Chronology of Fluctuating Sea LevelsSince the Triassic

BILAL U. HAQ, JAN HARDENBOL, PETER R. VAIL*

Advances in sequence stratigraphy and the development of depositional models havehelped explain the origin of genetically related sedimentary packages during sea levelcycles. These concepts have provided the basis for the recopition ofsea level events insubsurface data and in outcrops ofmarine sediments around the world. Knowledge ofthese events has led to a new generation ofMesozoic and Cenozoic global cycle chartsthat chronicle the history ofsea level fctuations during the past 250 million years ingreater detail than was possible from seismic-stratigraphic data alone. An effort hasbeen made to develop a realistic and accurate time scale and widely applicablechronostratigraphy and to integrate depositional sequences documented in publicdomain outcrop sections from various basins with this chronostratigraphic fiame-work A description of this approach and an account of the results, illustrated by sealevel cycle charts of the Cenozoic, Cretaceous, Jurassic, and Triassic intervals, arepresented.

F OR MORE THAN A CENTURY, EARTHscientists have accumulated geologicevidence indicating fluctuations in

the mean sea level during Phanerozoic time.In the early 20th century, Suess (1) re-marked on the apparently synchronous epi-sodes of deposition and nondeposition ofmarine strata in different parts of the worldand suggested that sea level rises and fallsmay be eustatic (global) in origin. Otherresearchers have since documented the sealevel histories of different parts ofthe world,and some of them have ascribed the appar-ent synchroneity of these events to episodesof global tectonics (2, 3).

Sea level fluctuations have important im-plications for organic productivity of theoceans and sediment distribution patternsalong the continental margins and in theinterior basins. Therefore, the study ofthesefluctuations is ofprime interest to hydrocar-bon exploration. Sea level changes are alsothought to control hydrographic-climaticpatterns and, indirectly, biotic distributionpatterns as well. Understanding thesechanges is ofconsiderable value in decipher-ing past oceanographic (paleoceanographic)conditions.Developments in seismic stratigraphy

during the 1960s and 1970s led to therecognition that primary seismic reflectionsparallel stratal surfaces and unconformities(4). On this basis, Vail ea al. (4) proposedthat sediment packages (depositional se-quences) bounded by unconformities andtheir correlative conformities represented

primary units with chronostratigraphic sig-nificance. Vail at al. used stratal geometriesand patterns of onlap, downlap, truncation,and basinward shifts of coastal onlap tointerpret sea level histories along variouscontinental margins. The apparent synchro-neity of sea level falls in widely separatedbasins led them to generate a series ofchartsshowing global cycles of relative changes insea level (4).With the assertion of the method by Vail

at al. that primary seismic reflections repre-sented time lines, seismic stratigraphy wasseen as a breakthrough for regional andglobal chronostratigraphic correlations. Ithas been particularly valuable in frontierareas where it aids in the predrill determina-tion of geological exploration parametersfrom seismic profiles. The original coastalonlap curves (4) were largely based on inter-pretations ofseismic sections with paleonto-logical age control from well data. Sincepublication of the Vail at al. method, sealevel curves have been a subject of livelydebate. The main criticisms of the curveshave centered on (i) the lack of adequatecorrections for local and regional subsidenceand thus the potential error in estimatingthe magnitude of sea level rises and falls (5);(ii) questions about the timing and theglobal synchroneity of some of the majorevents and their significance to the events inthe deep sea (5, 6); (iii) the need for updat-ing the sea level curves in view of the recentrefinements of time scales (6); and (iv) thenonpublication of supporting evidence (6,

7), much of which was considered propri-etary. Since the original publication, moreup-to-date versions of the global coastalonlap curves for the Jurassic and Cenozoichave been published (8), and some of theissues mentioned above have been ad-dressed. However, to reduce the depen-dence on proprietary seismic and well-logdata, the need was seen to develop alterna-tive criteria for identifing sea level fluctua-tions in easily accessible sections, wherelessons learned from seismic interpretationof sea level changes could be applied topublic domain outcrop data. The recentadvances in sequence stratigraphic concepts(9) and the development of depositionalmodels of genetically related sediments dur-ing various phases of the sea level cycles (10,11) have helped fulfill this need (see Fig. 1).The sequence-stratigraphic depositional

models, together with detailed paleontologi-cal data, enhance the ability to recognizegenetically related sediment packages in out-crop sections. They also provide indepen-dent avenues by which seismic and diversesubsurface data can be augmented and inte-grated. Sea level rises and falls are manifest-ed by specific physical surfaces that can beused to identify sequences in land-based andoffshore marine sections. In this way, sealevel changes can be documented in diverseareas that are within the public domain.[Studies listed in (11) cover quantitativemodels, applications in the field, chronostra-tigraphic basis, and the documentation ofthis methodology.] These developmentsrepresent a major step forward since the firstpublication of sea level curves (4).Over the past several years stratigraphers

at Exxon Production Research (EPR) haveattempted to produce a global stratigraphicframework that integrates state-of-the-artmagneto-, chrono-, and biostratigraphieswith sequences recognized in the subsurfaceand outcrop sections in different sedimenta-ry basins. These data have provided a new

The authors are on the scientific research staff of ExxonProduction Research Company, Houston, TX 77252.

*Prcsent address: Department of Geology, Rice Univer-sity, Houston, TX 77251.

SCIENCE, VOL. 235

on

Oct

ober

28,

200

8 w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from

generation of Mesozoic and Cenozoic cyclecharts that go beyond the resolution possi-ble with seismic stratigraphic techniquesalone.

In order to publish the new cyde charts(Figs. 2 to 5) without further delay, we havesummarized our results in this artide. Wedescribe our approach and the results for theMesozoic (the Triassic, Jurassic, and Creta-ceous are dealt with separately here) and theCenozoic. The sequence-stratigraphic con-cepts and depositional models have beenaddressed elsewhere (11).

Chronostratigraphic BasisThe accuracy of a widely applicable corre-

lation framework depends on the reliabilityofthe stratigraphy on which it is based. Our

objective has been to build a stratigraphicframework that is advertent to empiricaldata and is rigorous enough that quickmodifications are not necessary as singularnew items of data become available. Thechoice of the linear time scale provides anexample of this approach.

Traditional methods of constructing timescales, including some of the more recentattempts, have relied heavily on a few radio-metric dates that are used to "nail down"segments of the otherwise exrapolated lin-ear time scale (12). The choice of radiomet-ric dates that are used as 'tie points" oftendepends on an internally justifiable prefer-ence of the researchers. The result is that aseries of different and equally valid timescales can be constructed on the basis of adiffering choice of tie points. Even if thetime scale near the fixed tie points is valid,

the precision ofthe rest ofthe segments willdepend on the accuracy of the assumptionsused to extrapolate the time intervals. Suchassumptions may not always be warranted(13). The tie-point approach is a reasonableoption only when reliable radiometric dataare extremely sparse, as is still the case formuch of the Paleozoic.

Differing time scales can also result froman investigator's preference for a certain typeof radiometric technique. Examples are pro-vided by the differing linear scales basedexclusively on high-temperature radiometricdates, compared with those based largely onlow-temperature dates (14). Although theproblems inherent in various radiometricdating techniques are becoming better un-derstood (15), the adoption of one tech-nique to the exclusion ofothers introduces adistinct bias. It also ignores a large body of

TME

4c

zUl

2(A)

IH0LwLa

CANYON

A) U DEPTHSHALLOW DEPISHALLOW DEEP

Fig. 1. Sequence-strati-graphic concepts. Depo-sitional model showingsystems tracts during thedevdlopment of type 1and type 2 sequencesthat occur aftcer type 1and type 2 unconformi-ties, respectively. (A)The systems tracts in re-lation to depth. (B) Thesame features plottedagainst geologic time(legend below this figureexplains the symbols).

6 MARCH I987

B) N GEOLOGIC TIME

LEGEND

SURFACES

(SB) SEQUENCE BOUNDARIES(SB 1) = TYPE 1(SB 2) = TYPE 2

(DLS) DOWNLAP SURFACES(mfs) = maximum flooding surface(tfs) = top fan surface(tis) = top leveed channel surface

(TS) TRANSGRESSIVE SURFACE(First flooding surface above maximumregression)

SYSTEMS TRACTS

HST = HIGHSTAND SYSTEMS TRACTTST = TRANSGRESSIVE SYSTEMS TRACTLSW = LOWSTAND WEDGE SYSTEMS TRACT

ivf = incised valley filpgc = prograding complexIcc = leveed channel complex

LSF = LOWSTAND FAN SYSTEMS TRACTfc = fan channelsfl =fan lobes

SMW = SHELF MARGIN WEDGE SYSTEMS TRACT

tls Its fc fl

ARTICLES II57

on

Oct

ober

28,

200

8 w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from

potentially valuable analytical and empiricaldata.To produce a practical time scale with the

widest possible use, one must reconcile allreliable observations. Our linear scales, asshown on the Mesozoic and Cenozoic cyclecharts, are best-fit solutions ofthe analytical-ly sound and stratigraphically constrainedradiometric dates (16). We believe a solu-tion that does not overlook any potentiallyuseful chronological information generatestime scales that are more stable and utilitar-ian and that will resist the need for quickmodifications.The magnetostratigraphy (geomagnetic

polarity reversals) adopted for the cyclecharts is a combination of four differenttypes of paleomagnetic information of vary-ing quality. We have adopted a polarity scaleof geomagnetic reversals of the past 6.5million years that was identified in lavaflows and constrained by reliable radiomet-ric data (17). The polarity scale for theinterval from 6.5 to 84 million years ago isbased on stacked mean ages of magneticanomalies from three major ocean basinprofiles, calibrated to a best-fit radiometriclinear time scale (18).We have adopted the Oxfordian through

Barremian polarity scale by calibrating theM-series magnetic anomalies (MO-M29)against a best-fit numerical scale based onavailable radiometric dates for the Jurassicand Cretaceous. For the pre-Oxfordian in-terval, for which no sea floor magneticanomaly data are available, we constructed atentative polarity reversal model based on asynthesis of various paleomagnetic studieson outcrop sections of the Triassic throughCallovian age (19). This model may berevised as new magnetic data for this intervalbecome available.The magnetobiostratigraphic basis for the

Cretaceous through Quaternary interval hasbeen considerably refined in recent yearsthrough direct ties between fossil occurrencedatum events and magnetic polarity reversalevents. All such magnetobiostratigraphicdata available to us have been incorporatedin the cycle charts presented here. Direct

Fig. 2. Cenozoic chronostratigraphy and cydes ofsea level change. The linear time scale (in millionyears before present) is repeated on the lkft,center, and right of the cyde chart. Sectionsinclude magnetostratigraphy, chronostratigraphicsubdivisions, biostratigraphy, and sequence stra-tigraphy. Long- and short-term eustatic curvescomplete the chart. The key at the bottom of thefigure explains the relative magnitude of sequenceboundaries (type 1 and 2 boundaries are distin-guished) and condensed sections. Sources (cita-tions in the figure) for the Cenozoic magnetostra-tigraphy are listed in (17, 18) and for biostratig-raphy in (31). (Two halves of the figure arereproduced on facing pages.)

II58

correlations between reversals and biohori-zons (first and last occurrences of calcareousplankton) in the tropical to temperate re-gions are now available for much of theCenozoic and late Cretaceous (20). For

I

MAGNETO-STRATI-GRAPHY

CHRONO-STRATIGRAPHY -I

parts of the Neogene, such ties are alsoavailable for siliceous plankton (21, 22). Butfor the Jurassic and early Cretaceous, suchfirst-order correlations between polarity re-versal events and fossil occurrences are limit-

BIOSTRATIGRAPHY

E _~!I'-I E i~ zPLANKTONIC NANNOFOSSIL RADIOLARIAN DIATOM DINO-

FORAM BIOCHRONO- BIOCHRONO- BIOCHRONO- FLAGELLATE_ SoE AE BIOCHRONO- ZONES ZONES ZONES BIOHORIZONS

112 1 ! N ZONES )JNORTH PACC- 41 __OCE_ A - t jJ t ;. 4MITDTOLOW

orL- . VLAU M1 HO ZEw _|LAP .eYRlTH^

4

4U

L 1

L 6

7

112

,2

176

n

is

_.

C3

CB

C4

DI

Cl

cm

Cs

1

13

1E

t

cli

EU

CENE

LUzLuUl2

zLu20

0

Lu

00LU

-

W.a.I

ZANCUAN

MESIINIAN

TORTONIAN

--'.'2-Lu

aja

-1

cc

0:

eeu

IL

Luf0.

Luaa

e

eUs

SERRAVALLIAN

BURDIGAUAN

AQUITANIAN

CI4A1TIAN

30RA

RUPELIAN

PRIASONIAN

36.4_

BARTONIAN

42A

LUTETIAN

YPRESIAN

- U-

LUzU0

I4

Lu

-4-S ?Ellu

0 DANIAN

as --

-I MDOR REVERSED POLARITY EVENT

~ NOMAL PoLAIrry

= REEPD POLAITfY

o. TRUNCAT- EDm A IS

--mou_--AI

'Ix|N17

SPOINGASTERPENTAS

PEREGMIA

DODYWIOCYRT1ISPENULTWAA

DOYlUIOCYRBANrlE!MNULTIMUA

D -ARTUSPErc RWOM

_MO.N1G. FOHS4 .1-ROSUSTA E12

FOU ;111SPO 5le

a. IIISUETA

C. DISSMILIS

G. KUGLERI

C.- PUiPs-P.=CR

a. CERRO- l7

INVOLUTA PMO

T. ROHRI MO

G. LEHNERI PI2

G. SUBCONG-|LOaATAf Pl

NENal Pt

.PET- Fp

NE AS Ps

tPTPUOTE

SDLEEUDEDS-P USEUDO-MENARDE P4

IPUSISLL IPAUCIUATAI

M. UNCINATAE n

S. TRULOOB IPMCEN

0. PSEUDO-BULLOIDES one

c"D EXIUS

S HETIERO-

'S IlLEbiNlO

_D DOU-CZ1

_T CARINATUS

c-n S_ CSPEsOEuSI

"PM

Cv PtDSSDTEN"TUS

l T

. iFOREDSEeUS

_- ECRUCP17"

cC,su

SAc rof

DIAN

cr3

c- NPU

C.3A

1liNP12

DORCADOSPYRISALATA

CALOCYCLETTACOSTATA

SmCHORCORYSVVOLWWM

STICHOCORYSDELMONTENSIS

CYRTOCAPSELLATETRAPERA

LYCHNOCANONIAELONGATA

DORCADOSPYRISATEUCHUS

THEOCYRTISTUSEROSA

TMYSOYRTIS_BROMIA .

PODO.CYRTIS

AMPLA

THYRSOCYRT18TR-ACAIHA

DENTICULOPSIS [ESCESucTESHUSTEDTII A UCU

(A

DENTICULOPSISHUSTEOTIl/

DENTICULOPSISLAUTA

(_

DENTICULOPS;ISLAUTA

,.ACTW40CYCLUS

THALASSKOSIRAFRAGA

,,A

THALASSIOSIRASPINOSA

ROCELLAGEUDA

BOGOROMIVENIAPAINI

Cs

ROCELLAVIGILANS

CESTOOtSCUSRETICULATUS

81RADIATA

BRIGHTWELLIAHYPERBOREA

+ CRASPEDODLSCUSr.Si..r...^..BEKOMA UNDULATUS

INDARTENSIS

',,,AFTER STAINFORTH ET AL IN)ITAFER SLOW INN AND ERGOREN 10121

)AFTER BUKRY lUll

(Q)AFTER MARTINI 1i9R(fV AFTER SANFRUPPO. WESTBERG AND RIEDEL (11. 19NSRt;AFrER: GOMRO301N52.1N). GOMBOS AND CIESIELSKI 1163)

FENNER (1RI, ANE BARRON 1055)

" PSfUDOEXfCUfV^TUSR PERfORATU;

TEUgbOH PLECTItU;

t H-OMORPHUSS XOXA POROSUXH tOXoS;A-YVIATUW

E Ufc tCO_05lTUt^RCU^IUKISSE LOVIA SPt t

D COII"P EX

l-IH TUltFfRUIIDIItT8OPLOKUCkOENtSEtOf t S

* SfT^^D PHOSPHORITtC

MUtttllPIIIIOS

ClflCUT U ATAHOO_¢UC S^fPtUW

D OEBISfftDEIII

2

F

0

l

'TLAST OCC _ESE

VERSION 31IOtobw. `111

SCIENCE, VOL. 235

I00 TAAIk_ _-

C.12

19- .-

I -

:)A 6zi MAA;UCMMfifl-I Lu_ i,,

=. '-.___iTN-,5-uwLL--ML--F,

:]

10Oiw . !-

D. OOibcvn

M!. D- MUL-n-RAWATUS

CP"

H

Cft Mmm-

F TYMPANcp. 0,0""Ps

I PAACELLUSC" NP4

CL DANICMC" NM

CV" c Tim"

M. wvanmA

.1..

z

.1 .

j

t

i

.1--.j

II

III

I-

.711

I

I .1I

I

I

i

. -. - -.I M, 1.CW_UM P A I_

I

m

I

ci

r?DENITICULOPSISKMTSCHATICA

7;-

6.MAJUIMITAE

0.ACOSTAENSIS

0. MOAAFM

6

le

Is

20

25

30

n

a

|NTE

.1

- 30

G.CO-EPOENSISCIPEROENSIS

G. OPIMAOPIKIA

PZ2

P2t

- 40

so

a

so

a

-SO

- 56

HEMIAULUSI"CURtVUS

INOT ZONEO}

HYPERIONTlYAN

- a6

m .

mu,-,

C-0 0 aum-- CAUER"usC".m MII

:.= D- CALCAM

CW D- HAktATM

GiawaCRASPFDODISCUS

OBLONGUSBURYELLA CsCLINATA

on

Oct

ober

28,

200

8 w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from

ed to a few isolated studies (23).Much work needs to be done on the

Mesozoic sections before a dear picture ofthe relation between polarity reversal se-

quences and fossil zones can emerge. Until

SEQUENCE STRATIGRAPHY

lSd< IC

N

0

zz

a-

LU

4fDE13Su

io

4

wi

4

n

LbJ

ORDERI

UJ

0

Ic

U0

U

RELATIVE CHANGE OF COASTAL ONLAP

LANDWARtO BASNWARO

1.0 0.5

I-

W 3 g. X ? 3F g

that time, we must assign equal durations tothe subdivisions of such commonly usedzonal schemes as those based on anrmoniteswithin individual stages of much of theJurassic and Lower Cretaceous.

863>.US

lug .4

f7

-21

HS

9

NS

SW-IHS""

EUSTATIC CURVES

Bo no 166 so Omi- -.- - c =-- --7

F

£-

41 36.__65 T _ _--37.S-

4.1 -3B __L-_

3.l _

-41.2- HLS_ _ _ -

3.4 4-=-43-

323\-- -e: HS |

2.9- f---------

-2 = =.5__, , _-4zs - - - - - - - - -_2.9 t54-63-FF

.7 en'.=T'.-46qW.-. .............,,,a_.__ m. . .X.i, i i i . . . _ _ 1TR

1.2 TR t-

._SEQUENCE BOUNDARY TYPES

TYPE 2 x

TYPE I 1 10i1iai6

LOWSTAND FAN

DEPOSITS WITHOUT

COASTAL ONLAP

SYSTEMS TRACTSHS High-wWW Dopod4

TR Trm4U3133 D0o.o6

LSW L.,_.i W_.e

F r- of Kn.ow F..

SMW mmm-m6rgi. Wedge

8.u. HAQ. J. HARDENBOL, P.R. VAILR.C. WRIGHT. L.E. STOVER, G. BAUM,T. LOUTIT. A. GOMBOS. T. DAVIES,C. PFLUM. K. ROMINE. H. POSAMENTIER

6 MARCH I987

By tradition, the European stages of theMesozoic and Cenozoic have come to beaccepted as the basic units of chronostratig-raphy for worldwide correlations. We corre-lated these stages with our linear scale andmagnetobiostratigraphy by establishing thesequence-stratigraphic framework of thestage stratopes and other key referencesections from various parts of the world.The standard stages could then be integratedmore accurately into the cycle chartsthrough the biostratigraphic and physical-stratigraphic relations of the stratotypes. Atypical example of the sequence-stratigraph-ic approach is provided by the ChattianStage of the Upper Oligocene.The neostratotype of the Chattian at Do-

berg bei Bunde in West Germany consists ofabout 70 m of nearshore to offshore marinesands and marls (24). The litho- and biofa-cies, grain-size analysis, and paleoenviron-mental considerations suggest that two dis-tinct depositional sequences separated by anunconformity are present. The lower se-

quence consists of fining-upward and deep-ening transgressive deposits that are overlainby coarser high?stand, nearshore to littoral,deposits. This succession of depositionalpackages is repeated in the upper sequence,

although in a somewhat deeper setting. Thebiostratigraphic data permit a precise corre-lation of the neostratotype to two of theUpper Oligocene sequences (TB1.2 and1.3, Fig. 2) on the cycle chart. Such se-

quence-stratigraphic studies of the stage

stratotpes have helped us to position thestages more accurately within the standardchronostratigraphic framework.

Documentation of Sea LevelChanges Since the TriassicThe new generation of cycle charts of sea

level fluctuations is largely based on thestudy of sequences in outcrops that aug-ments the results from subsurface (seismicand well-log) data. These cycle charts are an

improvement over previously publishedones, which were based entirely on subsur-face information. The documentation of theMesozoic and Cenozoic sequences is derivedfrom outcrop sections in various parts of theworld. Much of this documentation, howev-er, comes from marine outcrops spreadthroughout central and westem Europe, andin the United States, along the Gulf andAtlantic coasts and in the western interior(Colorado, Utah, and Wyoming). All ofthese sections are within the public domainwhere the results can be studied and tested(25-28).

Ideally, sequence-stratigraphic analysis(recognition of depositional sequences relat-

ARTICLES II59

I- JR3S 0s * .2-

3.-3 - -.-3.2 t = 3 _w_ HS _t2M~~~~~~~~~~~~~~~~~~.;...2--/ - L831 -9,__--- 2-116 646

vs--(13.m1 12 HS- .t2.5 '-:z.;;;;;-4-2-,1 113.4- -.Tl

24 -':- : 47_~~~~ ~~ ~~~~~~~ 2. ii - -___ 11.- 7-ff- - = C ---155I!mIIfI3t 125 -

Is

\ :0 TR~~~~7 LONG TEM 8640SHO7 TEM

- LW____I

-~~~-3.5 2 ----'

L.. 14 (~~~~~~~~~~~~~~~~44513.8.2A \ HSr

2 .3 He'2b'4. ____

KEY TO RELATIVE MAGNITUDE

CONODESEOSEC7706

'6 6

*SEOUENCE AGES 6NPREVIOUS CYCLE CHARTS

0

0

6

a

I

i

r-

r.1--7--i

4..-.1f-::-

S-

Uf

IFF-U--31F-V -7777 7*tt.

1.2

*.S

4.4

2 -

.- 7'

---t

IIII

-l

-

-

1-L

id

r-

r-

r

4

01

I

'RI

0i11

4

I

I

I

7 ,A-

.t_

-

="

upi

ki

I

I

I

I

I

III

-

-1

-I

-a

-3

-a

-31

-4a

49

.60

a

so

al

u

0

n

r- I......

ff-4 If fIf, .,+, ",

t.t

33

36

Is

1-4

HS

SEQUENCEBOUNDARY

Mi.W

Modi..

m.iK. .

ITO IDO

:. -+ 7. . ..--'N ..;

I

on

Oct

ober

28,

200

8 w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from

ed to sea level rises and falls) should becarried out on seismic sections and well logsfrom an area, in combination with extensiveoutcrop studies of the same region. Thesedata can be integrated to establish an accu-rate chronostratigraphic framework and toidentify depositional sequsnce boundariesthrough litho- and biofacies analyses. Inpractice, however, all three types of data areusu,ally not available for all areas.Changes of relative sea level, which is the

combined effect of subsidence and eustacy,control the accommodation potential of thesediments and the distribution of facieswithin the system tracts. The details of thesequence-stratigraphic concepts and the cri-teria used to identify depositional systemstracts (Fig. 1) in outcrops have been dealtwith elsewhere (11). In marine outcropsthree depositional surfaces (systems-tractsboundaries) can usually be identified. Themost readily identifiable surface is the"transgressive surface" (Fig. 1), which oc-curs above the lowstand deposits that arecharacterized by sediments of the maximumregressive phase. The transgressive surfaceheralds the onset of a relatively rapid sealevel rise and marked lithologic changesassociated with the rise. In the absence oflowstand deposits, the transgressive surfacemay coincide with the underlying uncon-formable portion of the sequence boundary.The second most easily recognizable sur-

face in outcrops is the surface of maximumflooding, which is referred to as the "down-lap surface" on seismic profiles (Fig. 1). Thissurface is associated with the condensedsection that occurs within the transgressiveand highstand systems tracts. It depicts aninterval of depositional starvation when therapidiy rising sea level moves the sedimentdepocenters landward. Because ofthe lack ofterrigenous input, the condensed sectionmay be expressed as a zone of high pelagicfossil concentration, or as hardgroundcaused by lithification. The relative durationof slow sedimentation represented by thecondensed section increases basinward until,beyond the areas of terrigenous influence,the deposition in the deeper basin is theoret-ically a series of stacked condensed sections.The marked lithologic and faunal changesacross the condensed section can also bemistaken for the sequence boundary.The third recognizable depositional sur-

face in outcrops is the sequence boundary.In seismic sections this is expressed by the

Fig. 3. Cretaceous chronostratigraphy and cyclesof sea level fluctuation (see Fig. 2 and caption forkey and explanation). Sources for Cretaceousmagnetostratigraphy are listed in (18) and forbiostratigraphy in (33). (Two halves of the figureare reproduced on facing pages.)

downward (basinward) shift of coastal on-lap. In outcrops the sequence boundary maybe represented by an obvious unconformityor by more subtle changes, depending on

MAGNETO-STRATI-GRAPHY

CHRONO-STRATIGRAPHY

the position of the section along the shelf-to-basin profile and on the rate of relativesea level fall. For example, if the location ofthe section is more shoreward along the

BIOSTRATIGRAPHY

E . o E PLANKTONIC CALCAREOUS MACROFOSSIL AMMONOID DINO-Z C 0F F N F STAGES FORAM AND NANNOFOSSIL BlOCHRONO- BIOCHRONO- FLAGELLATE Z

0 (AzoO | O0 CALPIONELLID BIOCHRONO- ZONES ZONES BIOHORIZONS a.4E BIOCHRONO- ZONES (BOREAL) (TETHYAN

_ZONES GREAT BRITAIN REGION)65r- _ G-Pwuoo. bDmc=mmHz=7j - 65

70

76

86

Is

too

- Ml

Ml

M1

01

M3

M4

Mml

I

c- ..I

ICM14

CCas

_n0

uI-w0c

wa.

0

DANIAN

MAASTRICH-TIAN

i 74

CAMPANIAN

SANTONIAN

CONIACIAN

TURONIAN

CENOMANIAN

ALBIAN

AnT

APTIAN

SS!E

.ih WURIN A1 T IUSIG. EUGUSlNA "w INVER..

A MAYARONENSIS MC25I5IC.5A MURARG. CONTUSA z5 UAORATUS

G.GNSEI_ ARKHANGEL-GASEI" SKIELLA

Z CVMBIFORMISI

G VENTRICOSA

G. ELEVATA

D. ASYMMETRICA

D. CONCAVATA

D. PRIMITIVA

H. HELVETICA

W.ARCHAEO-CRETACEA

R. CUSHMANI

IR. REICHELIR. BROTZENI

R. APPENNINICAR. TICINENSIS

R. W T;illlI

T. PRIMULA

M. PLANltiIMA

T. BEtJAOUENSIS

IG. GORBACHtKE

I S. CAFRO

I G. BLOWI

I H. SIMIUS

IIBARREMIAN E' H SIGAU

11X.5_.

HAUTERIVIAN

121-

VALANGINIAN

RYAZANIAN

CAM11JURASSICIUI POATLANDA

CM17111

G. HOTERIVICA

|ICALPIONELLIOSCALPIONELLITES

3CALPION - _ELLOPSIS 2

CALPIONELLAELLIPTICA

IC)

CALPIONELLAALPINA

(s)

LITHRAJ'HIDfESACUTUbl

31 030

5"xI

I -D--CSPHAERA

CODLUMNATA

G TRICARINATA z

IG.CALCARATA_

CHI^STOZYGUS IIOWVEBNIU8tbWWS

FISUCOSTATUS

MICRAPNTH-

O TUSUS

LITHfAlPHIDITESDOLU

CRtTAXH"DUUONOEI

CALCICALATH IMAOBLONGATA

__-

N^( OCO" SCOIO"I

s g.

*UDI I JUDD$GESLI"UM GESLINIANUMGUE.tANGERI NAW.CULARE5JUKESOOWNER JUKESSROWNEI

R"OTOblAGE"SE RHOTOMAGENSEDIXONi DIXONI

NK)lEINTERtMEDIUS=DE"fTATUSK"'ETUSE.ETAT..NORNNUE"9ULLIENSIS

AUPiTfO15miLARCHERIPUZOsW4US

FLOIDUM-KrrcHEmiREGULRbilMEIOIDS=FARNHAMIEWSM

'LACUOE''

,Iuvta

. GlllhtoTW

O COSI.I.

P CIIUORIOIOE.

S CO.G-.CATU.;

C C"Ut

UDENTATUMA SEAA- coc cE G-fLUt

RUOEFOI"COSTATUbA PUL AIE O AfflCLCHELLA

RsaOCICTUI

I#VEIISU

1UN#^AMEDIPITREI

DlCHOTOlliTES

POLYPTYCHITS

ALUDUblSTENOMP14LU!_

K*ENRKOCHI

LAAPLUGHIPREPLWCMPHALUS

OPRIESSUS

NO&ZICGSTTA-

SAYMCRUWMSE

#-OOOSDPLU>CTUIbILORYI

I M.T.A.-Sc

TOKI

O AFTER ALLEMAN ET AL. T1971). PtEMOLI-SLVA AND BOLL) 1473). VAN HINTE (12).PFESMOU-SILVA ANO ROERSMA 1137J. REMANE 119721. RORASYNSKI ET AL. I1974.141,AND CARON (ISM)

AFTER: THERSTEIN 119701. SISSINGHE1147 MANIOVIT ET AL. 11771ROTH (1R3.1RO1. AND MONECHI ANO THIERSTEIN (ITMI,

O AFAR: RAWSOM ET AL. 19720. ANR KENNEDY l190.AFTER: VARIOUS AUTHORS IN CAVALIER AND ROGERS T19E).

AMEDRO 1A0.IUR.119RR. ROSASTN%KI ET ALU.1131, AND KENNEDY 11941.

VERSION 3.1IGEOctobE 1rSIM

SCIENCE, VOL. 235

CERATO-LITHO DES&IACULEUS

CAW"SBROONSONIA

PARCACALCULITESoescuFlus

L CIANO-RHABDUSCAYEUXN

REON"AROTITESANTHOP"ORCIS

;Z;;ASTER SFURCATUS

IWFILLIL71.015

9-L

;iII rf

I R.DEOM i .1

UTHEU]§En NUTRELDIENSMSI _ n

_1--1ml R.AlCAWn

I

. .-Fous Li-1 JACOBI GR^ ...ANn' _ _ J_r_I

(A

70 -.

is

so

as

90

95

100

105

110

115

120

125

130

IC291C3D1C31

i

u

r--l.OWEJ

H

TESTUDiNAROUS

SOCIAUS

CORANGUSNUM

fAANTELLi C..f--..PERMFLATA

DISPAR FALLAX3WA

INFLATUM SILENUSLAUTUS SWUCATUS

ANGUSTUS11

3:9

TARDEFURf.ATA_0,;

tGAPFtIFN110

iis

120

126

130

1160

m

mim

on

Oct

ober

28,

200

8 w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from

shelf, the probability of deposition of low-stand deposits is reduced. If located updip,the deposition of such deposits may beentirely precluded, so that the sequence

UZ

- 65

70

Is

s0

85

90

95

100

105

110

1120

120

126

-130

SEQUENCE STRATIGRAPHY

IIST ORDER)

O a

7 E

0

N0

LUz

LU

0

>-C.)1I..

(A 6!

-iCcO) O> 0

rlm

(A

I- 2- 11

N_

N

0r.XLDL

N

IC

0-J

ND4

z

N

0r.

CL

zN

cc

mN

N

XC6

45

4.4

4.3

42

4.1

3.5

34

3.3

3.2

3.1

2.72.6_

2.5

2.4

2.3

2.2

21

1.5

1 4

1.3

12

1.1I

RELATIVE CHANGE OF COAST AL ONLAP a,3ODz<w20

BASINWARD,0

.W

5U22 ~2-.-

1

RIll 412-

boundary is an unconformity that may coin-cide with the transgressive surface. Down-dip, the sequence boundary becomes con-formable and occurs within the overall

EUSTATIC CURVES 4

2

0

1120

B.U. HAGL J. HARDENBOL, P.R. VAIL,L.E. STOVER. R.C. WRIGHT, R. JAN DU CHENE

6 MARCH I987

shoaling-upward sediments. It is typicallycharacterized by a change from interbeddedprogradational deposits to more massiveaggradational deposits.

Similarly, significant falls of sea level aremanifested by prominent unconformitieswith erosional truncation caused by subaeri-al exposure. Type 1 unconformities (Fig. 1)produced as a result of rapid sea level fallsthat are greater than subsidence at the shelfedge may expose the entire shelf. The seawithdrawal below the shelfedge also signalsthe development ofincised valley systems onthe shelf that may be accompanied by low-stand fan (lowstand fan systems tract) depo-sition offshore if a source ofsand is available.The incised river system feeds the fan direct-ly, and the fan deposits therefore do notshow coastal onlap (Fig. 1).As soon as the regional subsidence begins

to outstrip the slowipg rate of sea level fall,relative sea level begins to rise, and backfill-ing of the incised valleys commences. Thelowstand facies (lowstand wedge systemstract) accumulate between the shelf edgeand the fan; these deposits may initiallydevelop a leveed channel complex. Eventual-ly, lowstand deposits prograde over the lev-eed channel complex and the fan deposits, asthe shoreline reaches its maximum basin-ward regression. As the global sea levelbegins to rise, the transgression of the shelfis marked by the transgressive surface, andthe landward back-stepping transgressive fa-cies (transgressive systems tract) begin to bedeposited. The transgressive deposits are inturn overlain by the prograding highstanddeposits (highstand systems tract) duringthe highstand phase. Short-term, higher fre-quency flooding events occur in all systemstracts and have been termed parasequencesor pacs (29) (Fig. 1).When the rate of sea level fall is slow, the

withdrawal ofthe sea is more deliberate, andthe whole shelf may not be exposed. Theresulting unconformity is less prominent(type 2 unconformity). In this case thelowstand fan and the leveed-channel depos-its do not develop. Instead, the shelfmarginfacies (shelf margin wedge systems tract)prograde directly over the shelf edge andonto the slope (Fig. 1).The application of sequence-stratigraphic

concepts (11) to outcrop sections has pro-vided the framework to identify and classifymajor, medium, and minor sequences. Inpractice, only sequences of major and medi-um magnitude are discernible at the regionalseismic level. Minor sequences are generallybeyond the resolution obtainable with seis-mic data alone, but they can be mapped bydetailed well-log studies and in outcropsections.We list here the only major areas and

ARTICLES II6I

LANDWARD1.0 1250 ;f0 lS0 1e 50 MA

10

10I7zD

esr

41

4.1

3.6

3.4

131

2A

2.4

2.3

1

2.2

I.q-

1.5

on

Oct

ober

28,

200

8 w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from

sections that were used for our documenta-tion. Study areas for the Cenozoic (25),Cretaceous (26), Jurassic (27), and Triassic(28) are listed separately.

Results and Description of theCycle ChartsA series of four cycle charts depict the

chronology of sea level fluctuations in theCenozoic, Cretaceous, Jurassic, and Triassicand cover the eustatic history of the last 250million years (Figs. 2 through 5). On eachcycle chart the linear time scale in millionyears before present is repeated in threeplaces. The cycle charts incorporate a vastcombination of information that helps tochronicle the sea level changes in a precisemanner. This information is presented infive sections. The first section ismagnetostratigraphy. Both the sea floormagnetic anomaly data, where available, andpolarity reversal sequence are included (17-19).The second section includes chron-

ostratigraphy. The columns contain the sys-tem, series, and stage designations (forthe Cenozoic, series and stage only). Wehave adopted the commonly used westemEuropean stages that have come to be ac-cepted as standard intemational chronostra-tigraphic units. Some of the suprastagechronostratigraphic designations (such asLias, Dogger, and Malm for the Jurassic,Fig. 4) that are still frequently used in someregions are also indicated within the stagecolumn (30).The third section depicts biostratigraphy

and includes two types of information. For-mally defined zonal schemes (based on firstor last occurrence of fossil taxa or their totalranges) are included. They differ, however,for the Cenozoic (31), Cretaceous (32, 33),Jurassic (34), and Triassic (35), dependingon the relative usefulness of the groups forsubdividing the particular stratigraphic in-terval. The second type of information inthese columns is the first and last occurrenceevents, or biohorizons, of those fossilgroups whose zonal schemes have not yetbeen formally established or become widelyaccepted (for example, the dinoflagellatebiohorizons for the Cenozoic through the

Fig. 4. Jurassic chronostratigraphy and cycles ofsea level fluctuation (see Fig. 2 and caption forkey and explanation). Sources for Jurassic magne-tostratigraphy are listed in (18, 19), and forbiostratigraphy in (34). The synthesized pre-Cal-lovian (older than 160 million years) magneticpolarity reversals model (shown in gray andwhite) may be subject to modification whenadditional magnetic data become available. (Twohalves of the figure are reproduced on facingpages.)

II62

MAGNETO-STRATI-GRAPHY

CHRONO-STRATIGRAPHY BIOSTRATIGRAPHY

3i AND ONXN*KWO- SKX:*~~~OHRONZONES OHMOMZONSX~~~~~~~~~SAE PRE| A 3A Y ZONS AN SU ZDN IN.W. EWM

ZONES] AN"AVONS INW.EROEPASMPOMAM A- COPE FT AL (EEXIAS. E _AGELLAT

rul aMN acm _m ImIm _. _,I __

M1

MIN

mm

K=

h=

.=

C""4

CM

amlCMIS

CM"

1-

I-Icma

cma

ii

= E

_I

0

HI.

zs.<

CRETA-CEOUS

(a)

Cl)

U,

LUCE

Lu

a

Lu

0:

I

0-

VS

L

XN I.RAM II

0.POR -

LANDN

KIMM-

XFORDIN

UM.LOVAN

-Wi--

ATHONIA

SAOCUAN

TOARAXAN

PUENS-

BACAN

no

1

NEAN'l

EUTANSON

®mucrn USAGE

I

I

It

P..~

P. KOWAIL oe

IS"

NEP&re

D,n

x

roPrAir,

w

imr.oP. OLA,

'mwmIL SnAK

or, vAa re

T.,

tjT. .

r

Of.,

. m

SASU A'IUM

D. HCATE"

3EE

__ __i(KC

LAP.

-J

-) ii

w!_,MAU!mAD

00oVSKwMEILA

POLYPOOOMNmA4ALA

HEXUM

t _ _EUOOXUSMIUTA

ROmmRA"ut

mmrn

TEMUMCATUMCOEATUM

coLAM

CALLOV W

L..s

PPOIRAClEU%WTIPAMUNSOM =WWWW-LD-.-_U ---

WODOMAINIL

0AawO

CONCVUM

k 0IHUM

0FLOPA

OWESU

T_MCSTTUM

on

JAAVOB

P. D _ A"_AN_ISUIUS . (i)RAlCOGTIfAT_

v_NGAI SPHAERA OXYNOTUM

DUNAOA .TUB

P2-A3lS.SIMCKSTATUM00

PAHAODOLfmWS AEGULATA

DUc Ca

VERSION 3.1IOcobtw, UN)

UAMICtS

PLNVI

.. im

0.840"IKA lm

VWlgv.

FALL.

VALCW.""M,w

.ALDA.

Ao&".

14 AMMOMIE 2z0_-_ E_AMD

I PUDOCOcDATA

CAUDIMMES

5 PUCATLZ

.c

a

w"IC"__^T 0

IX

FT

T-

TZ

In

10

- T

_

I

zI =Kwgill 22151 WTUBOMPHALUS

,,,-k3 L

cools"

In U, ONdPLFXL

mpj.P. SP.A

ulWulP. orrm. sro4vAVLnMMl I= -MIAmkr--

Mann" &-----JIMMMUML-

::L- -n-.-.

I LAEVRNNX" - L--.r_ _r- I

I

I

- 7

I

_

a

6

7

a

9

0

2

3

Is

IL

a

7

a

n

In

T-----Ms

*.

I..MAU.ay

is

or

_

ml

_GOANEO

on

Oct

ober

28,

200

8 w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from

Jurassic, and the radiolarians for Jurassic andTriassic). Much of this information shouldbe regarded as preliminary until future workconfirms the correlation ofthese events withthe global chronostratigraphic framework.The dinoflagellate occurrences, in particular,largely from western European sections,represent an aggregate of the data of EPRpalynologists (36) and will need to be cor-roborated elsewhere before they can be ap-plied on a wider basis.The fourth section contains the terminol-

ogy for sequence stratigraphy. It includessequence chronozones or cycles (megacy-des, supercycles, and cycles) (37) and scaledrelative changes of coastal onlap. The ages ofthe cycle boundaries and downlap surfacesare indicated in separate columns, as are thedepositional systems tracts (boundarieswhere fans have been observed are indicatedby "F" in this column, Figs. 2 through 4).Major, medium, and minor sequenceboundaries and condensed sections are iden-tified by the relative thickness of the linesdrawn through them (38). The unshadedtriangles within each coastal onlap cyclerepresent the condensed sections, depictingthe intervals of slow deposition after rapidsea level rise, the relative duration of whichincreases basinward.The long- and short-term eustatic curves

are plotted in the last section. The scale (inmeters) represents the best estimate of sealevel rises and falls compared with the pre-sent-day mondial mean sea level (39).

In the long term, the generally low sealevels of the late Paleozoic (Pennsylvanianand Permian), which reached their lowestpoint in the Tatarian, continued into theTriassic and early Jurassic. In the Hettan-gian, the sea level dipped to another mark-edly low position; the levels remained gener-ally low through much of the middle Juras-sic, rising somewhat in the Bajocian, butfalling again in the late Bathonian. Thetrend reversed itself in the Callovian, and thelong-term sea level continued to risethrough the Oxfordian, reaching a Jurassicpeak in the Kimmeridgian.

After a transient but marked decline in theearly Valanginian, the sea level began to riserapidly, remaining high through the remain-der of the Cretaceous. It reached its Meso-zoic-Cenozoic peak in the early Turoniantime. After this mid-Cretaceous high, agradual decline of sea level began in thelatest Cretaceous and continued through theCenozoic. With the exception of relativelyhigher levels in the Danian, Ypresian, Rupe-lian, Langhian through early Serravallian,and Zancian, this trend toward lower sealevels continues to the present time.

In the short term, the major sea level fallsoccurred at the base of Portlandian, in early

ARTICLES II63

on

Oct

ober

28,

200

8 w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from

Fig. 5. Triassic chronostratigraphy and cycles ofsea level change (see Fig. 2 and caption for keyand explanation). Sources for the synthesizedTriassic and early-middle Jurassic magnetc polari-ty reversal model are listed in (19). This modelmay be subject to future modifications. Sourcesfor Triassic biostratigraphy are in (35). (Twohalves of the figure are reproduced on facingpages.)

Aptian, mid-Cenomanian, late Turonian,late Maastrichtian, early Thanetian, latestYpresian, latest Bartonian, near the Rupe-lian-Chattian boundary, in Burdigalian-Langhian, in the late Serravallian, andthroughout the late Pliocene-Pleistocene in-terval. These short-term, but marked, sealevel falls are frequently associated withworldwide major unconformities. At leastsince the Oligocene, sea level drops may bein large part due to the increasing influenceofglaciation. This influence is manifested bythe relatively large variations in amplitude ofthe short-term sea levels since the mid-Oligocene.A total of 119 Triassic through Quatema-

ry sea level cycles have been identified in thenew generation Mesozoic-Cenozoic cydecharts. Of these, 19 began with major se-quence boundaries, 42 began with relativelymedium magnitude sequence boundaries,and 58 were represented by minor sequenceboundaries. As mentioned earlier, only thesequence boundaries of major and mediummagnitude can be identified generally atregional seismic level. Detailed well-log oroutcrop studies are usually necessary to re-solve the minor sequences.

CondusionsWe have described our approach in

chronicling the Mesozoic and Cenozoic his-tory of sea level fluctuations from variousparts ofthe world. Our objective has been tomake the cycle charts public in the mostexpedient manner possible. In this artide wehave not attempted to address the importantissues of the causes of sea level change, theabsolute magnitude ofthe sea level rises andfalls through time, the implication of thesechanges for the continental margin anddeep-sea sedimentary budgets, or their influ-ence on hydrography, climate, and bioticdistribution and evolution.A cursory comparison, for example, could

not establish a dear relation between ocean-ic sedimentation rates (40) and sea levelfluctuations. A comparison of Neogene sealevel cycles with known intervals of wide-spread gaps in deep-sea sedimentation (41)reveals, however, that these gaps are coinci-dent either with the downlap surfaces (con-densed sections) ofmajor and medium mag-

II64

MAGNETO- CHRONO-BITRTGAHSTRATi- STRATIGRAPHY lIOTR lTGAHRAPHY

- - oNNON ThMASSI TRIASSIC- SIOCMRONO~~~~~~~~~~~~~~~*ONCH4RONO AMMONITE AMMONIE PALYNO0- F

ZONES ZONES SICHNONO- sIOONO- OrSTAE n M-0 3OW ZONE ZONES SiOHORIOXNS 2

F~~~~~~~~~~~~~~~~~~~~~~A T1-|1N IILXORTs^H s -n_ _ I NllESAOETAL ATrl_ _____ninum ~~~~mOSfuR AND TOZER 11SM

II ~ ~ ~ ~ ~ ~ ~ i~~~ ANDMAID _ _ _ _

__ E.M TIDNTATA COIP RRAUIsN Ul

2U_21' 1 215M. COUJUSIAMPJS M. MACDR

2X ._ H. HOIRARE"I CT I_ ___TI u u- 2

E. N TATA 0. T. oUII C. INC RENATUI -

NORIU _AN W. PAUClM JlAG_C

7M It zlA 230_STRITA M. DAWORI M. PAULCE*I

_R 8USmRu E. AT0EPTENOV16 M515KE 0. JANOSHAUS

12~~ ~ ~ ~ KMACRWRJATU ANATROPITI 4hi

2_ P. POLY_RATHI-006 T. WELLON

CARNM ~~~~~~~~T.DiAEm T. SUBIJILLATUR

S ~ C' _____T_71 .SUT 7 euor

iCARNINZ TZ T_ OT. UM A. AUSTRIACUMT.DESATOYENER T. AONONERzn

E. L~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~IUNMUNRRI .M.MACLEARRIP HLU

IiSERIALSLA"2A. VAUE 26

-_^F X ~S UCRNT

- _ _ _~~~~~~~~N MNS EN PPOSEIDAON _ __GIS2

TATARSAN~ ~ ~ ~ ~ ~ ~~A.VAR

- UI9 -

IMLL OTHERS

VERSION 3.1lOctober, iSIS

nitudes (which represent periods of maxi-mum flooding of the shelves) or with se-quence boundaries (which represent sealevel drops), depending on whether thehiatuses are caused by carbonate dissolutionor by erosion and removal of sedimentsfrom the sea floor. A recent study has shownthat in the central equatorial Pacific, themajor breaks in Neogene sedimentation cor-respond to regionally correlatable and syn-chronous seismic reflectors (42). Whencompared to our sea level cycles, these re-flectors also correspond to condensed sec-tions ofmajor anS medium magnitude. The,reflectors are caused by carbonate dissolu-tion or diagenesis and are related to changesin the ocean chemistry (42). Obviously,there may be a cogent connection betweensea level fluctuations and shifts in the qualityofdeep ocean water. We suggest the follow-ing scenario to explain this correspondencebetween the two opposite phases (highstand

and lowstand) of the sea level cycle anddeep-sea hiatuses.During the highstand, after a prominent

sea level rise, the terrigenous sediments aretrapped on the inner shelf, starving the outershelf and slope. The sequestering of carbon-ate on the inner shelf may lead to reduceddissolved carbonate in seawater (43), andthe resulting rise in calcite compensationdepth (CCD) would lead to increased disso-lution in the deeper parts of the basins. Thisreduction in carbonate during highstandwould explain the correlation of dissolutionhiatuses with condensed sections (times ofmaximum flooding of the shelves). The sealevel elevation would also lead to climaticequitability and the weakening oflatitudinalthermal gradients (44), which in turn wouldresult in reduced current activity both at thesurface and on the sea bottom. After amarked sea level fall, on the other hand, theinner shelf is bypassed, and sediments are

SCIENCE, VOL. 235

on

Oct

ober

28,

200

8 w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from

directly transported to the outer shelf orslope. The resulting increase in carbonatecontent of the seawater and the lowering ofCCD would reduce carbonate dissolution.But the climatic inequitability and strength-ened thermal gradients during the lowstand(44) would lead to intensified circulationand increased bottom water activity, causingwidespread erosion on the sea floor. Thisprocess explains the correspondence of theerosional hiatuses to sequence boundaries.

REFERENCES AND NOTES1. E. Suess, The Face of the Earth (Clarendon, Oxford,- 1906), vol. 2, p. 535.

2. H. Stille, Grundfragen der vergkieihnden Tektonik(Bomtraeger, Berlin, 1924); A. W. Grabau, OsciUa-tions or Pulsations (16th International GcologicalCongress, Washington, DC, 1933); L. L. Sloss,24th Int. Geol. Congr. 6, 24 (1972).

3. L. L. Sloss, Gcol. Soc. Am. Bull. 74, 93 (1963).4. P. R. Vail ct a.,Am. Assoc. Pet. Geol. Mem. 26,49-

212 (1977).

6 MARCH I987

5. W. C. Pitman III, Geol. Soc. Am. Bull. 89, 1389(1978); G. Bond, Gcology 6, 247 (1978); A. B.Watts and M. S. Steckler, in Deep SeaDrilling in theAtlantic Ocean: Continental Margins and Pateenvir-pument, M. Taiwani, W. Hay, W. B. F. Ryan, Eds.(American Geophysical Union, Washington, DC,1979), Maurice Ewing Series, vol. 3, pp. 218-239;K B. Watts, Nature (London) 297, 469 (1982); A.W. Bally, Am. Geophys. Union Geodynamis Scr. 1, 5(1980).

6. N.-A. Momer, Geoloy 9, 344 (1981); N. Parkinsonand C. P. Summerhayes, Am. Assoc. Pet. Geol. Bull.69, 685 (1985); A. D. Miall, ibid. 70, 131 (1986);B. E. Tucholke, Soc. Econ. Pakontol.-Mineral. Spec.PuhI. 32, 23 (1981).

7. A. Hallam, Annu. Rev. Earth Planet. Sci. 12, 205(1984); A. D. Miall, in (6).

8. P. R. Vail and J. Hardenbol, Oceanus 22, 71(1979); P. R. Vail and R. G. Todd, in PetmrlumGogy of the Continental Shelf of Northwest Europe,L. V. Illing and G. D. Hobson, Eds. (Institute ofPetroleum Geology, London, 1981), pp. 216-235;P. R. Vail, J. Hardenbol, R. G. Todd, Am. Assoc.Pet. Geol. Mem. -36, 129 (1984).

9. Sequence is a widely used term in earth science, buthere sequence refers specifically to the depositionalsequence or the succession of sediments depositedduring a complete sea level cycle, that is, from a sealevel fall to subsequent rise and ending with the nextfall (Fig. 1). Sequence stratigraphy is broadly de-fined as the branch of stratigraphy that deals withdepositonal sequences of genetically related strata

deposited during the different phases (lowstand,transgressive, and highstand) of sea level cydes [seeR. M. Mitchum, P. R. Vail, J. B. Sangree,Am.Assoc.Pet. GeolMem. 26, 117 (1977)].

10. Various authors in C. Wigus et al., Eds., "Sea levelchange-an integrated approach," Soc. Econ. Paeon-tot. Mineral. Spec. Pubi., in press.

11. In addition to the references in (8), sequence-stratigraphic concepts, methodology, and clrono-stratigraphic basis have also been presented in thefollowing: J. Hardenbol at., 2ndInt. Conuf Pako-CeaW . (abstr.) (1985), p. 40; B. U. Haq etal., ibid.(abstr.), p. 40; T. S. Loutit et a., ibid. (abstr.), p.52; T. C. Moore at al., ibi. (abstr.), p. 55; B. U.Ha ctaal, Soc. Econ. Pan. M ral. (abstr.)(19 6), p. 121; Geol. Soc. Am. Abstr. Prorams(1986), p. 628. Details of quantitative models andtheorctical concepts ofsequence stratigraphy appearin M. T. Jervey, in (10); H. Posamentier, M. T.Jervey, P. R. Vail, ibid.; J. van Wagoner, ibid.Examples of applications of these concepts in thefield are included in P. R. Vail et al., Bul. Soc. Geot.Franc, in press; S. M. Greenlec, F. W. Schroeder,P. R. Vail, in North Amrican Continental Margins,R. Sheridan and J. Grow, Eds. (commemorativevolumes, Decade of North American GeologyBoulder, CO), in press; G. R. Baum, in (10); J.F.Sarg, ibid.; T. Loutit, J. Hardenbol, P. R. Vail, ibid.Details of the chronostratigraphic basis appear in B.U. Haq, J. Hardenbol, P. R. Vail. ibid.

12. For example, W. A. Bcrggren, Lethaia 5, 195(1972); J. Hardenbol and . A. Bergren, Am.Assoc. Pet. Gedt. Spec. Studies 6, 213 (1978); W. A.Berggren, D. V. Kent, J. J. Flynn, J. A. VanCouvering, Geol. Soc. Am. Bull. 96, 1418 (1985);W. B. Harland at al., A Geolic Time Scak (Cam-bridge Univ. Press, Cambridge, 1982).

13. Time scales are commonly extrapolated betweenfixed points, assuming constant sedimentation or seafloor spreading rates.

14. Compare, for example, the Paleogene time scalebased largely on low-temperature radiornetric dates(K/Ar dates on glauconites) [G. S. Odin t al., inNumerical Dating in Stratgrphy, G. S. Odin, Ed.(Wiley-Interscience, New York, 1982), part 2, pp.957-960], with one based exclusively on selectedhigh-temperature dates (K/Ar dates on bentonites)[W. A. Berggren, D. V. Kent, J. J. Flynn, J. A. VanCouvering, Geol. Soc. Am. Bull. 96, 1407 (1985)].

15. Various authors in G. S. Odin, Ed., NumerialDating in Strajraphy (Wiley-Interscience, NewYork, 1982), part 1, pp. 151-454.

16. In constructing the linear time scale, we used bothhigh- and low-temperature radiometric dates. Crite-ria for the selection of radiometric dates have beenthat they be analytically acceptable and stratifraphi-caily constrained. When reliable low- and high-temperature dates are available for the same strati-graphic interval (for example, the Aptian throughMaastrichtian), the older ranges ofthe low-tempera-ture dates show significant overlap with the youngerranges of the high-temperature ates.

17. I. McDougall, N. D. Watkins, G. P. Walker, L.Kristjansson, J. Geophys. Res. 81, 1501 (1976); I.McDougall, K. Saemundson, H. Johannesson, N.A. Watkins, L. Kristjansson, Geol. Soc. Am. Bull. 88,1 (1977); E. A. Mankinen and G. B. Dahrymple,J. Geophys. Res. 84, 615 (1979). (If not alreadyconverted, we have converted all radioinetric datesaccording to new decay constants.)

18. Composite late Cretaceous-Cenozoic polarity rever-sal scales have been developed from the marinemagnetic anomaly scale of J. R. Heirtzler et al. [.Geoph'ys. Res. 73, 2119 (1968)], with later refine-ments by J. L. LaBrecque, D. V. Kent, and S. C.Cande [Geoly 5, 330 (1977)] and W. Lowrie andW. Alvarez [ibid. 9, 392 (1981)]. The Callovian(late Jurassic) through Cretaceous polarity reversalscales have developed from magnetic anomaly pro-files ofR. L. Larson and W. C. Pitman III [Geot. Soc.Am. Bull. 83, 3645 (1972)], R. L. Larson andT. W.C. Hilde [. Geopys. Rcs. 80, 2586 (1975)], and S.C. Cande, R. L. Larson, and J. L. LaBrecque [EarthPlanct. Sci. Lett. 41, 434 (1978)]. To arrive at thelate Cretaceous-Cenozoic stacked mean ages for themagnetic anomalies (anomalies 34 through 1), weused the following sea floor anomaly profiles fromthe North Pacific: W. C. Pitman III, E. M. Herron,J. R. Heirtzler,J. Geopys. Rns. 73, 2969 (1968); forthe South Pacific: ibid., p. 2069; and for the SouthAtlantic: G. 0. Dickson, W. C. Pitman III, J. R.Heirtzler, ibid., p. 2987. Magnetic chron nomencla-ture (polarity chronozones) for the late Cretaceous-Cenozoic is largely by L. Tauxe a al. [Palaeogeogr.

ARTICLES II65

on

Oct

ober

28,

200

8 w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from

Paadima. PaPaecd. 42, 65 (1983)]. For thelate Jurassc-early Cretcs (magnetic anomaliesM26 dtrou«g Ml), the foilowingprofies were usedftomhe Niorth Pacific: T. W. W=d, N. Isczaki, J.M. W2 Ama,. Qq*. Uxion Geob. AMir.19, (1976); and from thewesta NorthAZ-tic: H. Schouten and K. M. Kltgord, Earth Planet.Sci. Lea. 59, 255 (1982). Magnetc pohrity chrono-zones for the Aptian-Callovian inteaval are afer J. LLaBrecquet at. [Palaege.gr Platdmao Pa-laed. 42, 91 (1983)1.

19, Pre-Caovian (Triassic throu midd JurssK)synthtic polarity reversal mdel is based on thefollwing: C. E. Helsley, Gel. Soc. Am. RxU. 80,2431 (1969); P. J. TaAm A=o. Pet. God. Bull.54, 1130 (1970); M. W. Mcliny and P. J.Burek, Natr (Lo,go) 232, 98 (191); K. M.Creer, ibid. 233, 545 (1971); D. M. Pechersky andA. N. Khrmov, ibid. 244, 499 (1973); C. E.Helsky and M. B. Steine, Gcot. Soc. Am. Bug. 85,457 (1974); E. Marton, P. Marton, F. Hdekr,EartkPla. Sa. Ltt. 48, 218 (1980); J. E. T. Channell,J. G. Ogg, W. Lowrie, Phiks. T7w. R. Soc. LondonSer. A 306, 137 (1982); F. Hi6ner and F. Heller,Geopbys.J. R. Amn. Soc. 73, 705 (1983).

20. Recent studies that havc prvded direct aorlationbetwccn the lte Cretaceous-C&nozoi fossil occur-rences and magnetic polarity reversals icude W. B.F. Ryan t al., Riv. [tat. Palte . Strw*r. 80, 631(1974); W. Alvare a atl., Gcd. Sac. Am. Bxull. 88,367 (1977); B. U. Haq, W. A. Brgren, J. AVanCouvering, Natwre (London) 269, 483 (1977); H.IL Thierstcin at a., Ge4y 5, 400 (1977); B. U.Haq et al., ibid. 8, 427 (1980); J. E. T. ChanniladF. Medizza, Earth Plant. Sci. Lett. 55, 419 (1981)-W. Lowrie at al., Ged. Soc. Am. Bull. 93, 414(1982); R Z. Poore et a., Gecdyy 10, 508 (1982);H. Stadncr and F. Allram, Ixte. Rep. Dep Sea Drill.Pr*. 66, 589 (1982); W. A. Brggren eta ., ibid.72, 675 (1983); G. Napolonct al., Ged. Sc. Am.Bul. 94, 181 (1983); R. Z. Pooreetal.,Paee rPaaoc aol. Pataoewl. 42, 127 (198 K . Hset at., Ged. Sac. Am. Bull. 95, 863 (1984); W. A.Bergren,D. V. Kene, J. J. Flynn, in Te Cbtrw4y~fekGalogicl Recet, N. J, Snellng Ed. (Man.Geol. Soc. London/Blackwell, Oxford, 1985), vol.10, p. 141; S. Monechi and H. R. Thiekrsen,Mar.Maropateontot 9, 419 (1985); C. E. Birton and J.Bloemendal,Ia Rep. Deep SeaDril. Pr,j. 90, 1294(1986); W. H. Lohana, ibid., p. 776.

21. H. Bolli, J. B. Saunders, K. h-Ne , Eds.,Plankton St (Cambridge Univ. Press,Cambridge, 1985).

22. L. H. Burck, Pro 2nd Work. Group Meet. It.Gol. Coard. Pro., Proj. 114, Baung 1977, Spec.Pubi. Ged. Res. De. Cen. 1, 25 (1978); F. Thecyr,C. Y. Mato, S. R. Hammond, Mar. Micr nt.3, 377 (1978); L. H. Burckle and J. Traincr,

7l v 25, 281 (1979); J. A. Barron, in(21), pp. 763 .

23. Dirct ties between Jurassic and Cretaceous fossilocurrences and magnetic polarity reversals are pro-vided by the followng stdies: W. Alvarez et at.,Garl. Sac. Am. Bull. -88, 367 (1977); J. E. T.Channcll, W. Lowne, F. Mediza, Erth Plane. Sa.Let. 42, 153 (1979); W. Lowrie a a., J. Gephys.les. 85, 3597 (1980); E. Marton, P. Marton, F.Heller, Earth Plant. Sc. Let. 48, 218 (1980); E.Marton, ibid. 57, 182 (1982); J. E. T. Channeil etat., in (19); F. Horner and F. HcIer,J. R. Aron.Soc. 73, 705 (1983); J. G. Ogg, Inst. Rep. Deep SeaDrill. Proj. 76, 685 (1983); B. Galbrun and L.Rasphis, C.R. Acad. Sc. 298, 219 (1984); T. J.B wer, Mr.Micfpaens., in press

24. H.-J. Anderson et al., Field Guide OlcO l Exar-swon, 1969 (Marburg Uniersity, Marburg, 1969), P.47, figue 8; Gion Gal. (Ser. 2), 37,69 (1971).

25. To document the Cenozoic sequences, oupshave boen studied in bothtnohwe Eure-an basins (Anglo-Bclgium and Pars basins , soutenFrance, West German, and Italy) and along theU.S. Atlantic and Gulfa. In Europe the Palo-gene sequeces have been documented in southerEngIfand the Isle of Wight. This documentationinld ithe reference secton for the Thanetian andde stratype section of the Baronian (in Bartonarea). Thec Belgian sectons include stratoypes ofYpresian, Ledian, Tongian, and Rupelian Sages. Anorthwestern Ge n s onnar Dobe providedthe documentation for the Chattuans s. IFrance, the Pans basin area has been of partuarsignificance, and sections studid incudec stratotypes or hypostratopes ofthe Sparnacian, Cuisaan,Lutetian,Auversian, Marinesian, and Stampian

1166

Stages. The sections in southe France indude thetrtypsOfAqiainandBidgaan(ohn

the Aq bain).In to P anstratoty!Qn norther Italy, the cxensive litea-turconNeogme s andhy inIwand Spa was invaliable in veig addt dx

Neogene seq cs. The U.S. sections include Pa-~ene outcrops along the Atlantic cokst in North

and a the Gulf coa inAbbNm and Gorgia. In addkion, Paoge se-quncn have also been stdied in outcrps on Southi ew Zcaland, and along the fnks of theOttway Mountains in Victoria, Austaa.

26. Docimentation for tie Cretaco sequees ismainly provided by sections in western anid centrlEuropean basins, the U.S. Gulf coast, and thewestm inteior. In Europe, the ower Cr eooussequenccs have boen saudied in the Suisse Romandearea of Switzehand and southeastern France. TheSwiss outcrops inclde the s psections oftheValannian and Hauterivian Stages. The Frenchlocalities include lower Creaceous section in theAreche, Yauchue, and Alpes dc Haute-Provcnceareas, induding the stratoype orhyp osection ofthe Beriasn, Aptian, Va -

Barrenian, and B lian p G -cooustcrop studics were undcrtaken in sectons inthe Alpes ar nd the Drofne areas of thePr ne. In aon, the stratotpes of the Cerio-manian (near Le Man) and Turonian (in the Tou-rane) and the muid-Cretcous outcrops ofdte Bou-onnais were also studied. In dte United States, themiddle and Upper Cretaceous sections have bensludied in the westem interior and centrl Texas.

27. The bulk of documentation for the Jurassic sc-quences came from sections in western Europe,particularly the minor sequences that have beendocumented only in England and France so far. TheEnglish sections incde outcrops in Dorst andSomerset in southern England and Yorkshire innoriern England. ntally a complte p -sion of Jurassic sences can be seen along theDorst coast: the ttangian dtoug Toarcan se-quences can be observed Pinhay Bay eastwardto Etdlifi and the Aalenian through Calovianbetween Burton Bradstock and Tidmoor Point; theUpper Jurassic and lowcst Cretaceous sequencesare exposed fiom the Isk of Porland to Swanage.In France, seauences of H throu eAaknian age, icluding the stratoqpc oarcian(near Thouars), have becn studied. In additon,sections in southrm Germany provided verificationof Rhaetian thrugh Toarcian sequences, whichincludc the s type section of the PlicnsbachianStage (near Pliensbach). Secdons niear thc Montsal-vens area of Switzerland have furnished fiutherconfirmation of late Jurassic sequences.

28. Triassic documentation was provided by sctions inSvalbard, on the Arctic iand of Bjorn0ya, in theDobomites in Italy, and from the Salt Rangc inPakistan.

29. J. van Wagonr, in (10); P. W. Godwin and E. J.Anderson,J. Geot. 93, 515 (1985).

30. The Jurassc-Crcetaos boundary is placed betweenthe Portandian and Ryazanian Stagcs, folwgthecomnmn British usage. Alternative boundary place-ment (for cxample, by French stratigraphers) isbetween the rthonian and Berasian Stages.Names of the commonly used suprastge subdivi-sions inluded in the stage column (that is, Zcch-stein, Buntandstein, Muschelkalk, Lias, Dogger,Mahm, Neocomian, and Senonian) are includedronlyfor convenice, and no fornal status is suggtd.

31. The Cenozoic biostratigraphy (Fig, 2) includespbnraminiferal zones as descn-bt in the

fi)l : W. Blow, . st Int. Cf Pxkt.Micrfik, 1967, Gca 1, 199 (1969); W. A.Berggren in (12); R. M. Staififth, Univ. Kax.Pale . Conib. 62, 1 (1975). Cakarous nanno-fossil zones are dscribed in E. Martini, Proc. 2ndPlano. Ceuf, 1970, Rome 2, 739 (1971); H.Okada and D. Bukay, Mar.Mipeontd. 5, 321(1980); D. Bukry, Soc. Eon. Pakontol.MAneul. SPC.PuH. 32,335 (1981) Radioria zones are citc inA. Sanfillipo, M. J. Wetberg-Smith, W. RK Riedel,Ie. Rep.Dep Sea Dri. Pro. 61, 495 (1981); (21),p. 631. Diatom zones ar from A. M. Gombos,acilri 5, 225 (1982); _ and P. Ciesielski,

I*nt. Rep. Deep Sea Dril. Proj. 71, 583 (1983); A.M. Gombos, ibid. 73, 495 (1984); J. Fenner, in(21), p. 713; J. A. Barron, ibid., p. 763.

32. F. Robasynski et at., Rev. Miopakont. 26, 145(1984).

33. Crcus bios=tigraphy (Fig. 3) includes: Plank-tonc foraminifra zones from I. Premoli-Silva andH.. Bolli, I%t R¢. Deep Sea Dril. Pej. 15, 499(1973)- I. Prcmfi-Silva and A. Boerama, ibid. 39,615 (1977); J. E. van Hinte, Am. Asoc. Pet. Gad.Bul. 60, 498 (1976); F. Robasynaski et i., Cab.Miaopleotl 1, 185 (1979); ibid. 2, 181 (1979);M. Caron, in (21), p. 17. Calpioreffid zones from F.Allman t al., Pro. 2nd Plakton. Cnfi, 1970,Rome, 2, 1337 (1971); J. Remane, in Introduaionto ieM , B. U. Haq atd A.Boersma, Eds. (E , Ncw York, 1978), p. 161;(21), p. 555, Cakarous nannofossil zones fr H.Thierstein, Ma.Mipeot. 1, 325 (1976); W.Sissingh, G.*M bouw 56, 37 (1977); H. Manivita t ., K. Nd Akad. Wet. B80, 169 (1977); P. H.Roth, Insi. Rip. Dee Sea Dril. Proj. 44, 731(1978); ibid. 76, 587 (1983). Boreal (British) am-monoid-beemnnoid zones from P. F. Rawson t al.,Gal. Soc. Lodm Spc. Rep. 9, 1 (1978); W. J.Kenndy,; Bul. Gel. Soc. Den. 33, 147 (1984).Tethyan, ammonoid zones fiom C. Cavalier and J.Roges, Eds., Bur. Rub. God. MA . Mem. 109, 1(1980); R. Bunardo, ibid. 125, 293 (1984); F.Amedro, Rev-. Mirpalewnd. 22, 195 (1980); Cre-taceousRes. 2, 261 (1981); Bul. Sac. Gad. Norman-dieAxssMa. Ham 71, 17 (1984); B. Clavel et a.,Edokae Gaol. Hde. 79, 319 (1986); W. J. Kcnnedy,cited earlier in this r (32).

34. Jurassk biostratigraphy section (Fig. 4) incdudesradiolarian evnts and zrelimar zones in: E. A.Pessagno,Mcpal o23, 56 (1977);and C. D. Bkom, ibid. 26, 225 (1980); P. 0.Baumgarte,Ini. Rep. Deep SeaDril. Proj. 76, 569(1983); E. A. Pessagno t al., Cushmn Foud.Foaminiferal Res. Spec. Pubt., in press. Cacaresnannofossil zones in T. Barnard and W. W. Hay,Eckegae Gal. Hdv. 67, 563 (1974); A. W. Medd,Mar.Micral. 7, 73 (1982); G. B. Ramilin A SL=VksA AfC sNdriSrn*nical Indec o~fCakarow NannofiesitA. R. Lord, Ed. (Horwood, Chicheter, England,1982), p. 17; P. H. Roth, A. W. Mcdd, D. K.Watkins, Ibit. Rep. Dcep Sea Dril. Proj. 76, 573(1983). Boreal (British) ammonoid zones in J. C.W. Cope t al., Gal. Soe. Lmndo Spe. Rep. 14, 73(1980); ibid. 15i 109 (1980).

35. Triassic biosatigraphy (Fig. 5) indludes radiolarlanzones in C. DE. bome, Bull. Am. Palee. 85, 5(1984). Conodont zones in L. C. Mosher,J. Pakton-to. 44, 737 (1970); W. C. Swcet atl., Gal. Soc.Am. Mcm. 127, 441 (1971); T. Matua, ih TheTeds, HerPadr Pabfi Paeoi o oc, K. Nakazawa andJ.M.Dickens, Eds. (Tokai Univ. Press, Tokyo, 1985), p.157. Thc Tethyan ammonoid zones in E. T. Tozer,Gad. Sun. Can. Misc. Re. 35, 8 (1984). NorthAmerian ammonoid zones after N. J. Silbering andE. T. Tozer, Geol. Soc.Am. Spe. Pap. 110,7 (1968).

36. The yde charts in e (mosdy dino-C22) biohorizonstathvebc synd=izedand compiled by EPR palynologis (L. E. Stovcr,M. Millioud, N. S. Ioannides, R. Jan du Chene, Y.Y. Chen, J. D. Shanc, and B. E. Morgan).

37. The cycle terminology adopted here (that is, Absaro-ka, Zuni, Teias, and thcir subdivisions) is based onthe sequence temiinology suggested by L L. Sloss,in (3).

38. The dashed lines drawn through the condensedsectons represent the surfaces ofmaximum Hooding(downlap surfaces on sismic profiles). Howeverthe relative thikness of th represents therelaive nitude ofthe condensed sections associ-ated with these surfaces.

39. Themtude term sea level variations onthe curves is estimated, as in the mtdhod describedby J. Hardenbol, P. R. Vail, and J. Ferrer [Ocanet.Aca 4 (suppl.), 31 (1981)], with a high value (inCampanian) adopted fron the estimate of C. G. A.Harrion [in Sea Leve Choie, R. Revele, Ed.(National Rearch Council, Washington, DC, inpress)]. The magnitude of short-term sea kvel risesand fills has been estimated from seismic and se-quence-stratigraphic data.

40. T. A. Davics, W. W. Hay, J. R. Southam, T. R.Worskey, Scince 197,53(1977); T. R. Worslcy andT. A. Davies, ibid. 203,455 (1979). _

41. J.A. Barronand G.Keller, Geolog 10, 577 (1982);G. Keller and J. A. Barron, Get. Soc. Am. BxU. 94,590 (1983); Geo, in prcss.

42. L. A. Mayer, T. H. Shipey, E. L. Winterer, Scienc233, 761 (1986).

43. W. H. Berger and E. L. Wmtcrer, Spec. PubI. Int.Assoc. Sedimtex. 1, 11 (1974).

44. B. U. Haq, MAr. Gad. 15. M25 (1973).

SCIENCE, VOL. 235

on

Oct

ober

28,

200

8 w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from

45. The Mesozoic and Cenozoic cyde charts presentedin Figs. 2 through 5 are the product of input andinterest ofmany colleagues, both inside and outsideEPR. The principal responsibility for the chronocu-static framework, however, rested with the authorsof this artide. Input for the individual cyck charts

varied; the collaborators are listed at the bottom ofeach cyde chart. We are grateful to all of theseparticipants and to many other colleaes for theirImportant input, without which s Inthesiswould have been far less detailed. We thank R. G.Todd and J. M. Widmier for their support for an

accurate global stratigraphic-eustatic framework andfor stimulating discussions on the subject. Thecharts were drafted by D. Thornton. We thankExxon Production Research Company for givingz uspernission to release the cyck charts and to pubishthis article.

Fertility Policy in China: Future Options

SuSAN GREENHALGH AND JOHN BONGAARTS

A wide range of social, economic, and demographiccriteria are used to evaluate China's present one-childpolicy and five alternative fertility policies that mightguide China's population control efforts until the end ofthe century when the one-child policy is scheduled to beabandoned. These criteria include the policies' macrode-mographic impact on total population size and popula-tion aging; their microdcmographic effects on the family'sability to support the elderly, its economic capabilities,and the position ofwomen; and their cultural acceptabil-ity to the majority Han Chinese population. The resultssuggest that the least desirable strategy is to retain thepresent policy; all the two-child alternatives performbetter than the current one-child policy in achieving thepolicy goals considered.

S INCE CHINA FIRST ADOPTED STRONG BIRTH CONTROL POLI-cies in the early 1970s, there has been a dramatic fall inChinese fertility. Under the wan xi shao policy (literally "late,

sparse, and few," a policy calling for later childbearing, longerspacing, and fewer children) in effect during the 1970s, the totalfertility rate, the most widely used fertility measure, dropped from5.93 births per woman in 1970 to 2.66 births in 1979 (1). The one-child policy introduced in 1979 has pushed fertility even lower: by1984 the total fertility rate had dropped to 1.94, slightly below thelevel required for population replacement (2).

In part because of the policies' demographic success and in partbecause of the political problems that stemmed from efforts toshrink family size very rapidly, in 1984 and 1985 China's leaderstook steps to relax birth-planning policy. An important shift inpolicy direction occurred in April 1984 with the issuance ofCentralDocument 7 by the Party's Central Committee. (Central Committeedocuments on various issues are numbered from 1 each year.) Underthis document the conditions under which couples may have twochildren were expanded and reforms were called for in policy, workstyle, organization, and ideology that were designed to increasevoluntary participation through better meshing of the policy withthe needs of the people (3). Further relaxation occurred in early1985, when the Central Committee changed its target for the year2000 from 1.2 billion to about 1.2 billion-an apparently smallchange, but one that indicates a notable increase in flexibility on thecritical issue of population goals.

6 MARCH I987

In May 1986 concern that changes in the age structure would putupward pressure on the birthrate during the coming decade wasreflected in Document 13, which supersedes Document 7 as theguiding statement on fertility policy (4). Although many of themoderate elements introduced by Document 7 are continued in thisdirective, its central message appears to be that cadres must takestronger measures to ensure that birth targets are met during theseventh 5-year plan period (1986 to 1990). In late 1986 mountingevidence that fertility was rising after several years of decline ledPremier Zhao Ziyang to advocate in an early December speech thata renewed emphasis be placed on the one-child limitation (5).Despite these indications that the policy is becoming more restric-tive again, the question of which elements from the more relaxedphase can be maintained in the late 1980s and the 1990s withoutjeopardizing achievement of the century-end target appears unre-solved at the political center. The optimal mix of policy elements isalso the subject of a lively debate among scholars.From speeches and articles that have been published in the past

few years, it is clear that the range of factors considered inpopulation policy-making continues to widen. The sparse evidenceavailable from the 1970s suggests that one macrodemographicconsideration-total population size-dominated the decision toinitiate the one-child policy in 1979. Although population quantity,combined with population quality, remain the central consider-ations in the mid-1980s, several new concerns have also been raisedin academic and official circles: the cultural acceptability of fertilitypolicy, its effects on the physical safety offemales, and its impact onthe rate of population aging (6).Were these new considerations to be brought directly to bear on

fertility policy, it is not clear what shape that policy would take. Notwishing to abandon the one-child policy for both substantive andpolitical reasons, China's leaders have tended to incorporate newfactors only by tinkering with the preexisting policy. With thebenefits of hindsight and the perspective of outsiders, we use thesenew criteria and other factors to evaluate the present policy and fivealternative policies that might guide China's population controlefforts until the end of the century, when the one-child policy isscheduled to be abandoned. We consider the macrodemographicimpact of these hypothetical policies on total population size andpopulation aging; the microdemographic effects on the family'sability to support the elderly, its economic capabilities, and thesocioeconomic position of women (two factors Chinese leaders do

The authors are associate and senior associate, respectively, of the Centcr for PolicyStudies, The Population Council, 1 Dag Hammarskjold Plaza, New York, NY 10017.

ARTICLES II67

on

Oct

ober

28,

200

8 w

ww

.sci

ence

mag

.org

Dow

nloa

ded

from