Long-distance correlation between tectonic-controlled

Long-distance correlation between tectonic-controlled, isolatedcarbonate platforms by cyclostratigraphy and sequencestratigraphy in the Devonian of South China

DAIZHAO CHEN* , MAURICE E. TUCKER , MAOSHENG JIANG* and JINGQUAN ZHU**Institute of Geology and Geophysics, Chinese Academy of Sciences, PO Box 9825, Beijing 100029,P.R. China (E-mail: [email protected]) Department of Geological Sciences, University of Durham, Durham DH1 3LE, UK(E-mail: [email protected])

ABSTRACT2

During the early Middle Devonian in South China, an extensive carbonate

platform was broken up through extension to create a complex pattern of

platforms, and interplatform basins. In Givetian and Frasnian carbonate

successions, ®ve depositional facies, including peritidal, restricted shallow

subtidal, semi-restricted subtidal, intermediate subtidal and deep subtidal

facies, and 18 lithofacies units are recognized from measured sections on three

isolated platforms. These deposits are arranged into metre-scale, upward-

shallowing peritidal and subtidal cycles. Nine third-order sequences are

identi®ed from changes in cycle stacking patterns, vertical facies changes and

the stratigraphic distribution of subaerial exposure indicators. These sequences

mostly consist of a lower transgressive part and an upper regressive part.

Transgressive packages are dominated by thicker-than-average subtidal cycles,

and regressive packages by thinner-than-average peritidal cycles. Sequence

boundaries are transitional zones composed of stacked, high-frequency,

thinner-than-average cycles with upward-increasing intensity of subaerial

exposure, rather than individual, laterally traceable surfaces. These sequences

can be further grouped into catch-up and keep-up sequence sets from the long-

term (second-order) changes in accommodation and vertical facies changes.

Catch-up sequences are characterized by relatively thick cycle packages with a

high percentage of intermediate to shallow subtidal facies, and even deep

subtidal facies locally within some individual sequences, recording long-term

accommodation gain. Keep-up sequences are characterized by relatively

thin cycle packages with a high percentage of peritidal facies within

sequences, recording long-term accommodation loss. Correlation of long-term

accommodation changes expressed by Fischer plots reveals that during the late

Givetian to early Frasnian increased accommodation loss on platforms

coincided with increased accommodation gain in interplatform basins. This

suggests that movement on faults resulted in the relative uplift of platforms and

subsidence of interplatform basins. In the early Frasnian, extensive siliceous

deposits in most interplatform basins and megabreccias at basin margins

correspond to exposure disconformities on platforms.

Keywords Cyclostratigraphy, Frasnian, Givetian, high-frequency cyclicity,

isolated platforms, platform carbonate correlation, South China.

Sedimentology (2001) 48, 57±78

Ó 2001 International Association of Sedimentologists 57

INTRODUCTION

Metre-scale, upward-shallowing cycles (or para-sequences) are the basic building units of thickshallow-water carbonate successions throughoutthe Phanerozoic and are commonly organizedinto relatively large-scale depositional sequences.Thick shallow-water carbonate successions arecharacterized by a hierarchy of stratigraphiccyclicity (e.g. Goldhammer et al., 1990, 1993;Elrick & Read, 1991; Osleger & Read, 1991;MontanÄez & Read, 1992; MontanÄez & Osleger,1993; Elrick, 1995; Balog et al., 1997; Strasser &HillgaÈrtner, 1998).

Many studies have documented Devonian plat-form carbonate cycles and sequences from aroundthe world (e.g. Read, 1973; Wong & Oldershaw,1980; Goodwin & Anderson, 1985; Dorobek, 1991;McLean & Mountjoy, 1994; Elrick, 1995, 1996;Yang et al., 1995; Garland et al., 1996; Lamaskin& Elrick, 1997). Sequence development and thehistory of relative sea-level change in the Devo-nian strata of South China have been documented(Chen & Chen, 1994a,b; 1995; Shen et al., 1994;Du et al., 1996; Gong et al., 1997), but detailedstudies of the high-frequency cyclicity of platformcarbonates have not been carried out. There isstill much dif®culty in correlating the platformstrata because of insuf®cient biostratigraphiccontrol.

This paper aims to: (1) describe and interpretdepositional environments of upper Middle andUpper Devonian carbonates across different plat-forms in Guangxi and Hunan provinces, SouthChina; (2) demonstrate different cycle types andtheir possible links to eustatic sea-level changes;(3) illustrate the cycle stacking patterns of indi-vidual stratigraphic sections, and integrate thesewith sedimentological, stratigraphical and diagen-etic features to identify and correlate depositionalsequences across the study area; and (4) documentthe interaction between tectonics and eustasy,and determine their control on cycle and sequencedevelopment, and platform evolution.

PALAEOGEOGRAPHY ANDSTRATIGRAPHY

As a result of the `Guangxi' (Caledonian) Orogenyat the end of the Silurian, marine waters recededfrom much of the Yangtze Plate. In South China,the sea only persisted from Silurian into Devo-nian times in southern Guangxi along a narrowbelt named the Qinfang Trough (Wu et al., 1987;

Zhong et al., 1992). The greatly enlarged land-masses formed by the orogeny created a complextopography, which exerted a fundamental in¯u-ence on the subsequent Devonian depositionalsystems, in conjunction with the reactivation ofbasement fault zones.

From the beginning of the Devonian, marinewaters gradually transgressed from the southwestto the northeast, reaching their farthest extent inthe Frasnian. In the early Givetian, the carbonatedepositional area expanded considerably, andclastics retrograded to a very narrow area closeto the uplands. From late Givetian to earlyFrasnian times, the carbonate platforms under-went intense fragmentation, and two major sets ofinterplatform basin were formed, intersecting incentral Guangxi (Fig. 1). The interplatform basinsin western Guangxi were dominantly orientedNW±SE, and show a grid-like pattern (Chen &Zeng, 1990; Zhang & Zheng, 1990; Zeng et al.,1992, 1995; Liu et al., 1993; Liu, 1998). Bycontrast, those in eastern Guangxi and Hunanmainly striked in a NNE±SSW direction, andshow some features of a strike-slip system (Shenet al., 1987; Jiang, 1990; Zeng et al., 1992; Liuet al., 1993; Liu, 1998). The formation of thisDevonian platform-basin system in South Chinawas probably related to a transform fault extend-ing into the continental margin during the open-ing of Palaeo-Tethys (Chen & Zeng, 1990; Zenget al., 1992, 1995; Liu et al., 1993; Liu, 1998).

This study will focus on the platform carbon-ates of Givetian and Frasnian age; namely theMintang and Rongxian (in part) formations in theLitang area of south-central Guangxi, the Tangji-awan (locally with Mintang), Guilin and Dongcun(in part) formations in the Guilin area, and theQiziqiao Formation in central Hunan (Fig. 2).These formations are conformable on lower Mid-dle Devonian carbonates in south-central Guangxiand terrigenous clastics in northern Guangxi andHunan Provinces, and are overlain by Fammeniancarbonates. On the basis of biostratigraphy, theseformations approximately correspond to theGivetian and Frasnian (Tan et al., 1987; Wu et al.,1987; Zhong et al., 1992; BGMR of Hunan, 1997);the correlation between the formations is given inFig. 2.

DEPOSITIONAL FACIES

Five major outcrops were examined on threeisolated platforms in Hunan and Guangxi Prov-inces: Ma'anshan (MA) and Xizhaikou (XZ) on

58 D. Chen et al.

Ó 2001 International Association of Sedimentologists, Sedimentology, 48, 57±78

the Guibei-Xiangzhong (GX) platform; Tangjia-wan (TJ) and Fuhe (FH) on the Guilin (GL)platform; Litang (LT) on the Guizhong-Litang(GZL) platform (see Fig. 1 for location). Continu-ous measured sections were obtained for eachoutcrop. Eighteen lithofacies types were identi-®ed in the Givetian and Frasnian, and these arearranged into ®ve major depositional facies. Inorder of increasing relative water depth they are:peritidal, restricted-shallow subtidal, semi-restricted subtidal, intermediate subtidal, anddeep subtidal/basinal facies (Table 1). In theplatform carbonate successions, high-frequencymetre-scale, upward-shallowing peritidal andsubtidal cycles (fourth- to ®fth-order parase-quences) can be identi®ed by the vertical faciesarrangements. These small-scale cycles can befurther arranged into larger-scale depositionalsequences (tens to hundreds of metres thick).

Fig. 2. Lithostratigraphic units of the Middle andUpper Devonian platform carbonate successions inGuangxi and Hunan Provinces. LC, Lower Carbonifer-ous. Stipple indicates terrigenous clastic formationsbelow the studied interval.

Fig. 1. Palaeogeographic setting ofSouth China in the Frasnian. Thearea was characterized by a complexpattern of platforms and interplat-form basins created by intense syn-sedimentary block-faulting inapproximately NNE±SSW andNW±SE directions. Localities: MA,Ma'anshan; XZ, Xizhaikou; TJ,Tangjiawan; FH, Fuhe; LT, Litang.Platforms: GX, Guibei-XiangzhongPlatform; GL, Guilin Platform; GZL,Guizhong-Litang Platform.

Correlation between carbonate platforms in the Devonian of South China1 59


Table

1.

Su

mm

ary

of

dep

osi

tion

al

facie

sof

Giv

eti

an

an

dF

rasn

ian

carb

on

ate

stra

ta,

Sou

thC

hin

a.

Lit

hofa

cie

su

nit

+th

ickn

ess

Lit

holo

gy

Sed

imen

tary

stru

ctu

res/

textu

res

Bio

taIn

terp

reta

tion

Occu

rren

ce

Per

itid

al

faci

esP

lan

ar

lam

init

e(0

á05±0á4

m)

Alt

ern

ati

ng

Ms/

pelo

idal

Ps

(or

Ws)

,or

dolo

-M

s/d

olo

silt

ite

Mil

lim

etr

e-s

cale

pla

nar

or

smooth

(cri

nkly

locall

y)

lam

inae;

¯at

intr

acla

sts;

desi

ccati

on

cra

cks,

dis

solu

tion

cavit

ies

locall

y

Rare

wit

hcalc

isp

here

sU

pp

er

inte

rtid

al

tosu

pra

tid

al

Low

er

part

of

Qiz

iqia

oF

m,

Ron

gxia

nF

m,

rare

inG

uil

inF

m

Wavy

lam

init

e(0

á05±2á5

m)

Alt

ern

ati

ng

pelo

idal

Ps

(or

Ws)

/(cry

pt-

)m

icro

bia

lite

;in

terg

row

n(c

ryp

t-)

mic

robia

lite

or

cry

ptm

icro

bia

lite

/m

icro

bia

lite

Cen

tim

etr

e-s

cale

un

du

lato

ryla

min

ae,

dom

al

lam

inae

or

LL

Hst

rom

ato

lite

sin

terc

ala

ted

locall

y;

fen

est

rae

locall

y;

inh

om

ogen

eou

sd

olo

mit

izti

on

Am

ph

ipora

,ost

racod

s,gast

rop

od

s;E

pip

hyto

nor

An

gulo

cell

ula

ria,

Roth

ple

tzel

la,

Urs

osc

up

ulu

s,calc

isp

here

s

Lam

inati

on

most

likely

mic

robia

lori

gin

wit

habu

nd

an

tcyan

o-

bacte

ria;

shall

ow

subti

dal

toin

tert

idal

Min

or

inT

an

gji

aw

an

Fm

an

dequ

ivale

nts

;abu

nd

an

tin

oth

er

fms

Fen

est

ral

lim

est

on

e(>

0á1

m)

Fen

est

ral

pelo

idal

Ms

toP

s,ra

reG

sL

am

inoid

,ir

regu

lar

an

dtu

bu

lar

fen

est

rae

com

mon

,sp

heri

cal

on

es

rare

;geop

eta

lfa

bri

cs

com

mon

;w

eak

lam

inati

on

inla

min

oid

fen

est

ral

zon

es

Ost

racod

s,gast

rop

od

s,calc

isp

here

s,A

mp

hi-

pora

com

mon

;m

inor

para

thu

ram

min

idfo

ram

inif

era

Sh

all

ow

subti

dal

toin

tert

idal

Don

gcu

nF

m,

Up

per

Mb

of

Gu

ilin

Fm

,u

pp

erm

ost

Qiz

iqia

oF

m

Ped

ogen

ic/m

ete

ori

calt

ere

du

nit

(0á0

2±0á4

m)

Ori

gin

all

ysu

bti

dal

thro

ugh

sup

rati

dal

lith

olo

gy;

ped

ogen

iccla

yst

on

e/s

ilts

ton

e

Ru

bble

or

cin

der-

like

fabri

cs,

®tt

ed

arg

illa

ceou

sse

am

s;la

min

ate

dcru

sts;

red

den

ing

locall

y;

fabri

cs

rela

ted

tod

isso

luti

on

(cavit

ies/

pip

es,

nep

tun

ian

dykes

or

cra

cks,

mic

rokars

tic

reli

ef)

;ra

realv

eola

rse

pta

lte

xtu

res;

vad

ose

cem

en

ts

Barr

en

;ra

rero

ot

mou

lds

Su

baeri

all

yexp

ose

dti

dal

¯at

thro

ugh

shall

ow

subti

dal

facie

s

Lim

ited

hori

zon

sin

Up

per

Mb

of

Gu

ilin

Fm

,R

on

gxia

nan

dQ

iziq

iao

fms

Res

tric

ted

shall

ow

subti

dal

faci

esS

trom

ato

lite

Bs

(0á2

±0á5

m)

Alt

ern

ati

ng

pelo

idal

Ps

(or

Ws)

/cry

ptm

icro

bia

lm

icri

tes;

inte

rgro

wn

(cry

pt-

)mic

robia

lite

s;u

suall

yin

terc

ala

ted

inw

avy

lam

init

es

Late

ral-

lin

ked

hem

isp

heri

c(L

LH

)la

min

ae

(10±35

cm

wid

e,

5~1

0cm

tall

)com

mon

,st

acked

hem

isp

heri

c(S

H)

(or

dig

itate

,colu

mn

ar)

rare

Ost

racod

san

dgast

rop

od

s,ra

reA

mp

hip

ora

Rest

ricte

d,

shall

ow

subti

dal

tolo

wer

inte

rtid

al

Ron

gxia

nF

m,

Low

er

Mb

of

Gu

ilin

Fm

60 D. Chen et al.


Pelo

idal

Gs/

Ps

(0á2

±1á5

m)

Pelo

idal

Ws

thro

ugh

Gs

wit

hm

inor

bio

cla

stic

san

dsp

ars

e(l

um

py)

intr

acla

sts

Fen

est

rae

locall

y;

rare

weak

lam

inati

on

Ost

racod

s,fo

ram

inif

era

,calc

isp

here

scom

mon

,sp

ars

eA

mp

hip

ora

an

dbra

ch

iop

od

s

Very

shall

ow

subti

dal,

inte

rtid

al

shoals

or

tid

al

cre

eks

All

fms

excep

tT

an

gji

aw

an

Fm

Ooli

tic

Gs

(0á2

±1á0

m)

Ooli

tic

Gs/

Ps

wit

hsu

bord

inate

pelo

ids

an

dbio

cla

sts

Rare

gra

ded

gra

ins;

rad

ial

an

dsu

per®

cia

looid

sd

om

inan

t,re

lati

vely

well

-sort

ed

;m

icro

fen

est

ral

fabri

cs

Sp

ars

eost

racod

san

dcalc

isp

here

sV

ery

shall

ow

subti

dal

toin

tert

idal

cre

eks

Rare

inQ

iziq

iao

an

dR

on

gxia

nfm

s

Am

ph

ipora

Fs/

Ps

(0á2

±3á5

m)

Am

ph

ipora

Fs

(or

Ws)

/P

san

dG

slo

call

yB

roken

Am

ph

ipora

,ra

nd

om

lyali

gn

ed

top

ara

llel-

pre

ferr

ed

ori

en

tati

on

;lo

call

yin

terc

ala

ted

inm

icro

bia

lla

min

ites

Am

ph

ipora

dom

inan

t,sp

ars

eost

racod

s,gast

rop

od

s,fo

ram

ini-

fera

an

dst

rom

ato

-p

oro

ids;

bio

turb

ati

on

Rest

ricte

d,

shall

ow

subti

dal

wit

hou

tst

ron

gcu

rren

tagit

ati

on

All

fms

excep

tR

on

gxia

nF

mat

Lit

an

g

Bio

turb

ate

dM

s/W

s(0

á2±1á5

m)

Bio

turb

ate

dM

s/W

sw

ith

pelo

idal

Ps/

Gs

len

ses

Mass

ive,

mott

led

ap

peara

nce,

an

ast

om

ose

dbu

rrow

netw

ork

s;se

lecti

ve

dolo

mit

izati

on

Rare

ost

racod

s,gast

rop

od

san

dsp

hero

idal

fora

min

ifera

(i.e

.u

mbell

ids)

Rest

ricte

dto

sem

i-re

stri

cte

dsh

all

ow

subti

dal

Abse

nt

inT

an

gji

aw

an

Fm

Foss

il-p

oor

Ms

(0á2

±1á0

m)

Foss

il-p

oor

Ms,

arg

illa

ceou

sse

am

sT

hin

-to

med

ium

bed

ded

,n

od

ula

rbed

din

glo

call

yR

are

thin

-sh

ell

ed

ost

racod

s,gast

rop

od

san

du

mbell

idfo

ram

inif

era

Rest

ricte

d,

qu

iet

shall

ow

subti

dal

Com

mon

inQ

iziq

iao

an

dG

uil

infm

s

Sem

i-re

stri

cted

subti

dal

faci

esO

stra

cod

Ps/

Gs

(0á2

±0á5

m)

Ost

racod

Ps/

Gs

wit

hm

inor

gast

rop

od

s,biv

alv

esh

ell

san

dp

elo

ids

Th

in-t

om

ed

ium

-bed

ded

;in

terc

a-

late

din

stro

mato

poro

idB

s;d

isart

icu

late

dan

dabra

ded

shell

s

Ost

racod

sd

om

inan

t,m

inor

gast

rop

od

s,biv

alv

es

Sem

i-re

stri

cte

d,

storm

-rew

ork

ed

dep

osi

tsor

shoals

beh

ind

bio

stro

mes

Tan

gji

aw

an

Fm

Gast

rop

od

Ws/

Ps

(0á2

±0á8

m)

Gast

rop

od

Ws/

Ps,

thin

arg

illa

ceou

sse

am

sin

terc

ala

ted

Th

in-b

ed

ded

;gra

ded

an

dlo

call

yp

refe

rred

ori

en

tati

on

;ra

resh

elt

er

fabri

cs

Abu

nd

an

tgast

rop

od

s,m

inor

ost

racod

san

dtu

bu

lar

fora

min

ifera

,ra

rete

nta

cu

liti

dsh

ell

s

Sto

rm-r

ed

ep

osi

ted

inse

mi-

rest

ricte

dd

eep

lagoon

or

intr

ap

latf

orm

basi

n

Low

er

part

of

Min

tan

gF

mat

Fu

he

``T

en

tacu

liti

d''

Ws/

Ps

(0á2

±4

m)

Ten

tacu

liti

d-l

ike

Ws/

Ps

Th

in-

toth

ick-b

ed

ded

,w

eakly

ori

en

ted

tubif

orm

shell

sT

en

tacu

liti

d-l

ike

fau

na

wit

hth

ick

wall

sd

om

inan

t;m

inor

ost

racod

s,gast

rop

od

s,fo

ram

inif

era

Sem

i-re

stri

cte

dd

eep

lagoon

on

pla

tform

s

Up

per

Mb

of

Gu

ilin

Fm



Table

1(C

on

tin

ued

).

Lit

hofa

cie

su

nit

+th

ickn

ess

Lit

holo

gy

Sed

imen

tary

stru

ctu

res/

textu

res

Bio

taIn

terp

reta

tion

Occu

rren

ce

Inte

rmed

iate

subti

dal

faci

esS

trom

ato

poro

id(d

olo

-)B

s(0

á5±15

m)

Str

om

ato

poro

idF

s/R

s(m

inor

Bi)

,ta

bu

lar

stro

mato

poro

idF

s/B

a;

skele

tal

Ws/

Ps

matr

ix;

stro

ngly

dolo

mit

ized

locall

y(e

specia

lly

inT

an

gji

aw

an

Fm

an

dequ

ivale

nts

)

Mass

ive

toth

ick-b

ed

ded

;ra

red

om

al

shap

ein

reli

ef;

overt

urn

ed

stro

mato

poro

ids

com

mon

,in

situ

gro

wth

posi

tion

rare

Bu

lbou

s,h

em

isp

heri

cal,

pla

ty/t

abu

lar,

bra

nch

ing

stro

mato

poro

ids;

cora

ls,

bra

ch

iop

od

s,gast

rop

od

s,fo

ram

inif

era

,ra

rebry

ozoan

san

dcri

noid

s;bio

turb

ati

on

Op

en

-mari

ne,

mod

e-

rate

lyd

eep

subti

dal;

pla

tform

inte

rior

bio

stro

mes

(or

bio

herm

s);

not

stro

ng

wave

resi

stan

tst

ructu

res

Inall

stra

ta,

bu

tqu

ite

abu

nd

an

tin

Tan

gji

aw

an

Fm

an

dequ

ivale

nts

Skele

tal

Ws/

Ms

(0á2

±2á0

m)

Div

ers

i®ed

skele

tal

Ws/

Ms;

spars

earg

illa

-ceou

sse

am

sin

terc

a-

late

d

Th

in-

tom

ed

ium

-bed

ded

Bra

ch

iop

od

s,cora

ls,

stro

mato

poro

ids,

gast

rop

od

s,cri

noid

s,ech

inoid

s,bry

ozoan

s,ra

rete

nta

cu

liti

ds

Op

en

-mari

ne,

mod

e-

rate

tod

eep

sub-

tid

al;

near

or

belo

wfa

ir-w

eath

er

wave

base

Tan

gji

aw

an

Fm

,lo

wer

part

of

Qiz

iqia

oF

m

Dee

psu

bti

dal/

basi

nal

faci

esN

od

ula

rsk

ele

tal

Ws/

Ms

(0á1

±2á0

m)

Arg

illa

ceou

ssk

ele

tal

Ws/

Ms

Nod

ula

r-bed

ded

;abra

ded

skele

tal

gra

ins

Div

ers

i®ed

bra

ch

iop

od

s,ru

gose

cora

ls,

cri

noid

s,gast

rop

od

s,bry

ozoan

s;ra

rete

nta

cu

liti

ds;

Th

ala

ssin

oid

-typ

ebu

rrow

s

Op

en

-mari

ne,

deep

subti

dal;

belo

wfa

ir-w

eath

er

wave

base

,ep

isod

icst

orm

agit

ati

on

Tan

gji

aw

an

Fm

,lo

wer

part

of

Qiz

iqia

oF

m

Ten

tacu

liti

dM

s/W

sin

terc

ala

ted

wit

harg

illi

tes

(0á2

±4á0

m)

Ten

tacu

liti

dM

s/W

s,in

terc

ala

ted

wit

hth

inarg

illa

ceou

sbed

sor

seam

s;h

igh

org

an

icm

att

er;

ch

ert

yn

od

ule

sor

ban

ds

(Liu

jian

gF

mon

ly)

Th

in-b

ed

ded

;u

pw

ard

-th

icken

ing

bed

din

gp

att

ern

sT

hin

-sh

ell

ed

ten

ta-

cu

liti

ds;

min

or

bra

ch

iop

od

s,calc

isp

on

ges,

gast

rop

od

s,fo

ram

inif

era

Very

deep

subti

dal

toin

terp

latf

orm

basi

n

Min

tan

gF

m,

low

er

part

of

Liu

jian

gF

m

Ms,

mu

dst

on

e;

Ws,

wackest

on

e;

Ps,

packst

on

e;

Gs,

gra

inst

on

e;

Bs,

bou

nd

ston

e;

Bi,

bin

dst

on

e;

Ba,

baf¯

est

on

e;

Fs,

¯oats

ton

e;

Rs,

rud

ston

e;

Fm

,F

orm

ati

on

;M

b,

Mem

ber.

62 D. Chen et al.


These depositional sequences were recognizedand correlated by cycle stacking patterns, andvertical facies trends combined with subaerialexposure features traceable across the differentplatforms.

Peritidal facies

Peritidal facies include planar laminite, wavylaminite, fenestral limestone and minor pedogenicand meteoric-altered units (Table 1). Planar lami-nites and wavy laminites (Fig. 3A) form caps toperitidal cycles and thin transgressive bases tosome peritidal cycles. Within the fenestral lime-stones, four fenestral types occur: laminoid, irreg-ular, tubular and spherical. Irregular (or tubular)fenestral limestones commonly grade into lami-noid ones, constituting upward-shallowing cycles.

Pedogenic/meteoric-altered units form the capsof cycles with host substrates either of peritidal orshallow subtidal facies. Karstic processes appearto have dominated over pedogenic processes, andcommon features include in situ breccias (rubble)(Fig. 4A and B), dissolution cracks and ®lls, andterra-rossa-type soils (Fig. 4C and D), with rarelaminated crusts.

In situ breccias can reach 40 cm in thicknessand are light pink-grey to dark brown in colour;their transition to the substrate is gradual, similarto the regolith illustrated by Szuczewski et al.

(1996; ®g. 3C and D). The breccias consist of cm-sized clasts of host rock either of peritidal facies,such as laminite and lime mudstone (peritidalpond deposit), or shallow subtidal facies such asAmphipora ¯oatstone/wackestone. Their matrixis a greenish to red dolomitic clay. Denselypacked clasts are generally irregular to subroun-ded in shape, with ®tted ferruginous rims in somecases (Fig. 4B). Vadose cements are common. Thebrecciation is probably related to periodic wettingand drying during subaerial exposure.

Dissolution depressions and pipes, anastomo-sing cracks and veins (Fig. 4C and D), and neptu-nian dykes are found at subaerial exposurehorizons. Dissolution residues of siltstone occurwithin the dissolution depressions, and dissolu-tion pipes below are ®lled with coarse blockycalcite. Anastomosing cracks in brecciation zonesare ®lled with granular ferroan calcite or dolo-mite, with some geopetal texture. Alveolar septaltexture is observed locally along the cracks andpossibly related to microbial activity around rootsystems. Some cracks are downward tapering,and range in width up to tens of centimetres andin depth up to 1á5 m. They are ®lled with subtidaldeposits of the overlying units, locally withbrownish to purple terra-rossa claystone at thebottom.

Clayey palaeosols form the top section ofsubaerial exposure-affected horizons (Fig. 4D).

Fig. 3. Common facies and cycle types in the study area. (A) Peritidal cycle. Mottled, bioturbated mudstone/wa-ckestone is covered by undulatory microbial laminite. See the network of irregular burrows in the lower part. Thecycle boundary is indicated by an arrow. Hammer for scale (37 cm long). Guilin Formation, Tangjiawan, Guilin. (B)Subtidal cycle. Thin- to medium-bedded skeletal wackestone/mudstone at the base changes into stromatoporoid¯oatstone/wackestone upwards; bulbous and columnar stromatoporoids are in their growth position. The black linemarks the cycle top. Hammer for scale (37 cm long). Qiziqiao Formation, Xizhaikou, Shaoyang County, Hunan.



They are generally 5±20 cm thick, grey to yellow-ish green in colour, and illite dominated. Rareroot moulds have been observed. In some cases,calcareous nodules occur within the claystone.Such clay soils are documented in Cretaceousand Triassic peritidal cycles and re¯ect a relat-ively humid climate during emergence (Decon-inck & Strasser, 1987; Strasser, 1988; Balog et al.,1997).

Rare laminated pedogenic crusts form the capsto fenestral limestones. They are light grey to pinkin colour and 2 cm to 10 cm in thickness, andconsist of ®ne laminae composed of micrite withcoated grains and peloids, along which greenishclayey bands are distributed.

Restricted shallow subtidal facies

Restricted shallow subtidal facies include: (1)stromatolite boundstone; (2) peloidal grainstone/packstone; (3) oolitic grainstone; (4) Amphipora

¯oatstone/packstone; (5) bioturbated mudstone/wackestone; and (6) fossil-poor mudstone/wacke-stone (Table 1). Stromatolite boundstones arecommonly present below, or intercalated withmicrobial laminites. They are commonly nucle-ated on scoured surfaces and intraclasts. Peloidalgrainstones/packstones form the lower or middleparts of peritidal cycles, and commonly gradeupwards into fenestral limestones or microbiallaminites. The rare oolitic grainstones are locatedin the lower or middle part of peritidal cycles,and generally grade into fenestral limestones ormicrobial laminites. The common Amphipora¯oatstones/packstones form the lower or middleparts of peritidal cycles or the cap to subtidalcycles. They are characterized by mono-speci®cspaghetti-like Amphipora stromatoporoids. Bio-turbated mudstones/wackestones (middle ofFig. 3A) form the lower or middle part of bothperitidal cycles and subtidal cycles. Fossil-poormudstones form the lower parts of peritidal

Fig. 4. (A) Brecciation during subaerial exposure. Note the cinder-like rubble coated with purple dolomitic argil-laceous seams. The original rock is Amphipora wackestone/mudstone, which is also reddened. Pen is 14 cm long.Guilin Formation. Tangjiawan, Guilin. (B) Polished slab from (A). Subangular to subrounded, red to purplish, in situclasts are coated with brownish, ferruginous crust. Black pebbles are also present (black arrow). Cavities are ®lled byblocky calcite with minor pendant calcite cements (white arrow). Matrix is dolomitic argillaceous mudstone. Scalebar is 1 cm long. (C) Funnel-shaped dissolution pit with insoluble residue (dolomitic siltstone). The light-colouredwalls (arrowed) with numerous dissolution veins beneath, are the result of leaching during subaerial exposure.Pencil is 13á5 cm long. Rongxian Formation, Litang, Guangxi. (D) An exposure pro®le showing the transition fromunaltered host rock to a brecciated zone with anastomosing veins and cavities, via an argillaceous brecciated zone,and ®nally to the dark greenish dolomitic siltstone/silty shale. The veins and cavities are ®lled with granulardolomite spar. The pencil is 14 cm long. Rongxian Formation, Litang, Guangxi.

64 D. Chen et al.


cycles or caps to subtidal cycles. Similar subtidalfacies have been described by Dorobek (1991),Elrick (1995, 1996), Garland et al. (1996) andLamaskin & Elrick (1997).

Semi-restricted subtidal facies

This less common facies association includes:ostracod packstone/grainstone, gastropod wacke-stone/packstone and `tentaculitid' wackestone/packstone (Table 1). The `tentaculitid' wacke-stones/packstones form the base of subtidalcycles (Fig. 5D), and are characterized by tenta-culitid-like shells (25±50% of the bioclasts) with

a relatively thick wall, suggesting a nekto-benthiclifestyle.

Intermediate subtidal facies

Intermediate subtidal facies include: stroma-toporoid (dolo-) boundstone and skeletal wacke-stone/mudstone (Table 1). Stromatoporoid (dolo-)boundstones form the lower or middle parts ofsubtidal cycles and peritidal cycles. Two sublit-hofacies: stromatoporoid ¯oatstone/rudstone andplaty/tabular stromatoporoid ¯oatstone/baf¯e-stone can be recognized (Table 1), with the lattergenerally grading into the former. Skeletal

Fig. 5. Common cycle types in the study area. (A, B) Peritidal cycles; (C, D) shallow subtidal cycles; and (E) deepsubtidal cycles. Bounding surfaces of cycles are marked by short lines on the right sides of logs. A, Litang; B±D,Tangjiawan; E, Ma'anshan. M, mudstone; W, wackestone; P, packstone; G, grainstone; F, ¯oatstone; B, baf¯estone;BI, bindstone.



wackestones/mudstones form the base or middleparts of some subtidal cycles.

Deep subtidal facies

Deep subtidal facies include nodular skeletalwackestone/packstone and tentaculitid mud-stone/wackestone intercalated with argillite(Table 1). Nodular skeletal wackestones/mud-stones are normally present at the base or in thelower part of subtidal cycles. Tentaculitid mud-stones/wackestones intercalated with argillitesform cycles several metres thick, displayingupward-increasing bed thickness.

CARBONATE PLATFORMDEPOSITIONAL MODEL

The Devonian platforms in South China aremostly isolated carbonate platforms, with rareclastic in¯uxes trapped within interplatformbasins (e.g. Shen et al., 1987; Wu et al., 1987;Zhong et al., 1992). The topography of the isolatedplatforms was variable, but generally with a sharptransition from platform to basin, particularlythose in western Guangxi. Platforms like GX, GLand GZL from east-central Guangxi to Hunanwere generally elongate in shape and arrangedroughly in a NNE±SSW direction (Fig. 1), follow-ing the orientation of the major fault zones. Theseplatforms commonly exhibited an asymmetricalpro®le with a sharp transition from platform tobasin on the western margin, but a relativelygentle slope on the eastern margin (Tan et al.,1987; Wu et al., 1987; Zhong et al., 1992). Reef orshoal complexes were commonly developed onwestern or southern margins, possibly related to

the SW-trending wind direction (Shen et al.,1987, 1994; Shen & Yu, 1996; Shen & Zhang,1997). The eastern, leeward margin was generallya distally steepened ramp with progradationtowards the interplatform basin (Fig. 6). Thisdepositional con®guration is different from theDevonian in the Great Basin region of the USA(a ramp system) (Dorobek, 1991; Elrick, 1995,1996; LaMaskin & Elrick, 1997), but similar to thatin western Canada and Europe (Burchette, 1981;Walls, 1983; Wendte, 1992).

Facies analysis from three isolated platforms(GX, GL and GZL) has revealed that extensiveorganic buildups with minor peritidal facies weredeposited in the lower Givetian, corresponding tothe the Tangjiawan Formation. Peritidal facieswere widely distributed from late Givetian toearliest Frasnian time, but then decreased involume until the end of the Frasnian. There arerelatively large volumes of subtidal facies atlocalities MA and XZ, and physiographicallythese two localities were located close to theeastern margin of the GX platform, particularlythe MA section. At TJ, the special `tentaculitid'wackestone/packstone lithofacies is interpretedas a deposit of a semi-restricted deep subtidalenvironment. At FH, carbonate platform facies areintercalated with thin, deep subtidal units in theGivetian, but rapidly pass up into basinal sili-ceous and pelagic carbonate facies in the Fras-nian. At LT, the dominant peritidal facies in theFrasnian are the shallowest of the measuredsections.

In all measured sections, depositional facies arecharacterized by micrite-rich deposits dominatedby wackestone/mudstone textures; wave andcurrent structures such as cross-bedding, ripplesand scours are rare; and only several layers of

Fig. 6. Depositional model of platforms in the study area. The platforms generally exhibit a sharp transition fromplatform to basin on the western side over a very short distance, and a relatively gentle slope on the eastern side. Reefand shoal complexes were commonly developed on the western margin, and biostromes on the eastern side. Theseplatforms can basically be thought of as eastward-dipping structures with distally steepened margins, which weredeveloped on tilt-fault blocks in an extensional tectonic regime.

66 D. Chen et al.


oolitic grainstone have been observed. This im-plies that the depositional facies package wasdeposited in a low-energy regime as a result ofeither wave-dampening caused by the intercon-nected basin system, or the effect of wavereduction by the reefs and shoals on the westernsides of the platforms. Low-energy conditionsacross platforms resulted in highly bioturbated,homogenized subtidal deposits and tidal-¯atmudstones.

HIGH-FREQUENCY CYCLICITYOF METRE-SCALE,UPWARD-SHALLOWING CYCLES

Types of metre-scale cycle

Peritidal facies through deep subtidal/basinalfacies (Table 1) are organized into metre-scale,upward-shallowing cycles averaging 1á5±3á0 mthick, lasting 20±100 kyr (Tables 2 and 3). Cycleboundaries vary from abrupt to transitional.Based on the intracycle facies arrangement andfeatures of bounding surfaces, two kinds of cycleare recognized: peritidal cycles, capped by peri-tidal facies, and subtidal cycles, capped byshallow through intermediate subtidal facies.

The peritidal cycles are the most common(Table 2).

Peritidal cycles

Peritidal cycles are characterized by upward-shallowing successions (intermediate subtidal orrestricted subtidal to peritidal facies) capped byperitidal facies (Figs 3A and 5A,B). Two subtypesof peritidal cycle are recognized: regressive-proneor asymmetrical cycles, and transgressive-proneor symmetrical cycles. Regressive-prone orasymmetrical cycles are regressive (upward-shal-lowing) facies successions commencing withintermediate subtidal, restricted subtidal or peri-tidal facies, and are capped by peritidal facieswith an obvious facies `jump' across boundingsurfaces; that is, there is an abrupt deepeningabove cycle boundaries (Fig. 3A).

Transgressive-prone or symmetrical cycles arecharacterized by laminite bases followed bynormal regressive (upward-shallowing) succes-sions with gradual facies changes across thebounding surfaces (Fig. 5A). The basal transgres-sive laminite is basically similar to the cappinglaminite but with an increase in bioturbation,lamina thickness and darkness of colour, and a

Table 2. Statistics of cycle distribution in measured sections, Hunan and Guangxi provinces.

LocalityTotalthickness (m)

Total cyclenumber

Mean cyclethickness (m)

Peritidalcycles

Subtidalcycles

Ma'anshan 694á50 250 2á78 144 (57á6%) 106 (42á4%)Xizhaikou 399á85 207 1á93 113 (54á1%) 94 (45á9%)Tangjiawan 530á72 271 1á96 195 (72.0%) 76 (28.0%)Litang 410á87 249 1á65 218 (87á6%) 31 (12á4%)Fuhe 130á28 76 1á71 39 (51.3%) 37 (48á7%)

Table 3. Estimation of cycle duration for the measured sections. Carbonate deposition at Ma'anshan and Xizhaikoubegan approximately one conodont zone later in the Givetian (~0á5 Myr) and at Tangjiawan half a conodont zone later(~0á25 Myr), compared with the southern part of Guangxi Province, where deposition started at the beginning of theGivetian. The time-scale is based on the data of Odin et al. (1982), Palmer (1983), Harland et al. (1989) and Fordham(1992) respectively.

Time-scale Ma'anshan Xizhaikou Tangjiawan Litang

Givetian 5, 6, 3á5, 4á8 Myr 128 106 >148 Cycle number23á4±43á0 kyr 28á3±51á9 kyr 22á0±38á9 kyr Cycle duration

Frasnian 10, 7, 10á5, 8á7 Myr ~105 96 ~128 114 Cycle number66á7±100 kyr 72á9±109á4 kyr 54á7±82á0 kyr 61á4±92á1 kyr Cycle duration53á7±62á2 kyr 61á9±71á8 kyr 33á9±39á2 kyr Mean cycle

duration



decrease in the degree of exposure. The laminaehave a wavy, undulatory appearance.

Subtidal cycles

Two varieties of subtidal cycle are recognized:shallow subtidal cycles, characterized by upward-shallowing successions composed of interme-diate subtidal, restricted (or semi-restricted)subtidal to shallow subtidal facies (Figs 3A, and5C and D), and deep subtidal cycles, upward-shallowing successions composed of deep tointermediate subtidal facies (Fig. 5E).

Shallow subtidal cycles are incomplete cycles,because the available accommodation space fromthe sea¯oor to sea-level was not fully ®lled.Their occurrence indicates that the accommoda-tion increase was faster than sediment aggrada-tion. They are present across platforms mainlywithin overall transgressive deposits. Thesecycles are basically upward-shallowing cycleswithout obvious evidence of prolonged subaerialexposure; only a tiny proportion exhibits fea-tures of subaerial exposure (e.g. in situ brecciasand microkarst) at cycle tops in the lower part ofthe Upper Member of the Guilin Formation at TJ,Guilin. This situation more commonly occurredin cycles composed of biostrome facies as aresult of their relatively high relief from thesea¯oor, or as the result of a sharp fall in relativesea-level. The thickest of all cycles, up to 15 m,is of this type.

Deep subtidal cycles are rare, and are onlyobserved at the bottom of the Qizhiqiao Forma-tion (or its equivalent) and at several horizons atFH. They are characterized by micrite-richdeposits with thin storm-generated layers and adeep-water biota, such as tentaculitids, in thelower parts of the cycles, representing low-energyconditions below fair-weather wave base. Theyare present mainly within overall transgressivedeposits along platform margin slopes or ininterplatform basins.

Cyclicity of metre-scale, upward-shallowingcycles

It is dif®cult to determine cycle duration fromthick platform carbonate successions as a result ofa lack of accurate radiometric age data and the`missed beat effect' (cf. Goldhammer et al., 1990).Here only an approximate estimation of cycleduration (Table 3) can be made with the giventime-scale of the Givetian and Frasnian (Odinet al., 1982; Palmer, 1983; Harland et al., 1989;

Fordham, 1992). It would appear that cycles inthe Givetian are in the range 20±52 kyr, whereasthose in the Frasnian were longer, in the range50±110 kyr. If the deposits in the Givetian andFrasnian are taken as a whole, the cycle durationis in the range 30±72 kyr (the bottom row ofTable 3). If `missed beats' (subaerial exposure anddeep subtidal missed beats) are taken intoaccount, the cycle durations would be shorterthan these estimated values. As with most metre-scale carbonate cycles, these Devonian ones arewithin the Milankovitch band, corresponding tofourth-and ®fth-order cycles (cf. Goldhammeret al., 1993).

The origin of metre-scale, carbonate cycles isstill a controversial topic, provoking much debate(e.g. Drummond & Wilkinson, 1993a4,5 ,4,5 b; Wilkinsonet al., 1996, 1997). Three mechanisms are com-monly cited to explain the repetition: (1) autocy-clic generation (e.g. Pratt & James, 1986; Cloydet al., 1990; Satterley, 1996), (2) episodic subsi-dence (e.g. Cisne, 1986; Hardie et al., 1991), and(3) high-frequency, glacio-eustatic ¯uctuations insea-level (e.g. Goodwin & Anderson, 1985; Gold-hammer et al., 1990, 1993; Balog et al., 1997;Strasser & HillgaÈrtner, 1998).

Although all three mechanisms cited above arepossible driving forces for cycle formation, theglacio-eustatic model does best explain the fea-tures of these Devonian cycles. Exposure-cappedperitidal cycles are interpreted as the result ofhigh-frequency sea-level fall below the platformsurface for one or several sea-level ¯uctuations(subaerial exposure `missed beats'; Goldhammeret al., 1990; Elrick, 1995). Subtidal cycles, correl-ative with updip peritidal cycles, re¯ect theincreased accommodation space downdip createdby high-frequency sea-level rise and differentialsubsidence, without complete ®lling of the spacebefore the ensuing sea-level rise. Symmetricalcycles with transgressive laminite bases are moreprobably related to symmetrical and low ampli-tude sea-level oscillations, allowing a degree ofkeep-up in the early transgressive part of a newcycle, and favouring the deposition of transgres-sive laminites (cf. Koerschner & Read, 1989).Glacio-eustasy is also indirectly supported by theevidence of continental glaciation in the LateDevonian (Caputo & Crowell, 1985). However,widespread exposure horizons in the lowerFrasnian platform strata are more probably tec-tonically controlled as suggested by the presenceof nearly synchronous siliceous rocks in basinsand megabreccias at basin margins (discussedlater).

68 D. Chen et al.


CYCLE STACKING PATTERNSAND DEPOSITIONAL SEQUENCES

In structurally isolated and deformed shallow-water carbonate outcrops with high-frequencycycles, it is dif®cult to use the traditionalsequence stratigraphic approach to determinethe depositional sequences in view of the lackof recognizable seismic-scale stratal geometries.However, one-dimensional stratal data, includingvertical cycle stacking patterns and systematicfacies changes, can be used.

Cycle stacking patterns and accommodationchange

The vertical stacking patterns of metre-scale,upward-shallowing cycles are mostly controlledby long-term changes in accommodation space;therefore they bridge the gap from individualcycles to the larger-scale depositional sequences,and permit the identi®cation of sequences andtheir component systems tracts (e.g. Goldhammeret al., 1990, 1993; Osleger & Read, 1991; Elrick,1995).

Fischer plots are useful graphical tools toillustrate vertical stacking patterns and long-termchanges in accommodation space (Fischer, 1964;Read & Goldhammer, 1988; Sadler et al., 1993),particularly for peritidal cycle packages. There is,however, some dispute over the usefulness ofthis method to identify long-term changes ofaccommodation space (Drummond & Wilkinson,1993a,b). Basically, the cycle thickness is a re¯ec-tion of changes in accommodation space, but therelationship is complicated by many factors suchas quasi-periodicity of cycles, variable sedimen-tation rates, incomplete shallowing to sea-level,non-linear subsidence rates, and missed beats.However, if extra information is included, theabove negative effects can be mitigated. Fischerplots are conventionally drawn by cumulativedeparture from mean cycle thickness against cyclenumber (no less than 50) (e.g. Sadler et al., 1993).In this way, thick cycle packages will positivelydeviate from the mean cycle thickness, and formthe rising limbs of the plots, re¯ecting a long-termincrease in accommodation space; whereas thincycle packages will negatively deviate from theaverage cycle thickness, and form the fallinglimbs, re¯ecting a long-term decrease in accom-modation space (Figs 7 and 8). In a more objectiveway, Fischer plots can be constructed with cumu-lative departure from mean cycle thicknessagainst cycle thickness (e.g. Fig. 9); however, the

long-term trend in this modi®ed Fischer plotshows little difference from the conventional plot,just with a more gentle rising limb and a steeperfalling limb. Precautions still need to be kept inmind when applying Fischer plots to the inter-pretation of subtidal cycles, although there arecases where such plots have been used success-fully to predict long-term changes in accommo-dation space for subtidal cycles (e.g. Osleger,1991; Osleger & Read, 1991). For deep subtidalcycles, the opposite trend may occur in theFischer plot (i.e. the lower part of Figs 7 and 8),if the sedimentation rate is low and the waterdepth is too deep for the sediments to record theaccommodation created by high-frequency, low-amplitude sea-level ¯uctuations (deep subtidalmissed beats). Because this kind of cycle onlycomprises a small part of the succession andperitidal cycles dominate (Table 2), the effect ofthis `subtidal cycle syndrome' is minimal. Even atsection FH, where facies are dominated by deep-water packages with numerous `subtidal missedbeats', the trend of accommodation changes(below the Liujiang Formation), although not thetrue re¯ection, can still be displayed with theFischer plot (Fig. 9). Moreover, the capping faciesof most subtidal cycles is a very shallow-waterfacies (i.e. Amphipora wackestone/packstone), sothat the cumulative lost information with aFischer plot is not signi®cant, especially in viewof the intercalation with peritidal cycles.

In order to alleviate the shortcomings andincrease the reliability of a single Fischer plotin predicting changes in accommodation space,other evidence such as vertical facies changesand stratigraphic distribution of subaerial expo-sure indicators have been integrated for thepresent study. Histograms of percentage of peri-tidal facies per cycle have been constructed, andcompared with the Fischer plots drawn frommeasured sections (MA, XZ, TJ and LT). Majorsubaerial exposure horizons and deep subtidalfacies indicators, such as black shale or nodularlimestone, are also labelled on the Fischer plots(Figs 7 and 8). This approach integrates theFischer plot of individual sections with the dataof vertical facies changes to provide more exactinformation for long-term changes in accommo-dation space.

Sequence identi®cation and correlationin platform carbonates

Sequence boundaries are the keys to determinethird-order depositional sequences. Detailed



outcrop study revealed that sequence boundariesare generally gradational over metres to tens ofmetres. These transitional boundaries areconformable stratigraphic zones composed ofstacked, high-frequency and thinner-than-averagecycles with upward-increasing intensity of sub-aerial exposure, rather than discrete, laterallytraceable surfaces de®ning the sequences (e.g.Goldhammer et al., 1990, 1993; Elrick & Read,1991; Elrick, 1995). These stacked exposure-capped cyclic successions represent platformsaffected by multiple episodes of subaerial expo-sure formed by high-frequency (fourth- to ®fth-order) sea-level ¯uctuations superimposed onlong-term (third-order) sea-level lowstands, ratherthan a single, long-lasting exposure event.Therefore, the sequence boundary zones shouldlie roughly halfway down to the lowest points onthe falling-limb sections of Fischer plots, repre-senting the maximum rate of accommodationloss. However, a more exact location of thesequence boundaries is largely dependent on

the evidence from vertical facies changes, i.e.the shallowest facies association with a relativelyhigh percentage of peritidal facies and obvioussubaerial exposure indicators; these sequenceboudaries are labelled in Figs 7 and 8.

Individual sequences are generally 15±90 mthick, and consist of 15±50, metre-scale, upward-shallowing cycles. The number of cycles is lesswithin the sequences of section FH, where moredeeper-water subtidal facies predominate (Fig. 9),and there may have been numerous subtidalmissed beats. Internally the sequences are basic-ally composed of a transgressive lower part and aregressive upper part (Figs 7 and 8). The trans-gressive deposits consist of thicker-than-average,upward-thickening cycles dominated by subtidalfacies, and the regressive deposits consist ofthinner-than-average, upward-thinning cyclesdominated by peritidal facies. The transitionbetween transgressive and regressive packages isgradational, so the maximum ¯ooding surface ofeach sequence is tentatively placed at the mid-

Fig. 7. Fischer plots of high-frequency metre-scale carbonate cycles combined with histograms showing percentageof peritidal facies per cycle in the Qiziqiao Formation (Givetian to Frasnian) at Ma'anshan and Xizhaikou. Majorsubaerial exposure features are labelled on the Fischer plots. The gaps within the curves are areas of no exposure inthe ®eld. Sequence boundary identi®cation is mainly based on the cycle stacking patterns (on descending limbs ofthe plot), vertical facies variations (i.e. where high percentages of peritidal facies occur) combined with exposureindicators. Sequence boundaries (SB) are generally transitional zones.

70 D. Chen et al.


point of the thickest cycle in the vertical succes-sion.

It is necessary to note that the bottom of theTangjiawan Formation in the Guilin area roughlycorresponds to the base of varcus conodont zone(Shen & Yu, 1996); and the Qiziqiao Formation incentral Hunan was deposited later (approximate-ly half a conodont zone). Stringocephalus (nolater than hermanni-cristatus Zone, Shen & Yu,1996) disappears in the third sequence above theclastics in the Guilin area, but in the second incentral Hunan, so there is one sequence less atMA and XZ. The top boundary of the QiziqiaoFormation in central Hunan approximately cor-responds to the base of the middle triangularisZone or a little higher (Ji, 1991). A transitionalunit named the Laojiangchong Member (BGMR ofHunan, 1997) exists between the ChanglongjieFormation and the underlying platform carbon-ates (top of the Frasnian), in which two small-scale sequences have been recognized (Muchezet al., 1996; Bai et al., 1997). These two sequencesare also recognized in the lower part of theDongcun Formation at TJ and Wuzhishan Forma-

tion at FH. At MA and XZ, two small-scalesequences are recognized in the uppermost partof the Qiziqiao Formation (one less at XZ due toincomplete measurement of the section, Fig. 7);their base therefore corresponds to the upperboundary of the Frasnian. At LT, the RongxianFormation is overlain by slope deposits of earlyCarboniferous age (with the coral Pseudouraliniasp.). The Frasnian±Famennian boundary isplaced at the top of the sequence below thesecond gap (Fig. 8) through comparison withFamennian sequences elsewhere (e.g. Johnsonet al., 1985; Chen & Chen, 1994a; Du et al.,1996). Moreover, the brachiopod Cyrtospirifer,an index fossil of the Upper Devonian, occurs inthe seventh sequence at TJ, and the sixth at MAand XZ, or a little lower. Futhermore, the Fras-nian±Famennian event with its dramatic bioticand sedimentological changes, is also used tode®ne the top boundary of the Frasnian.

From Givetian through Frasnian strata, ninesequences and sequence boundary zones can berecognized. However, at the MA and XZ sectionsthere is one sequence less, attributable to the later

Fig. 8. Fischer plots of high-frequency carbonate cycles and facies distribution (histograms showing percentage ofperitidal facies per cycle) of Givetian and Frasnian strata from the Tangjiawan (including Tangjiawan, Guilin andlower Dongcun formations) and Litang sections. Gaps on the plot are covered stratigraphic intervals in the ®eld. TheLitang section is dominated by the shallowest, peritidal deposits in these measured sections, and the Frasnian±Famennian boundary is at cycle 124. Sequence boundaries are marked with SB.



start of deposition in the early Givetian. Thecorrelation of these sequences is given in Fig. 10,based on internal facies organization, progressivefacies change and the biostratigraphic framework.

Sequence 1 is the ®rst carbonate sequenceoverlying clastic strata and is only developed inGuilin and farther south, because the Devonianmarine transgression started from the southernarea of Guangxi, and gradually extended north-ward. The internal succession is dominated by

subtidal cycles (including many dolomitizedbiostromes of stromatoporoid boundstone/rud-stone) with minor peritidal cycles above.

Sequence 2 is recognized from Guangxi up toHunan. Transgressive deposits are characterizedby thick subtidal cycles with many biostromes,particularly in the upper part, and are alsostrongly dolomitized. Some deep subtidal cyclesare observed in the lower part of the sequences atlocalities MA and XZ, and show a falling pattern

Fig. 9. Stratigraphic column fromGivetian to lower Frasnian strata atFuhe (with fossil assemblage datafor the age de®nition). The rightsaw-like curve is a modi®ed Fischerplot drawn with the cumulative de-parture from mean cycle thicknessagainst cycle thickness, in order to®t to the stratigraphic log. Episodicdeepening interrupted by microbiallaminite intervals in the MintangFormation (Givetian) are followedby a rapid deepening at the base ofthe Liujiang Formation (latest Give-tian to early Frasnian), recording along-term accelerating accommoda-tion gain at this time. The Fischerplot does not correlate very wellwith the accommodation changesrecorded on the platform, especiallyfrom the upper part of the MintangFormation (just above the coveredinterval), as a result of the effect ofdeeper-water sedimentation and`subtidal missed beats', when waterdepth was too deep to respond tosmall-scale, high-frequency sea-level changes. Sh, shale; M, mud-stone; W, wackestone; P, packstone;G, grainstone.

72 D. Chen et al.


on the Fischer plots; this is the result of the lowsedimentation rate and/or compacted clay-richhorizons within the cycles. Sequences 1 and 2were deposited at a time of rapidly increasingaccommodation space.

Sequences 3±7 record a long-term change froma slow to a rapid decrease in the amount ofaccommodation space available. Sequences 3±5contain relatively high numbers of peritidalcycles; subtidal cycles are concentrated withinthe lower transgressive parts, and generally arehighly dolomitized. Prolonged exposure is onlyrecorded towards the top of Sequence 5 in theshallowest succession at LT (Figs 5A and 8).Sequence 6 is the thinnest of the sequences ineach section, ranging from 15 to 30 m in thick-ness, indicating the smallest amount of accom-modation space. Exposure events represented bypedogenic and palaeokarstic horizons are ob-

served towards the top of this sequence in almostall localities (Figs 7, 8 and 10). These exposureevents occur locally at the top of subtidal cycles,indicating that their formation was abrupt andprolonged. These long-lasting periods of exposureand lack of accommodation space preventedfurther deposition on the platforms, and resultedin the absence of a regressive, upper part toSequence 6. Sequence 7 is transitional and showsan increase in the volume of subtidal facies and inthe thickness of the sequence, indicating adecrease in the rate of accommodation loss.However, the signi®cant loss of accommodationspace at the end of Sequence 6 still had anobvious impact on the deposition of thissequence; the exposure events observed locallyin the middle part of this sequence indicate thatthere was abrupt exposure or insuf®cient accom-modation space across the platforms (Fig. 10). By

Fig. 10. Correlation of Fischer plots between outcrops from three isolated carbonate platforms in Hunan and Guangxiprovinces. Nine sequences (third-order) were identi®ed in Givetian and Frasnian strata, of which the lower ®vesequences (Sequence 1±5) are Givetian, and the upper four sequences (Sequence 6±9) are Frasnian. There is a long-term reduction in accommodation space from Sequence 3±6 in the platform successions, but an opposite trend isobserved in the interplatform basin successions, as at Fuhe. The trend in the Fischer plot from Sequence 6 at Fuhe isnot a true re¯ection of accommodation change, but is a result of deeper-water sedimentation and the ``subtidalmissed beats'' effect; deep-water facies are indicated below the dashed line. The thicker line between Sequence 6 and7 is the most obvious boundary with widespread subaerial exposure. See Figs 7 and 8 for legend.



contrast, at FH, evidence for sporadic and gradualdeepening is recorded in Sequences 3 through 5,and sudden deepening is recorded in Sequence 6,with the deepest water facies occurring in thelower part of Sequence 7 (Fig. 9).

Sequences 8 and 9 record a long-lasting, rapidincrease in accommodation space with relativelythick cycles. The volume of subtidal faciesincreases and peritidal facies decreases throughthe sequences. Even the peritidal facies packagesshow an increasing degree of deepening asindicated by more fenestral limestones thanlaminites, particularly in Sequence 9. The top ofSequence 9 roughly corresponds to the boundarybetween the Frasnian and Famennian, whichoccurs in the Qiziqiao, Dongcun and Rongxianformations.

Sequence stacking patterns

The sequences identi®ed from cycle stackingpatterns and vertical facies changes can beroughly grouped into two categories: catch-upsequences and keep-up sequences (Fig. 10).

Catch-up sequences (Sequences 1±2 and 8±9)are characterized by packages of relatively thickcycles with a high percentage of intermediate toshallow subtidal facies, and even deep subtidalfacies locally within individual cycles. Theseform the ascending wave trains on the Fischerplots and indicate a long-term (second-order)increase in accommodation space. Thesesequences are dominated by subtidal cycles,particularly in the lower transgressive part ofthe sequences, and minor peritidal cycles occurin the more regressive, upper part. These featuresindicate that sedimentation rates on platformswere slower than the accommodation gains cre-ated by the combined effects of ®fth- throughthird-order relative sea-level changes, enhancedby the second-order, long-term relative sea-levelrise.

Keep-up sequences (Sequences 3±6) are char-acterized by relatively thin cycle packages with ahigh percentage of peritidal facies within indi-vidual cycles. These form the descending limbson the Fischer plots and indicate a long-term(second-order) accommodation decrease. At lo-calities MA and XZ, there are no obvious risinglimbs even in the transgressive deposits withinthese sequences. Subtidal cycles are only foundin the transgressive deposits, and exposure-capped peritidal cycles become more commontowards the top of the sequences. Furthermore, inSequence 6 the upper part of the regressive

deposits are mostly absent because of the stacked,prolonged subaerial exposure events towards thetop. These characteristics indicate that the sedi-mentation rate on the platform generally keptpace with accommodation gain created by thecombined effects of ®fth- through third-orderrelative sea-level changes, attenuated by asecond-order, long-term relative sea-level fall.The combined effects caused a maximum rate ofaccommodation loss on the platforms at the endof Sequence 6. Sequence 7 is transitional (i.e.deposited during a period of long-term stillstand)with characteristics of both keep-up and catch-upplatform sequences.

Within interplatform basinal successions (e.g.Wu et al., 1987; Jin, 1990), such as at locality FH(Fig. 9), a widespread deepening process wasstarted from Sequence 6, with the deepest waterrepresented by pelagic, hemipelagic and siliceousfacies in the lower part of Sequence 7, after whichthe water depth became shallow.

Long-term changes in cycle stacking patternsand tectonic implications

Long-term changes in cycle stacking patterns andaccommodation in these Devonian platform car-bonates are well illustrated with the Fischer plots(Figs 7, 8 and 10). Although there are somedifferences, comparison of Fischer plots con-structed from sections located on differentplatforms reveals a similar pattern (Fig. 10),suggesting similar, long-term trends in accommo-dation change. In these platform successions, arapid increase in accommodation space duringthe early Givetian (Sequences 1±2) was followedby a gradual to rapid decrease in accommodationspace from the middle to late Givetian (Sequences3±5), reaching a maximum rate of accommodationloss in the early Frasnian (the top of Sequence 6).Afterwards, the accommodation space changewas uniform or increased slightly (Sequence 7);this was followed by a rapid increase until theend of the Frasnian (Sequences 8±9). In detail,small differences in accommodation change arepresent between individual platforms, as can beseen from the Fischer plots of localities MA andXZ (both on platform GX) when compared withlocalities TJ (on platform GL) and LT (on platformGZL) (Fig. 10). These re¯ect small variations inthe subsidence rate of the three platforms.

One signi®cant phenomenon in Sequence 6and 7 is that the rapid loss in accommodationspace on the platforms coincided with a rapidgain in accommodation in the interplatform

74 D. Chen et al.


basins. The sharp deepening event in the basinindicated by siliceous deposits (bedded chertsand siliceous shales) and coarse debrites at thebasin margins, corresponds to the major subaerialexposure horizon between the two sequences onthe platforms (Fig. 10). This opposing trend inaccommodation change between platforms andbasins is well demonstrated at the FH section(Fig. 9), a pro®le recording the transition fromplatform to the interplatform basin, on the easternside of the Guilin platform.

This contrast between platforms and basins wasthe result of the extensional tectonic movements

noted earlier, whereby rapid deepening andincrease in accommodation space in basins, wasaccompanied by relative uplift of platforms,resulting in a large loss of accommodation thereand so shallowing and subaerial exposure(Fig. 11). During the fault movements, a jerkysubsidence of the basins (hanging walls) triggeredslope failure, and led to the deposition of debritesand slumps, followed by deep-water pelagicdeposition in the basin. The exposure discon-formity zones on the platforms (at the top ofSequence 6) therefore roughly correspond to thesudden deepening event indicated by the debrites

Fig. 11. Schematic diagram showing the tectonic controls on platform and interplatform basin deposition and de-velopment in an extensional regime. The rapid subsidence of interplatform basins controlled by faults would beaccompanied by the relative uplift of platforms, resulting in an accommodation gain in the basins, but an accom-modation loss on the platforms. The condensed exposure disconformity zones on platforms therefore correspond tosharp deepening in the basins. The arrows mark the relative movement direction of blocks. Wd0, initial water depthbefore fault movement; Wd1, water depth after fault movement.

Fig. 12. Comparisons of accommodation changes on platforms and in basins in the study area with the eustaticsea-level curve constructed by Johnson et al. (1985), for Givetian through Frasnian strata. Numbers (1±9) are thedepositional sequences documented in this paper. Note the opposite trend in accommodation changes, and thecoincidence of maximum rate of accommodation loss with accommodation gain in Sequence 6 (shadowed bar)between platforms and basins. Note the broadly opposite and similar trend, respectively, for platform and basins, inaccommodation changes, compared with Johnson et al.'s eustatic sea-level curve.



and related siliceous deposits in the basins. Thisopposite pattern in accommodation change ismost probably related to the tectonic activity atthe time.

The long-term changes in accommodation dis-cerned from cycle stacking patterns from differentplatforms in South China do not show a goodcorrelation to the pattern of eustatic sea-levelchange in Euramerica postulated by Johnsonet al. (1985). In fact, an opposite trend in accom-modation changes is mostly seen (Fig. 12). Thissuggests that local tectonics was a major factor inplatform generation and evolution in SouthChina, as seen in the opposing pattern of accom-modation changes between platforms and basinsin Sequences 6 and 7.

CONCLUSIONS

1 From the Middle Devonian, a complex patternof platforms and interplatform basins was wellestablished in South China. Five depositionalfacies including peritidal, restricted shallow sub-tidal, semi-restricted subtidal, intermediate sub-tidal, and deep subtidal facies, and 18 lithofaciesunits are recognized from measured sectionslocated on three isolated platforms in Givetianand Frasnian successions. These deposits arearranged into metre-scale, upward-shallowingcycles grouped into peritidal and subtidal cycles,which have an average cycle period between »22±110 kyr. This periodicity roughly falls within thetime-scale of Milankovitch rhythms, although themechanism of cycle formation is still uncertain.2 Nine third-order depositional sequences areidenti®ed, based on changes in cycle stackingpatterns, vertical facies changes, and the strati-graphic distribution of subaerial exposure indi-cators. The internal components of sequences aremainly composed of lower transgressive andupper regressive parts. Transgressive packagesare dominated by thicker-than-average subtidalcycles, whereas regressive parts are dominated bythinner-than-average peritidal cycles. Sequenceboundaries are generally conformable, transitionalzones composed of stacked, high-frequency, andthinner-than-average cycles with upward-increas-ing intensity of subaerial exposure, rather thandiscrete, laterally traceable surfaces.3 The identi®ed sequences can be furthergrouped into catch-up and keep-up sequence setsfrom the long-term (second-order) trend inaccommodation space revealed by Fischer plots.Catch-up sequences (Sequences 1±2, 8±9) are

characterized by packages of relatively thickcycles with a high percentage of intermediate toshallow subtidal facies, and even deep subtidalfacies locally within individual cycles, recordingincreased rates of long-term accommodation gain.Keep-up sequences (Sequences 3±6) are charac-terized by packages of relatively thin cycles witha high percentage of peritidal facies withinindividual cycles, recording increased rates oflong-term accommodation loss. This long-termsequence stacking pattern is tectonically con-trolled.4 The long-term increasing accommodation losson platforms from late Givetian to early Frasniantimes coincided with increasing accommodationgain in interplatform basins. This implies that,under an extensional regime, the deepening ininterplatform basins induced by increasing sub-sidence can be associated with the relative upliftof platforms, resulting in accommodation lossand shallowing there. The deep-water packagesin basins (i.e. between Sequence 6 and 7) there-fore roughly correspond to the stacked exposuredisconformities on the platforms.

ACKNOWLEDGEMENTS

This research was funded by the National NaturalScience Foundation of China (grant no. 49404023,and 49872043 to C.D.Z.). Fieldwork was gener-ously assisted by Professors Baoan Yin, Yi Wu,and many young colleagues in the RegionalGeological Survey of Guangxi at different times.We are very grateful to the K. C. Wong EducationFoundation and the Royal Society for the supportto the senior author to visit Durham University forone year. Thanks also go to the Institute of Geology& Geophysics, Chinese Academy of Sciences, forallowing the senior author's research absencefrom the Institute during the tenure of the Fellow-ship. The manuscript has bene®ted greatly fromreviews by Maya Elrick and Bruce Wilkinson.Numerous constructive editorial comments fromDr Ian Jarvis are highly appreciated.

REFERENCES

Bai, S.L., Bai, Z.Q., Ma, X.P., Wang, D.R. and Sun, Y.L. (1997)

Devonian Events and Biostratigraphy of South China.

Peking University Press, Beijing, China. pp. 62±113.Balog, A., Haas, J., Read, J.F. and Coruh, C. (1997) Shallow

marine record of orbitally forced cyclicity in a Late

76 D. Chen et al.


Triassic carbonate platform, Hungary. J. Sed. Res., 67,661±675.

Burchette, T.P. (1981) European Devonian reefs: a review of

current concepts and models. In: European Fossil ReefModels (Ed. D.F. Toomey), SEPM Spec. Publ., 30, 85±142.

Bureau of Geology and Mineral Resources (BGMR) of HunanProvince (1997) Multiple Classi®cation and Correlation ofthe Stratigraphy of China (43) ± Stratigraphy (Lithostrati-graphy) of Hunan Province. China University of Geoscience

Press, Wuhan, China. pp. 123±160.

Caputo, M.V. and Crowell, J.C. (1985) Migration of glacialcenters across Gondwana during Paleozoic Era. Geol. Soc.Am. Bull., 96, 1020±1036.

Chen, D.Z. and Chen, Q.Y. (1994a) Devonian sedimentary

evolution and transgressive-regressive patterns in SouthChina. Sci. Geol. Sinica, 29, 246±255 (in Chinese with

English abstract).

Chen, D.Z. and Chen, Q.Y. (1994b) Sequence stratigraphic

frameworks and sea-level changes of Early and MiddleDevonian, southern Guizhou. Sci. Sinica (Series B), 24,1197±1205 (in Chinese).

Chen, D.Z. and Chen, Q.Y. (1995) Dynamics of sedimentaryevolution for the Early and Middle Devonian, southern

Guizhou. Acta Sedimentol. Sinica, 13, 54±65 (in Chinese

with English abstract).

Chen, H.D. and Zeng, Y.F. (1990) Nature and Evolution of theYoujiang Basin. Sed. Facies Palaeogeogr., 1, 28±37 (in

Chinese with English abstract).

Cisne, J.L. (1986) Earthquakes recorded stratigraphically on

carbonate platforms. Nature, 323, 320±322.Cloyd, K.C., Demicco, R.U. and Spencer, R.J. (1990) Tidal

channel, levee, and crevasse splay deposits from a Cambrian

tidal channel system ± a new mechanism to produce shal-lowing-upward sequences. J. Sed. Petrol., 60, 73±83.

Deconinck, J.-F. and Strasser, A. (1987) Sedimentology, clay

mineralogy and depositional environments of Purbeckian

green marls (Swiss and French Jura). Eclogae Geol. Helv.,80, 753±772.

Dorobek, S.L. (1991) Cyclic platform carbonates of the Devo-

nian Jefferson Formation, southwestern Montana. In:

Paleozoic Paleogeography of the Western United States-II(Eds J.D. Cooper and C.H. Stevens), SEPM Paci®c Section,

67, 509±526.

Drummond, C.N. and Wilkinson, B.H. (1993a) Aperiodic

accummulation of peritidal carbonates. Geology, 21, 1023±1026.

Drummond, C.N. and Wilkinson, B.H. (1993b) On the use of

cycle thickness diagrams as records of long-term sea-levelchange during accummulation of carbonate sequences.

J. Geol., 101, 687±702.

Du, Y.S., Gong, Y.M. and Wu, Y. (1996) Devonian sequence

stratigraphy and sea-level change cycles in South China.J. China Univ. Geosci., 7, 72±79.

Elrick, M. (1995) Cyclostratigraphy of Middle Devonian car-

bonates of the eastern Great Basin. J. Sed. Res., B65, 61±79.

Elrick, M. (1996) Sequence stratigraphy and platform evolu-tion of Lower-Middle Devonian carbonates, eastern Great

Basin. Geol. Soc. Am. Bull., 108, 392±416.

Elrick, M. and Read, J.F. (1991) Cyclic ramp-to-basin carbon-ate deposits, Lower Mississippian, Wyoming and Montana:

a combined ®eld and computer modeling study. J. Sed.Petrol., 61, 1194±1224.

Fischer, A.G. (1964) The Lofer cyclothems of the Alpine Tri-assic. In: Symposium of Cyclic Sedimentation (Ed. D.F.

Merriam), Bull. Kansas Geol. Surv., 169, 107±149.

Fordham, B.G. (1992) Chronometric calibration of mid-Ord-

ovician to Tournaisian conodont zones: a compilation from

recent graphic-correlation and isotope studies. Geol. Mag.,129, 709±721.

Garland, J., Tucker, M.E. and Scrutton, C.T. (1996) Microfa-

cies analysis and metre-scale cyclicity in the Givetian back-

reef sediments of south-east Devon. Proc. Ussher Soc.,9, 31±36.

Goldhammer, R.K., Dunn, P.A. and Hardie, L.A. (1990)

Depositional cycles, composite sea-level changes, cycle

stacking patterns, and the hierarchy of stratigraphic forcing:Examples from platform carbonates of the Alpine Triassic.

Geol. Soc. Am. Bull., 102, 535±562.

Goldhammer, R.K., Lehmann, P.J. and Dunn, P.A. (1993) The

origin of high-frequency platform carbonate cycles andthird-order sequences (Lower Ordovician El Paso Group,

west Texas): Constraints from outcrop data and statigraphic

modeling. J. Sed. Petrol., 63, 318±360.

Gong, Y.M., Wu, Y., Du, Y.S., Feng, Q.L. and Liu, B.P. (1997)The Devonian sea-level change rhythms in South China

and coupling relationships of the spheres of the Earth.

Acta Geol. Sinica, 71, 212±226 (in Chinese with Englishabstract).

Goodwin, P.W. and Anderson, E.A. (1985) Punctuated aggra-

dational cycles: a general hypothesis of stratigraphic accu-

mulation. J. Geol., 93, 515±533.Hardie, L.A., Dunn, P.A. and Goldhammer, R.K. (1991) Field

and modeling studies of Cambrian cycles, Virginian Appa-

lachians ± Discussion. J. Sed. Petrol., 61, 636±646.

Harland, W.B., Armstrong, R.L., Cox, A.V., Craig, L.E., Smith,A.G. and Smith, D.G. (1989) A Geological Time Scale.Cambridge University Press, Cambridge.

Ji, Q. (1991) Conodont biostratigraphy and mass extinctionevent near the Frasnian±Famennian boundary in south

China. Bull. Chin. Acad. Geo. Sci., 23, 115±127 (in Chinese


Jiang, D.H. (1990) Sedimentary characteristics and evolutionof Middle and Late Devonian interplatform basin in south-

ern Hunan. Sed. Facies Palaeogeogr., 6, 21±29 (in Chinese


Jin, Y.J. (1990) The diachronism of the Dinghechong Forma-tion. J. Stratigr., 14, 131±135 (in Chinese).

Johnson, J.G., Klapper, G. and Sandberg, C.A. (1985) Devo-

nian eustatic ¯uctuations in Euramerica. Geol. Soc. Am.Bull., 96, 567±587.

Koerschner, W.F. and Read, J.F. (1989) Field and modelling

studies of Cambrian carbonate cycles, Virginia Appala-

chians. J. Sed. Petrol., 59, 654±687.Lamaskin, T. and Elrick, M. (1997) Sequence stratigraphy of

the Middle to Upper Devonian Guilmette Formation,

southern Egan and Schell Creek ranges, Nevada. In: Paleo-zoic Sequence Stratigraphy, Biostratigraphy, and Biogeog-raphy: Studies in Honor of J. Granville (`Jess') Johnson (Eds

G. Klapper, M. A. Murphy and J. A. Talent), Geol. Soc. Am.Spec. Pap., 321, 89±112.

Liu, W.J. (1998) Evolution of sedimentation on the South ChinaPlate in the Hercynian-Indosinian stage. J. Chengdu Univ.Technol., 25, 328±336 (in Chinese with English abstract).

Liu, W.J., Zhang, J.Q. and Chen, H.D. (1993) Geological fea-tures of Devonian sedimentary basins in South China and

their deposition and mineralization. Acta Geol. Sinica,

67, 244±254 (in Chinese with English abstract).

McLean, D.J. and Mountjoy, E.W. (1994) Allocyclic control onLate Devonian buildup development, southern Canadian

Rocky Mountains. J. Sed. Res., B64, 326±340.



MontanÄez, I.P. and Read, J.F. (1992) Eustatic control on early

dolomitization of cyclic peritidal carbonates: Evidence from

the Early Ordovician Upper Knox Group, Appalachians.

Geol. Soc. Am. Bull., 104, 872±886.MontanÄez, I.P. and Osleger, D.A. (1993) Parasequence stacking

patterns, third-order accommodation events, and sequence

stratigraphy of Middle to Upper Cambrian platform car-

bonates, Bonanza King Formation, southern Great Basin. In:Recent Advances and Applications of Carbonate SequenceStratigraphy (Eds B. Loucks and J. F. Sarg), AAPG Mem., 57,305±321.

Muchez, P.H., Boulvain, F., Dreesen, R. and Hou, H.F. (1996)

Sequence stratigraphy of the Frasnian-Famennian transi-

tional strata: a comparison between South China and

southern Belgium. Palaeogeogr. Palaeoclimatol. Palaeo-ecol., 123, 289±296.

Odin, G.S., Curry, D., Gale, N.H. and Kennedy, W.J. (1982)

The Phanerozoic time-scale in 1981. In: Numerical Datingin Stratigraphy (Ed. G. S. Odin), pp. 956±960. John Wiley &Sons, New York.

Osleger, D.A. (1991) Subtidal carbonate cycles: Implications

for allocyclic versus autocyclic controls. Geology, 19, 917±920.

Osleger, D.A. and Read, J.F. (1991) Relation of eustasy to

stacking patterns of meter-scale carbonate cycles, Late

Cambrian, USA. J. Sed. Petrol., 61, 1225±1252.Palmer, A.R. (1983) The decade of North American geology:

1983 geologic time-scale. Geology, 11, 503±504.

Pratt, B.R. and James, N.P. (1986) The St. George Group

(Lower Ordovician) of western Newfoundland: Tidal-¯atisland model for carbonate sedimentation in shallow epeiric

seas. Sedimentology, 33, 313±343.

Read, J.F. (1973) Carbonate cycles, Pillara Formation (Devo-nian), Canning Basin, Western Australia. Bull. Can. Petrol.Geol., 21, 38±57.

Read, J.F. and Goldhammer, R.K. (1988) Use of Fischer plots

to de®ne third-order sea level curves in peritidal cycliccarbonates, Early Ordovician, Appalachians. Geology, 16,895±899.

Sadler, R.M., Osleger, D.A. and MontanÄez, I.P. (1993) On the

labelling, length and objective basis of Fischer plots. J. Sed.Petrol., 63, 360±369.

Satterley, A.K. (1996) Cyclic carbonate sedimentation in the

Upper Triassic Dachstein Limestone, Austria: the role of

patterns of sediment supply and tectonics in a platform-reef-basin system. J. Sed. Res., B66, 307±323.

Shen, J.W. and Yu, C.M. (1996) Stratigraphic boundaries on

Devonian carbonate platform and reef complexes in Guilin,Guangxi. J. Stratigr., 20, 1±8 (in Chinese with English

abstract).

Shen, J.W. and Zhang, S.L. (1997) A Late Devonian (Frasnian)

coral-baf¯estone reef at Houshan in Gulin, South China.Facies, 37, 95±108.

Shen, D.Q., Chen, Y.Q. and Yang, Z.Q. (1987) Sedimentary

Facies. Palaeogeography and Controls on Ore Deposits ofthe Qiziqiao Formation (Late Middle Devonian), SouthChina. Geological Publishing House, Beijing, China.

Shen, J.W., Yu, C.M., Yin, B.A. and Zhang, S.L. (1994) Reef

complexes and sequences of Devonian carbonate platformsin Guilin. J. Stratigr., 18, 161±166 (in Chinese with English

abstract).

Strasser, A. (1988) Shallowing-upward sequences in Purbec-

kian peritidal carbonates (lowermost Cretaceous, Swiss and

French Jura Mountains). Sedimentology, 35, 369±383.

Strasser, A. and HillgaÈrtner, H. (1998) High-frequency sea-level ¯uctuations recorded on a shallow carbonate platform

(Berriasian and Lower Valanginian of Mount Salev, French

Jura). Eclogae Geol. Helv., 91, 375±390.

Szuczewski, M., Belka, Z. and Skompski, S. (1996) Thedrowning of a carbonate platform: an example from the

Devonian-Carboniferous of the southwestern Holy Cross

Mountains, Poland. Sed. Geol., 106, 21±49.Tan, Z.X., Dong, Z.C. and Tang, X.S. (1987) On the Qiziqiao

(Chitzechiao) Limestone. J. Stratigr., 11, 77±90 (in Chinese).

Walls, R.A. (1983) Golden Spike Reef Complex, Alberta. In:

Carbonate Depositional Environments (Eds P. A. Scholle,D. G. Bebout and C. H. Moore), AAPG Mem., 33, 445±453.

Wendte, J. (1992) Platform evolution and its control on reef

inception and localization. In: Devonian-Early Mississip-pian Carbonates of the Western Canada Sedimentary Basin:a Sequence Stratigraphic Framework (Eds J. Wendte,

F. A. Stoakes and C. V. Campell), SEPM Short Course, 28,41±87.

Wilkinson, B.H., Diedrich, N.W. and Drummond, C.N. (1996)

Facies succession in pertidal carbonate sequences. J. Sed.Res., 66, 1065±1078.

Wilkinson, B.H., Drummond, C.N., Rothman, E.D. and Died-rich, N.W. (1997) Stratal order in peritidal sequences. J. Sed.Res., 67, 1068±1082.

Wong, P.K. and Oldershaw, A.E. (1980) Causes of cyclicity in

reef interior sediments, Kaybob Reef, Alberta. Bull. Can.Petrol. Geol., 28, 411±424.

Wu, Y., Zhou, H.L., Jiang, T.C., Fang, D.N. and Huang, W.S.(1987) The Sedimentary Facies, Palaeogeography and Rel-ative Mineral Deposits of the Devonian in Guangxi. Guangxi

People's Publishing House, Nanning, China.

Yang, W., Harmsen, F. and Kominz, M.A. (1995) Quantitative

analysis of a cyclic peritidal carbonate sequence, the Middleand Upper Devonian Lost Burro Formation, Death Valley,

California ± a possible record of Milankovitch climatic

cycles. J. Sed. Res., B65, 306±322.

Zeng, Y.F., Chen, H.D., Zhang, J.Q. and Liu, W.J. (1992) Typesand main characteristics of Devonian sedimentary basins in

South China. Acta Sedimentol. Sinica, 10, 104±113 (in

Chinese with English abstract).

Zeng,Y.F.,Liu,W.J.,Chen,H.D.,Zheng,R.C.,Zhang,J.Q.,Li,X.Q.Li, X.Q. and Jiang, T.C. (1995) Evolution of sedimentation

and tectonics of the Youjiang composite basin, South China.

Acta Geol. Sinica, 69, 113±124 (in Chinese with Englishabstract).

Zhang, J.Q. and Zheng, R.C. (1990). Tectonic Pattern, Lithof-acies and Palaeogeography of the Devonian, SouthwestUpper Yangtze Area. Chengdu Polytechnic UniversityPress, Chengdu, China.

Zhong, G., Wu, Y., Yin, B.A., Liang, Y.L., Yao, Z.G. and Peng,J.L. (1992) Devonian of Guangxi. China University of Geo-

science Press, Wuhan, China.

Manuscript received 25 November 1999;revision accepted 9 May 2000.

78 D. Chen et al.


Documents

Long-distance correlation between tectonic-controlled