Mal'
Ahcl'Of!alcolllol., 2, 35-46, 1977.
Mudelsee, M. and M. Sehulz, The Mid-P1eistoeene climate
transition: onset of 100 ka eyele lags iee volume budd-up
by 280 ka, Eartlz Planet. Sci. Lett, 151. 117··123, 1997.
MudeIsce, M. and K. Stattegger, Exploring the strueture 01'
the mid-Pleistoeene revolution with advaneedmethods of
time-series analysis, Geol. Rundschau, 86,499-511,1997.
Muller, R.A. and G ..!. MaeDonald, Simultaneous presenee of
orbital inelination and eeeentrieity in proxy elimate reeords
from Oeean Drilling Program Site 806, Geolog)', 25. 3-6,
1997.
Mwenifumbo, C.l, .!.P. Blangy, Shor(-(erm speetral analysis
of downhole logging measurements from Site 704, Proc.
GDP, Sei. Results, 114, 577-585, 1991.
Paillard, D., The timing of Pleistoecne glaeiations from a
simple multiple-state elimate model, Nature, 391, 1998.
Park, l, S.L. D'Hondt, l W. King and C. Gibson, Late Creta
eeous preeessional eycles in double time: a warm-Earth
Milankoviteh response, Science, 261, 1431-1434, 1993.
Park, .!. anel K,A. Maaseh, Plio-Pleistoeene time evolution of
the 1OO-kyr eycle in marine paleoclimate reeords, J. Geo
plzys. Res., 98, 447-461,1993.
Pätzold, l und Fahrtteilnehmer, Bericht und erste Ergebnisse
über die METEOR-Fahrt MI5/2, Rio de Janeiro - Vitoria,
18.1. -7.2.1991, Bel'. PB Geow. Univ. Bremen, 17,46 pp,
1993.
Pätzold, land eruise partieipants, Report and preliminary
resll1ts of METEOR eruise M41/3, Vit6ria - Sa1vador de
Bahia, 18.4.1998 - 15.5.1998, Bel'. PB Geow. Univ.
Bremen, in press, 1999.
Pestiaux, P. and A. Berger, An optimal approach to the speetral
eharaeteristies of deep-sea elimatie reeords, in Milanko
vitell and Climate Part! edited by A. Berger, l Imbrie, l
68 Chapter 4
Hays, G. Kukla ancl B. Saltzman, Reidel, Dordreeht, pp
417-445,1984.
Petersen, N., T von Dobeneek and H. Vali, Fossil baeterial
magnetite in deep-sea sediments from the South Atlantie
Oeean, Nature, 320, 611-615, 1986.
Pisias, N.G. and T.e. Moore, The evolution of Pleistocene
climate: a time series approach, Earth. Planet. Sei. Lett.,
52,450-458, 1981.
Prell, W.L., Oxygen and earbon isotope stratigraphy for the
Quaternary of Hole 502B: evidence tor two modes of
isotopie variability, Initial Rep., DSDP, 68,455-464, 1982.
Raymo, M.E., The timing of major climate transitions, Pa
leoeeanography, 12,577-585,1997.
Raymo, M.E., Glaeial puzzles, Science, 281, 1467-1468, 1998.
Raymo, M.E., W.F. Ruddiman, J. Backman, B.M. Clement
and D.G. Martinson, Late Pliocene variation in northern
hemisphere ice sheets and North Atlantic deep water
cireulation, Paleoeeanography, 4, 413-446, 1989.
Raymo, M.E., D.W. Oppo and W. Curry, The mid-Pleistocene
elimate transition: a deep-sea carbon isotopie perspeetive,
Paleoceanography, 12, 546-559, 1997.
Robinson, S.G., Applications for whole-core magnetie suscep
tibility measurements of deep-sea sediments, Proe. GDP,
Sei. Results, 115,737-771,1990.
Robinson, S.G., M.A. Maslin andl.N. McCave, Magnetic sus
eeptibility variations in Upper Pleistocene deep-sea sedi
ments ofthe NE Atlantic: implications for ice rafting ancl
paleoeireulation at the last glaeial maximum, Paleoeean
ography, 10, 221-250, 1995.
Rossignol-Strick, M., M. Paterne, F.e. Bassinot, K.-e. Emeis
and G.I. DeLange, An unusual mid-Pleistocene mon8Oon
period over Afriea and Asia, Nature, 392, 269-272,1998.
Round, F.E., R.M. Crawtord and D.G. Mann, The diatoms:
biology and morphology ofthe genera, Cambridge Univ.
Press, Cambridge, 747 pp, 1990.
Ruddiman, W.F., M.E. Raymo, D.G. Martinson, B.M.
Clement and I. Backman, Pleistocene evolution: northern
hemisphere ice sheets and North Atlantic ocean, Pa
leoeeanography, 4, 353-412, 1989.
Saltzman, B. and M.Y. Verbitsky, Multiple instabilities and
modes of glacial rhythmicity in the Plio-Pleistocene: a
general theory of late Cenozoic climatic change, Clim.
Dyn., 9, 1-15, 1993.
Schmidt, A., T. von Dobeneck and U. Bleil, Magnetic
characterization of Holocene sedimentation in the South
Atlantic, submitted.
Schulz, H.D. und Fahrtteilnehmer, Bericht und erste
Ergebnisse über die METEOR-Fahrt M20/2, Abidjan
Dakar, 27 .12.1991 - 3.2.1992, Ber. FB Geow Univ.
Bremen, 25, 173 pp, 1992.
Shaekleton, N.J., A. Berger, W.R. Peltier, An alternative
astronomical calibration ofthe lower Pleistocene timescale
based on ODP Site 677, Trans. Royal Soc. Edinburgh:
Earth Sei, 81, 251-261,1990.
StabeIl, B., Variations ofdiatom flux in the eastern equatorial
Atlantic during the last 400,000 years (METEOR cores
13519 and 13521), Mar. Geol., 72, 305-323,1986.
Thiessen, W., Magnetische Eigenschaften des östlichen Süd
atlantiks und ihre paläozeanographisehe Relevanz, Ph.D.
thesis, Ber. FBGeow. Univ. Bremen, 41, 170 pp, ]993.
Thompson, R. and F. Oldfield, Environl11ental l11agnetisl11,
Allen and Unwin, London, 227 pp, 1986.
'riedel11ann, R., M. Sarnthein and N.l Shackleton, Astronomie
til11escale for the Pliocene Atlantie 8 130 and dust tlux
records ofOcean Drilling Program site 659, Paleoccano
graph)!, 9, 619-638,1994.
Vali, H., O. Förster, G. Al11arantidis and N. Petersen, Magncto
tactic bacteria and their magnetofossils in sediments, Fllrth
Planet. Sei. Lett, 86, 389-400,1987.
Villareal, TA., Abundance ofthe giant diatol11 Ethmodisells
in the Southern Atlantic Oeean and the eentral Pacific gyrc,
Diatom Res, 8, 171-177, 1993.
Volat, lL., L. Pastourct and e. Grazzini, Dissolution and car
bonate fluctuations in Pleistocene deep sea cores: a re
view, Mar. Geology, 34, 1-28, 1980.
von Dobeneck, T and F. Schmiede!', Using rock magnetic
proxy records for orbital tuning and extended time series
analyses into the super- and sub-Milankovitch bands, in
Uye ofProxics in Paleoecanography: Examplesfi'o/ll the
SOllth Atlantie edited by Fischer, G. and G. Wefer, Sprin
ger-Verlag Berlin Heidelberg, in press, 1999.
Wefer, G. und Fahrtteilnehmer, Bericht über die METEOR
Fahrt M 6/6, Libreville - Las Palmas, 16.2.1988
24.3.1988, Ber. FB Geow. Univ. Bremen, 3, 97 pp, 1988.
Wefer, G. und Fahrtteilnehmer, Bericht über die METEOR
Fahrt M12/l Kapstadt - Funchal, 13.3.1990 - 14.4.1990,
Ber. FB Geow. Univ. Brelnen, 11,66 pp, 1990.
Wefer, G. and cmise participants, Report and prelil11inary
results of METEOR cmise M34/3, Walvisbay - Recife,
21.2 - 17.3.1996, Ber. FB Geow. Univ. Bremen, 59,
153 pp, 1996.
Wiseman, J.D.H. and N.!. Hendey, The significance and
diatom content of a deep-sea floor sampie from the
neighbourhood of the greatest oceanic depth, Deep-sea
Res, 1,47-59,1953.
Yiou, P., C. Genthon, M. Ghil, J. Jouzel, H. Le Treut, I.M.
Barnola, C. Lorius and Y.N. Korotkeviteh, High-frequency
paleovariability in climate and CO2 levels from Vostoek
ice eore reeords, 1. Geophys. Res., 96, 20365-20378,1991.
Yoder, lA., S.G. Ackleson, R.T. Barber, P. Flament and
W.M. Balch, A line in the sea, Nature, 371,689-692, 1994.
69
Terrigenous flux in the Rio Grande Rise area during the past 1500 ka:Evidence 01' deepwater advection 01' rapid response to continentalrainfall patterns? (Paleoceanograplzy, 14, 84-95, 1999)
Franz X. CJingele, I Frank Schmieder,2 Tilo von Dobeneck,2 Rainer Petschick,3
amI Carsten Rühlemann2
Abstract. Sud~1ce sediment sampies and three gravity COlTS from the eastern terrace of the Vema Channc1, fhe westernnank of the Rio Grande Rise, and the Brazilian continental slope were investigated tor physical properties, grain size, anclclay mineral composition. Discharge of the Rio Doce is responsible tor kaolinite enrichments on the slopc south of 2()"and at intermediate depths of the Rio Grande Rise. The long-distance advection of kaolinite with North Atlantic DeepWater from lower latitudes is of minor importance as evidenced by low kaolinite/chlorite ratios on the Mid-AtlanticRidge. Cyclic variations of kaolinite/chlorite ratios in all our cores, \Vith maxima in interglacials, are attributcd tn 100vand high-latitude forcing of paleoclimate on the Brazilian mainland and the related discharge of the Rio Doce. A longterm trend to\Vard more arid and "glacial" conditions from 1500 ka to present is superimposed on the glacial-intcrglacialcyclicity.
1. Introduction
The Rio Crande Risc!Vema Channel/Hunter Channel regionis a key area in the western South Atlantie to study spatial asweil as temporal variations in the history of deepwater masses.Southern source bottom water passes from the AI'gentine to theGrazil Gasin via the Vema and Hunter Channels. Above, southward flowing North Atlantie Deep Water (NADW) is recordedfrom 4000 to 2000 m water depth, overlain by Upper Cireumpolar Water (UCPW) and Antaretie Intermediate Water (AAIW)[PelerSOI/ al/(! SlrallIIlW, 1(90). The Rio Crande Rise (RCR),wh ich reaches water depths of 800 m, offcrs the opportunity tosampie sediments situated in different water masses and look fortracers of changes in thermohaline circulation.
Numerous studies have focussed in partiClilar on variations inthe propagation of Antarctic Gottom Water (AAGW) andNADW in the South Atlantic using various tracers from benthicforaminifera assemblages [lI/fackel/seil ef al., 1993; SchIIlied! alld
MackeIlseIl, 1997], Cd/Ca ratios in benthic foraminifera [Boy!e,
1988, 1994; Oppo al/d Rosellf!w!, 1994], carbon-13 eomposition[Omy ef a!., 1988; Dup!essy ef al., 1988; IlIackellsell cf al.,
1994]. diatoms [Jolles alld .Johl/soll, 1984], grain size [Masse cf
al., 1994]. and elay minerals [Biseaye, 1965, (,halll!ey, 1975,JOlles, 1984; Dieklllalll/ cf al., 1996).
On the basis of their strong latitudinal and reciprocal distribution patterns in surface sediments of the South Atlantie, kaolinite and chlorite were inferred to be useful tracers ofthe majordeepwater masses in the vicinity of the RCR. An advection of
I Baltic SeH Research Institute, Rostock-Warncmucnde, Germany.
2 University 01' Bremcn, Bremen, Germany.
.\ University 01' Frankfurt, Frankfurt, Germany.
Copyright 1999 by the Allleriean Ceophysieal Union.
Paper number I998PA9000 12.0883-8305/99/1 998PA900012$12.00
chlorite with southern source dcepwatcr (AAI3W) was found byIiiscaye [1965] and confirmed by later studies [Jolles. I()84;Pelschick cl al., I99CJ]. The OCCUITence 01' 'I kaolinite maximumat intermediate depths (above 4000 m) was attrlbuted to advection 01' this mineral by NAD\V [ClwlII!e)" 1975). The propagation of kaolinite with NADW is 'I well-documented feature inmany cores 1'rom the eastern South Atlantie [f)ie!w/{/1I11 cl 01.,1996].
;\ Iternatively, JOlles [1984] suggested an isopycnal transportmodel for the kaolinite enriehment on the RCR. On the basis 01'the assumption that the suspension load 01' the Nf\DW in thewestern South Atlantic is too small to sustain a subslantialenrichment (clearwater minimum), kaolinite input by the RioDoee, southward transport with the I3razil current, deposition onthe Sao Paulo (Santos) Plateau, and resuspension and isopycnalflow to the RCR was proposcd. These different interpretationshave implieations on the evaluation 01' temporal changes 01'kaolinite content in sediment cores, which record either oeeanicprocesses such as NADW fluctuatlons, sea level ehanges, orvariations in fluvial discharge.
We examined 51 surface sampies from two transeets toevaluate potential sources and reeent propagation 01' clayminerals (Figure I). One profile sampies the continental slopeoff the Rio Doce. The second transect runs from the slope acrossthe Santos Plateau, the western and castern flank 01' the RCR, tothe Mid-Atlantic Ridge (MAR).
The temporal variation in the supply 01' clay minerals c1uringthe past 200 kyr was investigated in sediment eores from theslope (CeoB 2110) and the castern terrace of the Vellla Channel(CeoB 2822). A complete reeord of clay mineral supplycovering the past 1500 kyr was found in core GeoB 2821. Thiscore is situated in a key position on the western flank of theRCR just above the present AABW/NADW boundary. Crossspeetral analysis on the kaolinite/chlorite ratio of this eore wascarried out to deterllline phase relationships to orbital cycles andeompare phase angles 01' related paleoeeanographic and paleoclimatie proxies.
70 Chapter 5
-45' _40' _35' _30" _25" _20' -15'
kaolinite/chlorite 0 < 1 c. 2-5 • > 10
ratio -ll. 1-2 ~ 5-10' .... _-
E
Argentine Basin
_20"
-25~
-30'
-35'
Souree GEBCO.
w
_45" _40"
Brazil Basin
-25" -20' _15'
-20'
-30"
MidoceanRidge
10 kaolinite/chlorite9 ratio8
_J _65
4
32
1
o
/
__ 8 _Santos Rio GrandePlateau Vema Rise
Channel HunterChannel
0
E2SL
"- 2<lJU
Qj 3ro;';
4
50"W 45"W 40"W 35"W 30"W 20"W 15-W
Figurc 1. Area 01' investigation. core locations, and kaol initeichloritc ratios 01' 5 I surface sampies in the investigation area. Depth contours afterGeneral Bathymetric Chart 01' the Oceans (GEBCO). Kaolinite/chlorite ratios are also projected on a hypothetical transect Ii'om the Santos Plateauacross the Rio Grande Risc (RGR) to the Hunter Channel and Micl-Atlantic Riclgc (MAR). Extrcmd<aolinite/chlorite values (>10) from sites clircctlyotT thc Rio Doce are cxcluclecl from the projection.
2. Material and Methods
Sediment surface sampies were recovered with giant boxcorer and multicorer during Meteor cruises M 15/2 [Pätzold etal., 1993J, M23/2 [Bleil et al., 1993J, M29/2 [Blei! et al., 1994J,and M34/3 [Iflerer et al., 1996]. Sample treatment and claymineral analyses are given in detail by Petschick et al. [1996].Sediment cores (Table 1) were retrieved with a gravity corer oncruises M23/2 (GeoB 2110-3/4) and M29/2 (GeoB 2821-1 andGeoB 2822-2). Core GeoB 21 10 from 3008 m water depth onthe continental slope above the Santos Plateau (Figure I) con-
sists 01' grey hemipelagic clays with varying amounts 01' carbonate. Core GeoB 2822 from the eastern terrace 01' the VemaChannel (4267 m water depth) comprises terrigenous muds withsmall amounts 01' calcareous nannofossils and foraminifera.Increasing amounts 01' carbonate characterize core GeoB 2821from the western flank 01' the RGR (3927 m water depth).
The cores were sampled at 5 cm intervals. Samples weretreated with 10% acetic acid to remove carbonate, sieved over a63 ~lm mesh, and split into >2 ~lm and <2 pm fractions by settling technique. The weight percentage 01' the sand fraction (>63pm) was negligiblc. Weight percentages 01' the silt (2-63 ~lm)
and clay fraction «2 ~lm) were used to compute a silt/clay ratio.
Terrigcnous flux in tbc Rio Grande Risc arca during tbc past 1500 ka 71
Table I. Location of Scdimcnt Cores
Waler Lalilllde, CrlliseCore Deplh. III Longillldc Rcport Agc Model
Geo13 3008 28°389'8. Bleil cl al. Bleil Cl al. [1993J and2110-3/4 45°312'W [1993J C. Rlihlclllann,
(lIllpliblishcd dala, 1(93)GeoB 3927 30°271 '8, Bleil cl a!. F. SchIllieder,2821-1 38°481 'W [1994J (lIllplihlished dala, 19(4)GcoB 42()7 300 14.3 's, Bleil cl a!. Bleil cl a!. r199412822-2 3,)°08.5 'W [1994J
The clay fraetion «2 pm) was analyzed by X-ray diffraetion(CoKu radiation) on oriented mounts for the four clay mineralgroups kaolinite, smectite, illite, and chlorite following standardproeedures given in detail by Pc/schick e/ af. [1996]. Theseprocedures assume that the four main elay mineral groups addup to 100% in the fraction <2 pm and involve the use of theweighting faetors of Biscayc [1965]. As a eonsequence ourabsolute percentages for the individual elay minerals cannot becompared to those of .lanes [1984], who used weighting faetorsof I-Iea/h (lild Pisias [1979]. However, general patterns of elaymineral distribution are similar.
Magnetic suseeptibility was measured on the arehive halvesof the eores at a I cm spacing using a Bartington InstrumentsM,S,2.C loop sensor. Lomb-Scargle Fourier transform [Loll/h,I97G; Scargfc, 1982, 1989] embedded in the Spectrum program[Schuh ami S/al/eggcr, 1997] was used for all spectral analysis,This speetral estimation method ean be direetly applied tounevenly spaced geologieal times series. Cross-spectral analysisfor GeoB 2821 was performed using Welch' s [] 967] overlappedsegment average (WOSA) procedure to reduce spectral peaksoriginating fl'om random fluctuations.
2.1. StratigraphyofGeoB2110
A preliminary stratigraphic framework was established onshipboard counts of the planktonie foraminifera Gfoboro/alia
Il/enardii [Muli/za, 1993]. On the basis ofthe cyc]ic appearaneeof this speeies, Ericson (lild Woflin [1968] defined abiostratigraphic zonation scheme using a letter notation from Z(Holoeene) to Q in order of increasing ages. Zones Z to U ean bedirectly correlated to oxygen isotope stages with following agesat zone boundaries: ZN, ]2 ka; Y/X, 80 ka; X/W, ]30 ka; WIV,185 ka; and V/U, 370 ka, Though minor shifts in stageboundaries occur because of the poor resolution of the G,mel1ardii eounts, the preliminary stratigraphic fi-amework waseonfirmed for site GeoB 21] 0 by earbonate stratigraphy afterDall/u/h [1977]. Oxygen isotope measurements on planetie andbenthie foraminifera were earried out on a Finnegan massanalyzing technique (MAT) 25] mass speetrometer (Figure 2).The measurements of the planktonic foraminifera Gfobigeri
noides sacculifcr were partly eomplicated by low sedimentationrates and enhaneed carbonate dissolution but enabledidentification of oxygen isotope stages 6,0, 6.2, 7,0, and 7, I.Supplementary measurements of Uvigeril1a spp. near isotopestage 4 clarified the stratigraphie position of stages 4.0 and 5.0.In addition to isotope stratigraphy, some eharaeteristie patternsof the high-resolution magnetic susceptibility record wereeorrelated to the SPECMAP stack (Jmbrie el al., 1984] andsupplied four more tie points (near oxygen isotope stages 2.2,
3,0, 5,5, and 6.4), Because of different phase lags of plünktonlcand benthic foraminifera and magnetic susceptibility, no phaseinformation can be dedueed from this combined age model.
2.2. Stratigraphy 01' GeoB 2821
The Cl. Il/enardii counts for core GeoB 2821 elid not producca reasonable pattern. This is due 10 the wide sampie spaeing ofthe shipboard sampies and the low sedimentation rates of thisCOlT. Since no 0180 stratigraphy is available an agc modelderived by orbital tuning of the susceptibility rccord \\as ustdhere, This cyclostratigraphy was established in the framework ofa stratigraphical synthesis of 12 sediment cores from the subtropical South Atlantic Ocean (subtropical South Atlantic susceptibility (SUSAS) stack) [VOll Dobcneck and Sc/I/I/icder,
1998], based on cyclic and highly coherent magnetic susccptibility logs. The eorrelation of suseeptibility and carbonate content is inverse as in many marine environments [e.g., Robinson,
1990]. This indicates that the rock magnetic signal is a result ofvarying dilution of the (terrigenous) magnetic sediment fractionby glacial-interglacial carbonate dissolution cyclcs. On the basisof a detailed stepwise alternating field demagnetizatlon of theNatural Remanent Magnetization (NRM), three paleomagneticage marks could be identified for GeoB 2821 (Figure 2), Asimple age model generated by linear interpolation of thesereversal ages alreacly shows spectral characteristics 01'Milankoviteh cyclicity (100 anel 41 kyr cyeles), suggestingclimatic forcing. Correlation to a precisely dated paleoclimaterecord was necessary to improve the age-depth relation prior toorbital tuning, The 8 13C record of South Pacifie Ocean DrillingProgram (ODP) Site 806 (T. Bickert, unpublished manuscrJpt,1998) exhibits a eontinuously high pattern similarity with thesusceptibility reeord of GeoB 2821 and was therefore used asage referenee. The preclominant 40 kyr cyclieity of susceptibilitywas then synchronized with astronomically calculated obliquityvariations [Bcrgcr al1d LOIl/rc, 1991] as a target curve. A phaseshift of -30 0 (-3.4 kyr) was found by cross-spectral analysis ofmagnetic susceptibility and 8 I go for one of the SUSAS records[VOll Dobel1eck alld Schmiedcl', 1998] (8 180 data by Bickerl
[1992]). This value corresponds to those ealeulated by Imbrie el
al. [1993] for early proxy responses in the SouthernHemisphere, As the SPECMAP stack [Imbrie cl af, 1984] isassumed to lag obliquity by 7.9 (7,4 - 8,2) kyr, a net lag of 4.5kyr was applied for the tuning ofmagnetie susceptibility.
From the observation that faintly visible precessional peaksin the speetra gradually sharpened at both refinements of the agemodel the obliquity-tuned primary susceptibility record wasfiltered (15 - 47 kyr) to extract obliquity and precession cyclesand matched to a eomposed target curve of normalized obliquityand preeession.
2.3. Stratigraphy of GeoB 2822
According to Cl. mCllardii stratigraphy [.lahll, 1994] coreGeoB 2822 reaches zone V of Ericsoll alld Wolfill [1968]. Noisotope data were available to confinn these ages. Pa]eomagneticmeasurements showed a uniform normal polarity throughout theeore, The more detailed glacial-interglaeial cycles revealed bythe susceptibility record justify minor shifts of assumed stageboundaries ZN (isotope stages 1/2) and WIV (617), As in theother cores, kaolinite/chlorite and siltlclay (grain size) ratiosgenerally correlate with the glacial-interglacial susceptibility re-
72 Chapter 5
G menaJ(iii Susceplibilily (10-3 SI) K/C ratio (1180 (',\JnVS PDß)G
SiltlClay
200
300
~
~ 400
mou
500' "\_
·----S~
600 ~-~
700
800'
1.5 1.0 0.5 0.0
G mcnardii Susccptibility (10.3 SI) I<lC ratio SiltlCI8y ChRM lnclinalion n CI:RM Declirwtiofl n Polar;ty
0 200 400 600 SOO 07 0.6 0.5 OA 03 02 0.5 10 1.5 2.0 1.2 10 0.8 0.6 ·90 -45 45 ·90 90 180 270o· -=:::::::= ~ ~ ~ ~~-=--=--
~5-:~-
-~-=--=tr;eOB 2821 =:----~ .-<;.' .~
100 L~ J "~~~
.~~->
-=~j:c- -~~------=
==-----~ ~~ J~ (200~~~ C~ -<---~ -~ -> t-------~ )~ /5 '"S- C '( '"K
.c
=~.3 ~ §
? ~, m~ =- ~~-~ '-.~ '----...__._-~ ..... -
~ ~3 -~:~-=~ ro
~ __~~J- E
====3(\J
--=.._.____-.'1,.,
l:::J
1iil~__~ L_-
72
~~
f~ ~~ ~ -=;,
~ ~ -=r~ 1, ~ -4
700 ~~
800 ----- ~
G. menBrdii Susceptibilily (10-3 SI} K/C ratio (S01001hed x 3) SiitfClay Polarity
1000.L..--------.l....--------.l....--------.l....-------~
900
10 20 30 40 06 05 DA 0.3 0.2 01 0.0 0.2 DA 0.6 0.8...,
2·4
-;;~ 5
---==-
1.0 14 1.2 1.0 0.8 0.6 04
Terrigenous flux in the Rio Grande Rise area during the past 1500 ka 73
Table 2. Data of Cross Spectral Analysis
r'CI'SIiS RCI'!'!'scd ODP677 (yISO-33±8(13) 094 -31±10(15)
V!'!'SIIS R!'I'CI'scd S'PFCMAP Slock ö·ISO-5±3(5) 0')') -32±3(5)
100 kyr- 1 Bandk °Phi
093 -46± 12
T M,kyr kyr
500 50
220 ]()
400 3()
388 2.0
1525 50
500 50
41"X7
23 kyr I Bandk °Phi
08'\
11±14
-42±10(13)
41 kyr I Bandk °Phi
005
087
086
-6±7
16±15
-202(3 )
096
099
087
0.96
Kaolinite/eh lorite(Geoß 2821-1), 0-500 kaMaximum DDVostok ice eore[JoIl2!'1!'I n/., 1')')41%,NADW index[Rn)'/IIo!'1 nl., I(NO]Kaol in ite/eh lorite43° S [Di!'k/llllllll !'I o/., 1996]
Kaolinite/eh lori tc(GcoB 2821-1 ). 0-1500 kaKaolinite/eh lorite(GeoB 2821-1 ), 0-500 ka
Abbreviatiolls are k, cohereney; °Phi, phasc angle with 80% eonfldenee interval (in parentheses 05'%);1: maximum age used in the calculation; :.\1,
sall1ple interval. The 8 "0 SPECMAP stack was taken from l/IIb!'ie ('/ o/. [10841, ami the 8"0 record Oecan Drilling Program Site (ODP) 677 flOmSiloeklelol1 el o/. [1090].
cord with the exception of stage 6 (Figure 2). Still, individualmaxima in the susceptibility rccord within stage 6 cOITespond tomaxima in the precessional index around 133, 160, and 185 ka.Minima can be related to preeessional minima and peaks in thekaolinite/chlorite record. Nevertheless, absolute values for theminima in susceptibility are rather low for 21 glacial stage. Sincethis is 21 featurc unparalleled in the othcr COlTS it cannot bc satisfactorily explained by 21 common mechanism. Further detailedpaleomagnctic investigations would be required to investigatcthis local, timc-restricted phenomena. The final age models ofall corcs were obtained by linear interpolation between stratigraphie tie points (Figure 3).
3. Results and DisCllssion
3.1. Surface Distribution
Sinee kaolinite and chlorite are the proxy clay mineralscommonly associated with propagation of NADW and AABWwe use the kaolinite/chlorite ratio to delineate the extension ofthe deepwater masscs as demonstrated in the approach ofPetschick et al. [1996]. Additional sources of kaolinite shouldrefiect in exceptionally high kaolinite/chlorite ratios. Kaolinite/chlorite ratios of the surface sampIes are depicted in Figureland reveal some characteristic features of sediment input anddistribution. Extremely high values between 10 and 80 arerecorded off the mouth of the Rio Doce and confirm the role ofthis river as an important kaolinite source as already stated by
Figlll'c 2. (opposite) Parameters lIsed for the stratigraphie classifieation01' the COlTS. Age models tor cores GeoB 21 10 and GeoB 2821 are basedon shipboard counts ofGloborola/in menardii. Oxygen isotopes in coreGeoB 2110 were measured on the planctie toraminifera Globigeril1oidessncclIlz!er and some bcnthie Uvigeril1a peregrina. The stratigraphieJi-amework 01' GeoB 2821 is based on paleomagnetic analyses and orbitaltuning 01' the magnctic suseeptibility reeord. Also depicted arekaolinitechlorite ratios and grain size (ratio 01' silt/elay weight percentages),whieh nuctllate in a glacial-interglacial pattern. Note that the reeords 01'
suseeptibilty and grain size are reversed for bettel' graphic eorrclation.
DeMelo et al. [1975]. The highest values art": found on tht": sht":lfand upper slopt":. However, it is important to note that enoughriver suspension is carried through the water column to keepkaolinite/chlorite ratios above 10 at 4000 m watt":r depth. Evensouth of 25°, valut":s exceed 2 below 4000 m water depth. Below4000-4100 m in the Vema and Hunter Channt":ls and the abyssalplains of the Argentine and Brazil Basins, kaolinite/chloritt":ratios fluctuate from 0.5 to land confirm the concept of AABWas 21 source and carrier of chlorite [Biscaye 1965; .lalles, 1984;Petschick et al., 1996]. Some of the river suspension introducedby the Rio Doce is carrit":d south by the Brazil current.Scavenging and incorporation into fecal pellets are believed toremove most of the fine-grained material from the water COlUll1l1rather rapidly [Deuser et al., 1983]. Accul11ulations of kaoliniterich sediments are found on the continental slopes off Cabo Frioand with decreasing values above and on the Sao Paulo andSantos Plateaus from 2500 to 3000 111, Since the rivers south ofCabo Frio contain relatively little kaolinite the Rio Doce has tobe regarded as the major source for these deposits [Jolles, 1984].
A characteristic enrichment of kaolinite was also found atintermediate depths (3000-4000 m) on the RGR by Chamley[1975] and JOlles [1984], with the maximum centered at 3300 to3600 m. Such an enrichment is confirmed by our sampIes, whichcover 21 depth range from 4500 to 2900 m water depth on thewestern flank of the RGR. A similar feature is observed Oll thetransect from the eastern slopt": of the RGR to the Hunter ChallneI. East of the Hunter Channel toward the MAR kaoIinite/ch lorite ratios are lower « I) at comparable water depths(Figure I). Only on the top of the ridge maximum values of 1.0are reached_ This has important implications concerning theorigin and propagation of kaolinite.
JOlles [1984] proposed two models to explain the kaoliniteenrichment at mid-depths of the RGR: (1) advective transport ofkaolinite from low latitudes with NADW and (2) advection ofkaolinite along isopyenals from resuspension of kaolinite-richdeposits on the Sao Paulo Plateau. Though he did not rule outthe NADW transport 1110del complt":tely, the isopycnal mixingmodel was bettel' suited to explain his observed elay mineral dis-
74 Chapter 5
Srnectite % IIlite% Kaolinite % Chlorite % Kaolinite/Chlorite ratio
6 82010o 10 20 30 404020 30403020
o ·-r·-~~--t~~~"';:!'~~~~~r~~~==>".bdT~~~~~~,_Iiiiiä""''''''":1
so
100 5
ro-"<lJ01<i
150
50
100
'"-"<lJ01<i
150
200
250-1...-------'--------'--------'--------'- -'
Terrigenous flux in tbe Rio Grande Rise area during the past 1500 ka 75
tribution. Kaolinite/chlorite ratios determined on our sedimentsurface sampie set yield arguments for both transport models.Thc symmetrieal nature 01' the kaolinitc cnrichment at middepths01' both sides 01' the RGR eould support thc eoncept of advectivetransport 01' kaolinite from low latitudes with NADW.Consequently, similar kaolinite/chlorite ratios should beexpected at comparable depths 01' the MAR. However, only aslight rise in kaolinite/chlorite values is recorded on the top,whieh may rcsult from minor adveetion 01' kaolinite from lowlatitudes. On the other hand, although the transport of kaoliniteto the western slope 01' the RGR along isopyenals tl'om the SaoPaulo 01' Santos Plateaus may explain the kaolinite enrichmenton the western side 01' the RGR, it is hard to imagine a mechanism that transports kaolinite-rieh suspensions around 01' acrossthe RGR to the "backside." 1'0 explain the symllletriealnature ofthe kaolinite enrichment on the RGR, we alternatively suggestthe injection 01' kaolinite-rich suspensions into intermediatedepths 01' the NADW off the mouth 01' the Rio Doce and shortdistance transport southward instead 01' long-distance advcetionfrom lowcr latitudcs. This idca is supportcd by high kaoIinite/ehloritc ratios down to 4000 III wafer depth off thc rivermouth, dClllonstrating that kaolinitc-rich suspensions are able torcach decpwatcr Icvels. Injeeted at 20 oS, these suspensionscould be advceted south with the NADW anel be e1epositeel onthe first obstaclc, whieh is thc RGR.
Our results suggest that a substantial part 01' the terrigenousseelimentation on the RGR originates from freshwater input anelsuspension supply by the Rio Doce as evidenced by the kaolinitecnriehment at intermcdiate depths 01' thc RGR. Thc qucstionarises of whcther tcmporal variations in kaolinite supplyprimarily record pulses 01' fluvial discharge 01' rather representvarying intensities of short-distance NADW flow. 1'0 aelelressthis problem and look at the temporal variations 01' the kaolinite/chlorite ratios, we investigateel three seeliment COlTS.
3.2. Temporal Variations 01' Clay Mineral Input
Core GeoB 2110 was taken from 3000 m wafer e1epth at thecontinental slope above the Santos Plateau. Kaolinite/chloriteratios in surface sediments reach a maximum at this e1epth (Figure I) because 01' advection anel deposition 01' kaolinite from thenorth with the Brazil curren!. Significant fluctuations 01' chloriteflux can be ruled out 1'01' this site because the only potentialcarrier of chloritc, AABW, always stayeel weil below thcsce1epths in thc southwestern Atlantie [Cuny, 1996]. Sitc GeoB2110 shows highest kaolinite/chlorite ratios (4-8) in interglacialstages I, 5 and 7 and lowest ratios (± 1) in glacial sections (Figure 3). This is still in aeeordance with the model 01' .Iones
[1984], who proposeel a southward transport 01' kaolinite fromthe mouth 01' the Rio Doee only during times 01' high sea level(interglaeials) anel a direet injeetion 01' the river load into intermediate and deepwater levels during times 01' low sea level
Figllre 3. (opposite) Relativc clay Illineral percentagcs and kaolinite/chlorite ratios of cores GeoB 2110, GeoB 2821, ami GeoB 2822versus agc. In ordcr to producc a maxilllulll glacial-intcrglacial contrastthe cutoff for the shading 01' kaolinite/chlorite ratios was set at anaverage valuc 01' 1.0 101' the cores within the North Atlantie Decp Water(NADW) (GcoB 2110 ane1 GeoB 2821) anel at 0.6 for the core within theAntarctic Bottom Water (AABW) (GeoB 2822). Isotope stages 1-7flmbric cl a/., 1984] are shaded.
100 kyr 41 kyr rkyr
,
19 kyr
CO 152.. 1 1
E: /\:::J
10'-U :! ~Q)0.. I 1I 1\Cf)
-0 5 T'/ '\Q)+-' Y 1\CUE JI 1\t5
~/1 1\
W 0 I +-1 I II--~-l
0 0.02 0.04 0.06
Frequency [1/kyr]
Figllrt' 4. Speetral analysis ofkaolinitclchlorite ratio orcorc CieoLJ 2110.In addition to speetral peaks near I(JO and 41 kyr, strang response topreecssional löreing hccomes visible in the modifJed pcriodogralllcalculatcd by Lomb-Scargle Fourier transrorlll using a Welch window.Thc cross dcpicls (, dB bandwidth (horizontal) ami 80'% cunlidcnceinterval (vertieal).
(glaeials). However, our surface data show high kaolinite/chlorite ratios e10wn to 4000 m wate I' e1epth also in recent(interglacial) sediments off the Rio Doce. Furthermore, if scalevel exerts a major control on the input of kaolinite tiom theRio Doce, minima in kaolinite/chlorite ratios could bc alsoexpcctcd at the e1eep sites GeoB 2822 and 2821 during highstands. Howevcr, both e1eep sites show high kaolinite!chloriteratios during intergiacials. Thereforc we hypothesi/e that apotential effect 01' sea levcl ehanges on the input of kaolinite toall our core sites is overridden by fluctuations in the totalamount 01' river e1ischarge.
Spectral analysis of the kaolinite/chlorite ratio 01' core GeoB21 10 (Figure 4) strengthens this assumption. Although the eombined ehronostratigraphy e1escribed above does not preserve theexact phase relation 01' different proxies, a Fourier transformyields information on speetral characteristics within the uncertainties 01' the dating proeedurc. The resulting spectru111 indicatesthat there is a strong response of the kaolinite/chlorite ratio toprecessional foreing. There is abundant evidence (summarizedby Delvfenocal [1995] and Gingeie el al. [1998]) that low-Iatitude aridity-humielity cycles documenteel in eolian flux, monsoon intensity, 01' river discharge respond to variations in lowlatitude insolation, which in turn result from Earth' s orbitalprecession. Therefore we concluele that the kaolinite/chloriteratio at site GeoB 21 10 is significantly influenced by low-Iatitude river discharge.
The spectrum also e10euments i00 and 41 kyr variance(Figure 4). The dual nature 01' high- and low-Iatituele forcing 01'low-Iatituele climate has been recognized by DeMenocal [1995]
in various sites around the African continen!. Mechanismstransporting high-Iatitude signals, namely the pronounced 100and 41 kyr cycles, to low latitueles are believed to be eoolingNorth Atlantic surface water anel fostering shifts in atmosphericcirculation anel migration 01' vegetation belts. Especially the 100kyr frequency 01' eeeentricity is prominent in many aridityhumielity reeorels [Paslourel el al., 1978; DeMenocal el al.,1993; Gingeie el al., 1998].
76 Chapter 5
Age [ka]
0 200 400 600 800 1000 1200 1400
3 1.8
f2 1.6 c:I'-c.o 00.- 14 ,
4 (\)0 ......0 o'~ 1.2 ~.
m G)CD
0 1.0 00.- 0]
0 5 N~ 0.8 (XlE.-- N0 --'oe
Io:J --'oe
vo 0.6
6
0 200 400 600 800 1000 1200 1400
0.5
o0.060.02 0.04
Frequency [1/kyr]
---A k __ I,-
- ----- - -----~ -A --~-
--I
I
I I AII I I I -I
:l~I I
I
I~I I
I I I
+- I I I
80%level
0.06 00.02 0.04
Frequency [1/kyr]
o
10 10m2.E::::J 0 0'-.-üQ)0.
Cf)
"0Q) -10
CDE
:;:::;(()
w-20
>,ücQ)'-Q)
..coU 0.5"0~cu::::J
g 0
Figure 5. Cross-spcctral analysis ofkaolinite/chlorite ratio 01' core GeoB 282 I (black lines) and the reversed8 180 recorcl 01' Ocean Drilling Program(OOP) Site 677 (white lincs) [Shacklelol1 el a/.. 1990]. While the past 500 kyr are dOlllinated by eccentricity-relatecl 100 kyr cycles (A), an increasedalllount ofvariance in the obliquity band near 41 kyr appears in the spectra 01' the 0-1500 ka records (B). Both time serics are intel1JOlated to 5 kyrsteps, the average time resolution 01' the kaolinite/chlorite recorcl 01' GeoB 2821. The analysis was performed using We!ch's [1967] overlapped segment average (WOSA) procedure with two ovcrlapping segments and a Welch window. The sign ofthe 8 180 record ofOOP 677 was inverted to attainclirectly intelvretable phase relations. Crosses depict 6 clB bandwidth and 80% confidencc interval. The resu1ting negative phase angles shown inTable 2 indicate a lead with respcct to 8 180. High squarcd cohcrcl1cy is calculatecl in the 100 and 41 kyr band.
Tenigenous flux in tbe Rio Grande Rise area during tbe past 1500 ka 77
Core GeoB 2822 from the eastern terraee of the Vema Channel (4300 m water depth) is believed to represent sedimentationwithin the AABW during interglaeial as weil as glaeial times.The adveetion of chlorite from the south is doeumented in thelowest kaolinite/chlorite ratios «1) of all investigated eores.Although of low amplitude, a signifieant glaeial-interglaeialpattern in the kaolinite/chlorite signal ean be reeognized, withhigher values in the interglacial seetions (Figure 3). Evidencefrom grain size and sediment texture from the area [Jolmsoll emdRaslllllssell, 1984; Masse el !l1., 1994] indicates a more vigorousnow of AABW at the transition from warm to cold periods andwithin glacials. !nereased adveetion of chlorite with intensifiedAABW activity dming glacials could explain the observednuctuations in kaolinite/chlorite ratios of eore GeoB 2822. Sincethe site is weil below the direet innuenee of NAOW the adveetion of kaolinite from low latitudes is an unlikely source forGeoB 2822. Therefol'e we remain with the Rio Ooce as theprimaJ'y kaolinite souree also for this eore.
Cyelie variations are also reeorded in kaolinite/chlorite ratiosof COlT GeoB 2821 (Figure 3). High values correspond to lowsin the suseeptibility reeord (Figure 2) and thus interglacial stages(Figure 3). Thc clay mineral proxy nuetuates in tune with manypalcoclimate proxies dominated by high-Iatitude orbital foreing.A shift in the dominant perioel of variation from 41 to 100 kyr isvisible near 1000 ka (Figure 3).
Cross-speetral analysis of the kaolinite/chlorite ratio of coreGeoB 2821 was performed versus thc reversed SPECMAP 8 180stack and reversed 8 180 reeord of OOP 677 for the past 500 kyrand and versus the reversed 8 180 record of OOP 677 for the past1500 kyr. It appears that the elay mineral proxy is eoherent withglobal iee volume in the 41 and 100 kyr periods of obliquity aneleeeentrieity (Table 2 and Figures 5 and 6). Resolution and sedimentation rates of the eore are too low to investigate a preeession-related variability. Moreover, statistiea! data (Table 2) forthe complete kaolinite/chlorite reeord are difficult to interpretsinee signifieant variation in the 100 kyr banel oeeurs onlyduring the past 350 kyr, where 41 kyr eyeles, elominating theolder section, are redueed (Figure 5). As a consequenee of thisrefleetion of the mid-Pleistocene elimate transition [e.g., RliddiIllall el a!., 1989] we eoncentrate on the seetion from 0 to 500 kafor the !00 kyr eyclc anel make use of the whole reeord for the41 kyr period. As statistical data (Table 2) and phase wheels(Figure 6) indicate, maxima in GeoB 2821 kaolinite/chloriteratios lead maxima in NAOW nux by 5800 years in the 100 kyrcycle (0 to 500 ka) and 4 I00 years in the 41 kyr eycle (0 to 500and 0 to 1500 ka). These results are consistent with characteristic values of early proxy responses in the Southern Hemisphere[Imbrie el a!., 1993].
NAOW production and propagation have varied in orbitaltime scales, though the relation to climate is not simple [Clirry,1996]. In the western Atlantie (Ceara Rise), only NAOW wasrecoreled in some peak intcrglacials, whereas in some glacials,southern source deepwater reached 3200 m water depth [ClIny,1996]. Investigations in the eastern part of the South Atlantic[Dieklllallll el al., 1996] have shown that the propagation andsouthwarel extension of northern source deepwater (NADW)were more pronounceel during interglacials, thus providing morekaolinite during warm periods. At first glance, kaolinite/chloriteratios of core GeoB 282 I fit nicely in the concept of alternatingintensities of interg!acial NADW [Diekmallll el al., 1996] and
glacial AABW [Masse el al., 1994] now. Within the error margins (Table 2), phase angles of kaolinite/chlorite ratios in the100 kyr band of eeeentrieity are similar in the castern anelwestern Atlantie. The implieations for the RCiR are tllat eitherthe kaolinite/chlorite ratio is eontrolleel by ehanges in deepwaternux or that the high-Iatituele forcing meehanisms, whieh controleleepwater nux and low-Iatitude elimate in the eeeentricity baneI,do not show a signifieant time lag. However, there are strongarguments opposing an exelusively deepwater-controlled elaymineral supply for sites GeoB 2821 as weil as GeoB 2822.
First, AABW aetivity was strongly reduced after 350 [Masse'el al., 1994] 01' 275 ka [JOhllSOll alld Ra.s·1ll11,,·sell, \984], respee..tively, wh ich would lead to rising kaolinite/chlorite ratios f"rom350 ka to present. Instead, we find deereasing kao)inile!chloriteI'atios in COlT GeoB 2821.
Second, in the 41 kyr band, maxima of kaolinite/chloriteratios of eore GeoB 2821 signifieantly lead minima in ice volume, maxima in 'Yr,NADW tlux [Raymo el al., 1990] andmaxima in kaolinite/chlorite ratios in the eastern South Atlantie[Dieklllallll el ClI., 1996] (statistical elata in Table 2). Moreover,the phase angle of kaolinite/chlorite ratios of core GeoB 2821 inthe 41 kyr perioel is elose to the maxima of aD (deuterium) inthe Vostok iee core [JOlizel el ClI., 1994]. Maxima in the aDreeorel inelicate perioels of inereased atmospheric temperatures inAntaretiea [Waelbroeck el al. , 1995], whieh are relateel to intensifieel evaporation anel preeipitation in lower latitudes. Consequently, we suggest that temporal ehanges in the kaolinite/chlorite ratio of the RGR eores may be rather interpretedas an atmospherie signal of preeipitation anel rivcr input than asa proxy of eleepwater fluetuations.
4.Paleoceanographic and PaleoclimaticImplications
Summarizing the evidenee from elay mineralogy, grain size,anel magnetie susceptibility, we finel that the flux of terrigenousmatter on the RGR shows eyclie variations at least in the 100anel 41 kyr periods. fnterglaeial stages are eharaeterized by thedeposition of fine-grained, kaolinite-rieh terrigenous matter oflow magnetie suseeptibility. They reeord perioels of inercaseelhumidity on the South Ameriean mainland and runoff of the RioDoee. High nuvial eliseharge was also reeoreleel for the Orinoeoanel Amazon Rivers eluring interglacials [B01vles ami Fleischer,1985]. C1apperlon [1993] reported more humid eonditions forinterglaeial stage 5 in Patagonia. The eoneept of increaseelhumielity in interglaeial perioels (100 kyr perioel) is in aeeordanee with results from multiple investigations on the Afrieaneontinent (summary in DeMenocal [1995]). Runoff of majorAfriean rivers anel monsoon intensity fluetuates in the 100 kyrperiod of global iee volume and 23 kyr perioel of low-Iatitudeinsolation [PaSIOl/rel el al., 1978; Zachariasse el al., 1984;Rossignol Sirick, 1985; Pokras, 1987; Gillgele el al" 1998]. Ourdata from the RGR yield similar results for the South Amerieaneontinent at 20 0 S in the 100 and 41 kyr periods. In eore GeoB2 110 the resolution is sufficient to record fluetuations of kaolinite nux in the 23 kyr precessional band. They are believeel torepresent humidity anel discharge variations of the Rio Doeeinitiated by low-Iatitude insolation changes.
Minima in the preeessional index eorrespond to inereasedinput of kaolinite anel may be interpreted in terms of higher
78
1OO-kyr cycle
Max. Eccentricity
Max. Kaol./Chlor.43"S SE-Atlantie
I/
Max. ieeSPECMAP
Max. ieeSPECMAP
Chapter 5
41-kyr cycle
Max. Obliquity
Min. iee SPECMAP
.-i.~=:::::::;;;~~~IE~M'ax."%NADW" flux (41"N)-- Max. f<aol./Ch!or.
43°S SE-Atlantie
Figurr 6. Phase relallonships hctween kaolinite/chlorite ratios ofco]T Gcoll 2821 (0-)00 ka), 'X, N!\DW /lux IRo)'1I1O ('{ 111.. 1<J()Ojlaolinitc/chloritcratios 01' a core 43°S in the SE i\tJantic IDiekllloli1i cl al., 199ö!, ami thc maximum ol~D (deuterium) in the Vostok iee cme l.JolI~cI Cl 111., 1')9.\] in tlw100 ami 41 kyr pcriods 01' ccccntrieity and obliquily (statistical data sec Tablc 2).
humidity in the drainage area ofthe Rio Doce. It is interesting tonote that at about the same latitude west of the Andes in northern Chile, precession-related humidity/aridity cyclcs are recordcd, which show exactly the opposite pattern [Lall/)'. cl al.1998]. There, maxima in the precessional index correspond tomore humid eonditions in the sediment record. Inereased precipitation there is related to a shift in the position ofthe SOllthernwesterlies and indicates a nonsynchronous behavior of SouthAmerican palcoclimate on an orbital timescale.
Glacial stages are characterized by maxima in magnetic susceptibility, illite, and grain size (more silt) and lows in kaolinitesupply. Cold and dry conditions have been reported for theheadwaters of the Rio Doce during the last glacial [Be/ding andLiehle, 1997]. The lack of dilution by fine-grained river suspensions. a more vigorous AABW flow, and the increased input ofsilt-sized dust, rich in magnetic partielcs and i1lite, may havecombined to form the terrigenous signal of the cold periods.Previous studies have shown that major fluctuations in dust fluxappeal' to be of global signifieanee with maximum inputs duringcold and arid periods [elell/ens and Prell, 1990; DeAlenoeal 1'1
al., 1993]. A potential souree area for dust in the southwesternAtlantie is Patagonia. Patagonian dust is derived from loessdeposits, whieh contain a high pereentage of titanomagnetite andillite [Teruggi, 1957; Bonorino, 1966; Zarale and Balsi, 1993].Although insignifieant for the total terrigenous mass balance, theinput of Patagonian dust in cold periods eould seleetivelyenhanee the suseeptibility signal preformed by glaeial carbonatedissolution events. lt mayaiso eontribute to an illite enriehmentof glaeial eore seetions.
It appears that the eomposition of the terrigenous matter onthe Brazilian slope and also on the RGR responds to ehanges inhumidity/aridity on the adjacent South Ameriean eontinentrather than to variations in deepwater adveetion. The dual influ-
ence of high- and low-Iatitude foreing on the palcoelimate in theBrazilian hinterland is believed to result in periodieal disehargeof the Rio Doee. Consequently, the supply and deposition ofkaolinite on the slope and the RGR are reeorded for the majorMilankovitch periodieities.
Multiple potential sources complieate the interpretation ofthemain elay mineral eomponents smeetite and illite. Nevertheless.individual peaks in the downeore reeord of smectite in eoreGeoB 2821 are fi'equently assoeiated with maxima in kaolinite/chlorite ratios and thus with warm and humid periods,whereas i1lite shows a reciproeal pattern. More evident is a longterm trend in smeetite and illite percentages (Figure 3). Illitepereentages inerease while smeetite percentages, as weil asaverage kaolinite/chlorite ratios, decrease from 1500 ka topresent. If the paleoelimatie interpretation of the elay mineralproxies is eorreet, a trend towarcl more arid and glaeial eonditions during the Pleistoeene is evident. These findings are eonsistent with results from the Afriean eontinent [Del'vfenoeal.1995]. Here proxies for Afriean aridity reeonstruetecl fromvarious ODP sites around the eontinent show a gradual trendtoward more arid and cooler eonditions from the Plioeene topresent with prominent shifts arouncl 2800, 1700, ancl 1000 ka[DeMenoeal, 1995].
5. Conclusions
Clay mineral analyses revealed that the Rio Doee (Brazil) is amajor souree of kaolinite for marine sediments at 20o-30oS inthe western Atlantie. Charaeteristie enriehments of kaolinite areobserved at intermecliate depths of the eontinental slope and onthe flanks of the RGR at 3000-4000 m water depth.
The supply of kaolinite with NADW is minor as evideneeclby the eomparison of kaolinite/chlorite ratios from the MAR and
Tenigenous Dux in the Rio Grande Rise area during the past 1500 ka 79
thc RGR, Conscqucntly, pattcrns of tcrrigcnous scdimcntation in
this part of thc Atlantic rccord climatic conditions on thc South
Amcrican hintcrland.
Low-Iatitudc insolation fostcrs pcrioelical elischargc of thc
Rio Docc in thc prcccssional 23 kyr baneI, as cvidcnccd in kao
linitc/chloritc ratios on thc Brazilian slopc, Cyclic variations of
kaolinitc/chlorite ratios on thc slopc anel in a corc from thc
wcstcrn flank of thc RGR arc cohercnt with global icc volumc in
thc 41 anel 100 kyr pcrioels, Thcy are also bclicvcd to rccord
fluctuations in thc dischargc of thc Rio Doce and mirror humid
ity conelitions on the adjaccnt South Amcrican hintcrland, Hu
miel pcrioels arc cocval with warm intcrglacial phascs, whcrcas
ariel pcriods correspond to colel, glacial stagcs, Thc elualmodc of
high- anel low-latitude foreing of low latituele clilllate is consis,
tent with silllilar findings li'om thc African contincn! A long
time elccrcasc in smcctitc content and kaolinite/chlorite rarios
from 1500 ka to present is bclieved to elocul11cnt a trcnd toward
more arid and cooler climatc conelitions for subtropical southcrn
latitudcs of South America,
Acknowlcdgmcnts, The dcdication and etrort of crew and mastcr o!"R/V Meleol' is greatly acknowledged. We are grate!"ul to r:. Bassinot andD, Dobson 101' construetive reviews allCl hclpl'ul commenls. All thc dalaused in this manuseript are arehived in the inl'ormatiou systemPANGAEA/SEPAN (www.pangaea.de). F, Sehmleder was supported bythe Deutsche Forsehungsgeilleinschat't in the framcwmk o!"Ciroduierteukolleg 221, This is contribution 249 of spccial research p1"O.leet SFB 261.
References
Balsam, W I,., B, I" Otto-Bliesner, ami 13, c:.Deaton, Modem ami Last ()Iaeial Ivlaximumeolian sedimcntation pattems in the AtlanticOcean intcrprcted ti'om sediment iron oxidecontent, l'al!'Oceanogmphy, I IJ, 493-507,1995,
ßehling, H., and lvI. Lichte, Evidence of dry amicold climatic conelitions at glaeial times intropical southcastern Brazil, Qllal, Res., 48,348-358, 19'J7
Berger, A" ami M,F, Loutre, Insolation valuesfor the cllmate of the last 10 million years,Qllal, Sei, !?e\', IIJ, 297-317,1991.
Biekert, 1'., Rekonstruktion der spätquartären130denwasscrzirkulation im östlichenSüdatlantik übel' stabile Isotopc benthischerForaminiferen, l3er. {'achhereich Geowiss.Univ, l31'ell/en, 27,205, 1992,
Biscaye, PE., Mineralogy and sedimentation ofreeent deep-sea clay in the Atlantie Oeeanand adjacent seas ancl oceans, GSA 131111.,76,803-832, 1%5,
Bleil, U" anel Cruise Pal'ticipants, Report andpreliminary results ofMeleol' cruisc 23/2 Riode Janeiro - Recife, 27,2,1993 - 19,3,1993,Bel', Fachbereich Geowiss, Univ. l31'ell1ell,43, 133, 1993,
Bleil, LJ" and Cruise l'articipants, Report andpreliminary reslJits of Meleol' eruise 29/2Montevideo - Rio de Janeiro, 15,7,1994 08,08,1994, Bel', Fachbereich Geowiss,Ulliv, l31'ell/en, 59, 153, 1994,
Bonorino, F, Ci., Soil clay mineralogy of thePampa Plains, Argentia, .J. Sedilllelll, Res"36,1026-1035,1966,
Bowles, F, A., and P, Fleischer, Orinoeo andAmazon River sediment input to the easternCaribbean Basin, M(/I' Cleol., 68, 53-72,1985
Boyle, E, A" Caelmium: Chemieal tracer ofdcep-water paleoeeanography, Paleoceallogmphy,3,471-489,1988,
Boyle, E. A" A comparison of carbon isotopcsallCl cadmium in the modern and glacialmaximum ocean: Can we aCCOlull for theeliserepaneies? Cal'boll Crelillg ill Ihe CllacialOceall: COl1slmil1ls 011 Ihe ()ceal1 's Roll' il1Global Chal1ge, NA 1'0 ASI 5'el' C. vol. I 17,edited by 11., Zahn et al., pp. 105-144,Springer, New York, 1994.
Chamley, H., Inlluence des eourants profonds aularge du bresil sur la seelimentation argeleuserecent, lXI' COl1gl', 1111, Sedill1el1lol. Nice,Abslmcls, 13-19,1975,
Clapperton, e. M" Quatemm)' Cleologr alldGeoll1olphology 0/ SOlItli Alllerica, 779 pp"Elsevier, New York, 1993,
Clemens, S, e., and W, L. Prell, Late Pleistoeene variability of Arabian Sea summcr mon-
soon winds and continental aridity: Eolianrccords !"rolll the lithogcnic cOlllponcnt o!"deep-sea sedilllcnts, hdeocelll1ogmpliy, 5,109-145,1990
Curry,W, 13., Latc Quaternary dcep cireulationin the western equawrial Atlantic, Tlle SoulliAIlalllic: Preselli al1d Pasl ('iI'Clilaliollcdited by Ci Wel'cr et al., pp. 577-598:SprinGer-Verlaa New York 1996
Curry, \V, H, J c', Duplcssy: L. D Labeyrie,allCl N, ,I, Shaekleton, Changes in thedistribution of Ol3C o!" deep water LC02between thc last glaciation and the Holocene,Paleoceallogmpliy, 3, J 17-341, 1988,
Dalllllth,.I. E" Late Quaternary sedimentation inthe western eqllatorial Atlantic, Geol. Soc.;/111, Bull" 88,695-710,1977,
DeMclo, U., C, P, SlIllllllerhayes, allCl J P, Fllis,Salvador to Vitor!a, southcastern 13razil, IV,COIIIl'ib, 5'edillleI1I01., 4,78-116, 1975,
DeMenocal, P., Plio-Pleistocene Ati-iean el imate, Sciel1ce, 270,53-59, 1995 ,
DeMenoeal, P, 13" W, F, Rudcliman, and E M,Pokras, Intluences 01' high- anel low-Iatitudeprocesses on Ati'iean terrestrial climate:Pleistocene eolian records from equatorialAtlantie Ocean drilling program site 663,Poleoceallogmphy, 88, 209-242, 1993,
Deuser, W. G., P, G, Brcwcr, T. D. Jickells, andR, F, Commeau, l3iological control 01' thercmoval 01' abiogenic partieles from the surface oeean, Sciellce, 219, 388-391, 1983,
Diekmann, 1'3., R, Petsehick, 1', X. Gingeie, D,K, Fütterer, A, Abelmann, R, Gersonde, U,Brathauer, and A. Maekensen, Clav mineralIluetuations in late Quaternary sed-iments ofthe southeastern South Atlantie: ImplicationsJor past changes of deepwater advcction, nieSoulh Allolilic: Preselli ami Pasl Cil'Clilolioll,edited by G. Wefer et al., pp. 621-644,Springer-Verlag Ncw York, 1996,
Duplcssy, ,I, c., N, .I. Shaekleton, 11., G. Fairbanks, I... D" Labeyrie, D, Oppo, and N,Kallel, Deep water source variations duringthe last elimatic cycle allCl their impact on theglobal deep water cireulation, Paleoceonogmphy, 3, 343-360, 1988,
Ericson, D, B" and G, Woll in, Pleistoceneclimates and chronology in deep-sea sediments, Sc/enee, 162,1227-1234,1968,
Gingeie, F, X" P .I. Müller, allCl R, 11., Schneider, Orbital forcing of ti'eshwater input in theCongo-Fan area: Clay mineral cvidenee ti'omthe last 200 kyr, Poleogeogr, Paleoclimalol,Poleoecol., 138, 17-26, 1998,
Ciri1'1'in, ,I, ,I" H, Windom, anel E, D, Goldberg,The distribution of elay minerals in theWorld Ocean, Deep Sea Res, Oceanogr,Absll'" 15,433-459, 1968,
Heath, G, R" and N, G, Pisias, A method tor thequantitative estimation of elay minerals in
North Pacific dcep-sea sediments, C/m's C/oJ'AIincl',,27,175-184,1'J79, ,.
Imbrie, J, .I. D, Ilavs, D, G, Martinson, ;\,Mclntyre, anel AC', Mix, The orbital thcorv01' Pleistoccne climatc: Support ('rom a re'vised ehronology of the marine 0 180 reeord,in Milollkovilch ami Clilllflle, Pari I, editcdby ABerger ct al., pp, 269-305, D, Reielel,Norwell, Mass" 1984.
Imbrie .1., et al., On the slructure allCl origin 01'major glaciation cyeles, 2, Thc 100,OOO-yeareyelc, Poleoceollogmphy, 8,699-735, 1993,
.lahn, B" Menardii stratigraphy, in Repol'1 IIlIdPrelill1inorr ReslIlls 01' Meleol' Cmise 29/2MOlllevidc() - Rio dc' JOlleil'o, 15 7.1994 08,08.1994" edited by U, Bleil anel CruiseParticipants, Ber, Fachbereich Geowiss,Ulliv, I3rell/ell, 59, 153, 1994.
,lohnson, D. A" and K. A. I,asmussen, LateCenozoic turbidite anel contourite depositionin the southern Brazil Basin, Mol', Geol., 58,225-262, 1984,
,Iones, G, A" Adveclive transport of clay minerals in thc region of the Rio (jrande Rise,lviaI', Geol., 58,187-212,1984,
,Iones, G. A" and D, A, ,lohnson, DisplacedAntaretie diatoms in Vema Chal111el sediments: Late Pleistoeene/Holocene lluctuations in AABW 1l0w, MIII', Geol, 58, 59-88,1984
Jouzel, .1., C. Lorius, .I, R, Petit, c:. Ritz, M.Stievenard, P. Yiou, N, I. Barkov, V, M,Kotlyakov, and V, Lipenkov, The climatierecord from Antarctic ice now extcnds baekto 220 kyr BI', in LOllg-lenll Climillic VllrillliollS, NATO ASI SeI', , I, vol. 22, edited by ,1.e. Duplessy anel M ,-1'. Spyrielakis, pp. 213237, Springer, New York, 1994,
Lamy, F" D, Hebbeln, anel G, Wcfcr, LatcQuaternary preeessional eyeles of terrigenoussediment input off the northern ehieo, Chile(27,5°S) and paleoclimatie implications,Paleogeogr, Paleoclimalol, Paleoecol., 141,233-251, 1998.
Lomb, N,R" Least-squares ti-equeney analysis 01'unequally spaeeel elata,Asll'Ophys, S'pllce Sci"39,447-462,1976
Maekensen, A" D, K. Fütterer, H, Grobe, and G,Schmiedl, 13enthie foraminiferal assemblagesti'om the eastern South Atlantic Polar Frontregion between 35° and 57°S: Distribution,ccology and tossilization potential, Mal',Micl'Opaleolllol., 22,33-69, 1993,
Mackensen, A" H, Grobe, I-I.-W, I-lubberten, andG. Kul111, 13enthic foraminiferal assemblagesand the ol3C-signal in the Atlantic sector 01'the Southern Oeean, Carboll ()lC!ing ill IheGlacial Ocellll: Conslminls Oll Ihe Ocean '5
Roll' in Global Chllnge, NATO ASI SeI', C.vol. 117, editeel by 11., Zahn et al., pp, 105144, Springer, New York, 1994,
80
Masse, L., l-C. Faugercs, M. 13crnat, 11. Pujos,ancl M.-L. Mezerais, 11 600,000-year recorcl01' IIntarctic Botlom Water activity in1'errecl/i'om sediment textures ami struclures in asediment core /i'om the Southern Brazil Basin, Paleoceal/ogmphy. 9, 1017-1026, 1994.
Mulitza, S., Stratigraphy, in Reporl ami Prelill1i1/(11)' l?eslIlls ol Mcleor Cmise 23/2 I?io deJal/ciro - I?cci/i', 27.2.1993 - 19.3.1993,editecl by U. Bbl, ami Cruise Participants,Ber. Fachbercich Geowiss. Ul1iv.Brell1en, 43,133,1993
Oppo, D. W., ami Y. Rosenthai, Cd/Ca ehangesin a deep Cape 13asin eore over the past730,000 years: Response 01' circumpolarclccpwatcr variability to Northcrn Hemisphere icc sheet mclting'" Paleoceal/ogmphy, 9, 661-675, 1994.
Pastouret, L., 1-1. Chamley, G. Delibrias, l C.Duplcssy, allCl l Thiecle, Latc Quaternaryelimatic changes in western tropical A/i'icadeduced 1'rom dcep-sca sedimentation otlNiger delta, Oceallo/. //cla, 1,217-232, 1978.
Pätzold, l, ami Cruise Partieipants, Berichte underste Ergebnissc über die MClcor-FahrtM 15/2 Rio de .1aneiro - Vitoria, 18.1.1991 7.2.1991, Ber. Fachbereich Geowiss. UI/iv.Brell1cl/, /7,46, 1993.
Pctcrson, R. G., and L. Stramma, Uppcr-Ievelcireulation in the South Atlantie Oeean,ProgI'. Oceal/ogr.. 26, 1-81, 1990.
Petschick, R., G. Kuhn, allC] F. X. Gingeie, Claymineral distribution in surf~lcC sediments 01'the South At1antic: Sourccs, transport andrelation to oceanography, Mal' Geo/., 130,203-229, 1996.
Pokras, E. M., Diatom record 01' late Quatcrnaryclimatie change in the eastern equatorialAtlantic and tropieal Al1'ica, Poleoccol/ogmplly, 2, 273-286, 1987.
Raymo, M. E., W F Ruddiman, N. .I. Shackleton, and D. W. Oppo, I:volution 01' AtlanticPaeitie ü13c: gradients ovcr the last 2.5 ll1.y.,Eonh PIonei. Sci. Lcll., 97,353-368,1990.
Robinson, S, G., Applications 1'01' whole-coremagnetic susceptibility mcasurcmcnts 01'
Chapter 5
dcep-sea sediments, Proe. ()cean /)ril/. Progralll Sci. Res"lls, I 15,737-771, 1990.
Rossignol-Striek, M., Mediterranean Quaternarysapropels: IIn immcdiate response 01' theAfriean monsoon to variation 01' insolation,Poleogeogr. Palcoelill1olo/. Paleoecol., 49,237-263, 1985.
Ruddiman, W.F., M.E. Raymo, D.G. Martinson,13.M Clement, ami .I. Baekman, Plcistoeeneevolution: Northern Hemisphere ice sheetsami North Atlantle oeean, Paleoceooograplly, 4, 353-412, 1989.
Seargle, l[)., Studies in astronomical time seriesanalysis, 11, Statistical aspects 01' speetralanalysis 01' uncvenly spaced data, Asrropllys.J, 263, 835-853, 1982
Scargle, .1.0., Studies in astronomical time scriesanalysis, 111, Fourier translorms, autoeorrclation funetions, and cross-eorrelation functions 01' uncvcnly spaced data, ASlropllys. J,343, 847-887, 1989
Schmiedl, Ci., and A. Mackcnscn, Late Quatcrnary palcoproduetivity ami decp water cireulation in the eastern South Atlantie Oeean:Evidenec li'om bcnthie foraminifera, Poleogcogr. Polcoclill1a!ol. Paleoeco!., 13IJ, 43-80,1997.
Schulz, M., ami K. Statlcgger, Speetrllm: Speetral analysis 01' uncvenly spaced palcoclimatic time scries, COIlljJ. Geosci., 23, 929945,1997.
Shaekleton, N. .1., A. Berger, amI W. R. Peltier,An alternative astronomieal ca libration 01' thelower Pleistoecnc timcscale based on OJ)Psite 677, hans. R. Soc. Edinb"rgll Eorlll Sci.,81,251-261,1990
Teruggi, M. E., The nature ami origin 01' Argentine Locss, J Scdilllcnlo/. Pelm/., 27/3,322332, 1957
von Dobeneek, 1'., and F. Schmiedcl', Using rockmagnctie proxy rccords 1'01' orbital tuning andextenclecl time series analysis into the superami sub-Milankoviteh bands, Proxies in PaIcoccanograplly, cdited by G. Fischer and G.Werer, Springer, New York, in press, 1998.
Waelbroeck, c., .I . .Iouzel, L. L.abcyric, C.Lorius, M. Labraeheric, M. Sticvcnard, andN. l. 13arkov, A cOlllparison o1'the Vostok ieedeuterium record and serics jj'om SOLilhernOecan eore MD 88-770 ()Ver the last two glacial-interglaeial cyeles, Clim. Dyn., 12, 113123,1995
Wcfcr, G., and Cruisc Participants, Rcport am!preliminary results 01' Meleor-eruise M34/3Walvis Bay - Recife, 21.02. - 17.03.19%,Bcr. Fachbereich Geowiss. Univ. Brelllen.79, 168, 1996.
Zachariasse, W, .1., R. R. Schmidt, ane] R. J W.Van Lcuwen, Distribution 01' 1'oraminiferaanel calcareous nannoplankton 111 Qualernarysedimcnts 01' the castern Angola Basin in response to climatic and oceanlc fluelualions,Nelli. J. Sea /(es., 17, 250-275, 1984.
Zarate, M., ami A. Balsi, Late PlcistoceneI-lolocene eolian c1cJlosits 01' the soulhcmBuenos Aires province, Argentiml: 1\ preliminary model, QIIIII. 1111., 17, 15-20, 1')93.
F. X. Gingeie, Baltlc Sea Research Institllte,Secstralle 15, D-I8119 Wal'l1c11lüncle, (lermany.(c-mail: jj·anz.gingele@io-warnemucnde,dc) .
R. Petschick, Fachbereich Geowissenschaften, Univcrsitiit Frankfurt Scnckcnberganlage32-34, D-60325 Frankfurt, Germany.
C. Rühlcmann, F. Schmieder, anel T. V[)II
Dobcneck, Fachbereich Geowissenschaftcn,Universität Bremen, Bibliotheksstrasse, D28334 Bremen, Gcrmany.
(Received April 9, 1998;revised October 23, 1998;aeeepteel Oetober 27, 1998.)
Summary and perspectives
Two closely related main objectives delineate the
framework ofthis thesis:
• Establishing representative age models for secii
ment series from the South At1antic Ocean based
on high-resolution magnetic measurements.
• Extraeting paleoeeanographie and paleoelimatie
information from the elimatically eontrolled
magnetic signals by means of specially adapted
statistical ana1yses.
The rcsults achieved substantiate magnetie cyclo
stratigraphy as a potentially very powerful tool for
detai1ed, accurate age modeling. Proceeding from
magnetostratigraphic or oxygen isotope dating high
resolution ehronostratigraphies were established by
orbital tuning of susceptibi1ity records. In the sub
tropica1 and western tropica1 South At1antic Oeean
as in many other open ocean environments these
signals reileet the varying ratio of climatically mo
du1ated terrigenous and biogenic sediment accumu
1atiol1.
Whi1e the SUSAS reeords are essentially do
minated by carbonate dissolution cycles due to en
hanced iniluence of southem source waters during
glacia1s, the CEARIS series primari1y mitTor sea
level re1ated ehanges in tenigenous input. These
regional depositiona1 settings had to be taken into
account in the tuning procedure as they affect tbe
pbase relations between driving orbital variations
and environmenta1 magnetic response and therefore
tbe definition of suitab1e target signals.
Age contro1 is a crucia1 precondition for every
pa1eoceanographic study. Consequent1y, tbe mag
netic cyc10stratigraphies presented 1ed to severa1
interdisciplinary co-operations. Tbey enab1ed a
temporal analysis ofc1ay mineral input during PIeis
tocene times recorded in a SUSAS core from tbe
western Rio Grande Rise. In this area surfaee
sampIes identify the Rio Doce as a major source of
kaolinite. Cyclic variations of kaolinite/chlorite
ratios in the 41 and 100 kyr bands were shown to
be eoherent with global iee vo1ume. They are
interpreted to originate from increased humidity in
the Soutb American hinterland and enhanced runoff
of the Rio Doce during interg1aeials. Due to 10w
sedimentation rates, the strong preeessional compo
nent which must be expected under these eonditions
as a result oflow-latitude insolation changes is not
preserved in this core. However, for sediments from
the Santos Plateau high-resolution susceptibility
measurements permitted the refinement of an
oxygen isotope stratigraphy and the identification
of a strong 23 kyr periodicity in kaolinite/chlorite
ratios. This finding validates the link between low
1atitude climate changes and kaolinite input.
Extended time series ana1yses ofmagnetic sus
ceptibility focused on climate variability in the fre
quency bands above and be10w the main Mi1anko
vitch frequencies. Higher sedimentation rates a1
lowed a detai1ed investigation ofsub-Mi1ankovitch
phenomena for sediments from the Ceara Rise.
Using techniques ofbandpass filtering, spectra1 and
bispectra1 analysis high frequency pattems could
be related to periodic harmonics and combination
tones of orbital frequencies as wel1 as to millennial
Dansgaard-Oeschger cycles. The direct Iinkage of
tenigenous sedimentation to tropica1 climate mir
rored in enviroill11enta1magnetic data opens promis
ing perspectives for sub-Mi1ankovitch reconstruc
tions in these latitudes.
Magnetic cyclostratigraphy also enabled detailed
age modeling for the low sedimentation rate depo
sits from the oligotrophie 'ocean desert' ofthe sub
tropical South Atlantic yielding comp1ete Pleisto
cene time series. The main, stil1 enigmatic feature
of this interval of Earth's climate history is the
change from variations with aperiod of 41 kyr to
the late Pleistocene 100 kyr ice age cycles in the
course of the mid-Pleistocene climate transition
82 Chapter 6
(MPT). Visualized by evolutionary spectral analysis
the changing fi'equency content ofthe SUSAS cores
reproduce in great detail the Pleistocene climate
evolution known from benthic oxygen isotopes.
Compared to 21 stationary ETP model the mag
netic proxy signals document an alternating ex
change of spectral energy between the 41 and
100 kyr bands. Preceded by the occurrence of 21
single 100 kyr cycle at about 1150 ka, the major
shift towards reduced response to obliquity forcing
occurred at about 650 ka, synchronous with the
onset of 100 kyr cyclicity. These results substantiate
threshold models which explain the late Quaternary
100 kyr cycles by postulating 21 decreasing atmo
spheric pCO? level.
In addition, magnctic mineral concentrations re
corded an important paleoceanographic feature not
reported from oxygen isotopes and hence obviously
not directly related to global ice volume. In the mid
die Pleistocene average susceptibilities are enhanced
by 40%, a consequence ofreduced carbonate depo
sition as magnetite accumulation remained constant.
Temporal analysis of the SUSAS west-east asym
metry permits to define a MPT interim state of in
creased carbonate dissolution, most likely resulting
from enhanced influence of southern source deep
waters. This interval exactly fills the time span bet
ween the first occurrence oflarger glacial ice shields
at about 920 ka and the onset ofthe 100 kyr cyclic
ity.
At the end of the MPT interim state, a terminal
event documented by various unusual lithologies
was observed in several cores. This '530 ka event'
is presumably related to strong changes in the ocean
circulation system. Aglobai character ofthis event
is suggested as it occurs coincidentally with other
unusual climate excursions which hint at extremely
warm and humid climates in Asia and Africa.
These new insights to the nature of the MPT were
essentially drawn from a rock magnetic view on
Quaternary marine sediment series. Magnetic cyclo
stratigraphy and advanced signal analyses were thc
tools to decipher the sediment history ofa key region
for the understanding of the global thermohaline
circulation which, due to very low sedimentation
rates, had been largely excluded from carlier
paleoceanographic studies.
Danksagung
Herrn Professor Dr. Ulrich Bleil danke ich für die Vergabe und Betreuung der Dissertation, sein stetes, anregend kritisches Interesseam Fortgang der Arbeiten und ganz besonders für das mir in schwierigen Zeiten entgegengebrachte Verständnis.
Für die freundliche Übernahme des zweiten Gutachtens danke ichHerrn Professor Dr. Gerold Wefer.
Die intensive Zusammenarbeit mit Dr. Tilo von Dobeneck hat mirgroße Freude bereitet. Gut, daß wir die gleiche 'Wellenlänge' habenund so manches Mal zur richtigen Zeit 'in Phase' gedacht und gehandelt haben! Danke, Tilo!
Für die nette Arbeitsatmosphäre geht ein dickes 'Dankeschön' anmeine Kolleginnen und Kollegen in der Marinen Geophysik, an dieTechniker Liane Brück, Heike Piero und Christian Hilgenfeldt, unsere Mädels Katharina Däumler und Andrea Schmidt, den unvergessenen 'Ehemaligen' Dr. Harald 'Harri' Petennann, Dr. ThomasFrederichs, Dr. Karl Fabian und natürlich an meinen liebenZimmerkollegen Jens Funk. Ein 'Danke' auch an Professor Dr. Volkhard Spieß und seine Arbeitsgruppe für die gute und produktiveNachbarschaft, an Dr. Franz GingeIe für die gute Zusammenarbeitam Rio Grande Rise und an Dr. Torsten Bickert für seine steteDiskussionsbereitschaft und auch dafür, daß er mir zum Teilunveröffentlichte Daten zur Verfügung gestellt hat. Stellvertretendfür all' die anderen netten Leute im Fachbereich Geowissenschaftender Universität Bremen, die das Arbeiten hier im Hause und aufdem Schiff so angenehm gestalten, möchte ich an dieser Stelle der'guten Seele' des SFB's, Gisela Boelen, danken, die wir alle kennenund lieben und die hier insbesondere für die jungen Leute (undzum Glück auch für die nicht mehr ganz so jungen) so viel tut.
Petra, Dirk und Sebastian Rahe haben mir manch' erholsame Stundein der Heide geschenkt und waren immer für mich da, wenn mirder Himmel auf den Kopf zu fallen drohte. Einen ganz lieben Dankdafür!
Nicht zuletzt möchte ich von Herzen den Menschen danken, die inden Jahren meiner Dissertation mit mir zusammen gewohnt undgelebt haben: Eleonora Uliana, Andrea Spies, Andre Janke, BerndLaser und Markus J osten.