Tectonophysics, Elsevier

279

182 (1990) 279-300

Science Publishers

B.V., Amsterdam

Joint analyses of calcite twins and fault slips as a key for deciphering polyphase tectonics: Burgundy as a case study 0. Lacombe

t, J. Angelier

r, Ph. Laurent

1 and Ch. Tourneret

2, F. Bergerat

’ Laboratoire de Tectonique Quaniitative, DCpartement de Gkotectonique, U.R.A. 1315 C.N.R.S.,

2

lJmversit6 P. et M. Curie,

Paris (France) .’ Laboratoire de Tectonique et Gkochronologie,

(Received

October

U.R.A. 1371 C.N.R.S., UniversitP des Sciences et Techniques du Languedoc, Montpellier (France) 23, 1989; revised version

accepted

March

5, 1990)

ABSTRACT Lacombe, O., Angelier, J., Laurent, a key for deciphering polyphase

Ph., Bergerat, F. and Toumeret, Ch., 1990. Joint analyses of calcite twins and fault slips as tectonics: Burgundy as a case study. Tectonophysics, 182: 279-300.

Determination of regional paleostresses based on calcite twin analysis has been successfully applied for the first time in the Burgundy platform, France. Paleostress tensors obtained with this method are internally consistent along the profile studied. Moreover, the same directions of paleostresses are independently inferred from fault striation analysis. The results clearly show that the late Mesozoic-Cenozoic evolution of the area has involved polyphase deformation, including (1) NNE-SSW extension, probably late Mesozoic in age, (2) N-S compression related to a major erogenic event in the Pyrenees and Provence, late Eocene in age, (3) E-W extension related to the Oligocene rifting event (the Rhine-SaBne rift system), and (4) WNW-ESE compression related to the Late Miocene westward thrusting of Jura. Calcite twin and striated fault analyses are two complementary tools for paleostress determination: their combined application will allow accurate mapping of paleostress trajectories in intraplate tectonics setting. Moreover, calcite twins can allow stress and paleostress determination when macroscopic features are not observable, which is often the case in very weakly deformed areas and in drill holes.

Introduction During

or in the southern Rhinegraben Laurent, 1988) have been carried

the past ten years,

several

methods

that the method

of

(1984) is

is to demonstrate that results provided by calcite twin analysis are valid and consistent at a regional scale, and that combining fault slip and calcite twin studies provides at the present time the most

Cisternas (1978), Etchecopar et al. (1981) Angelier and Goguel (1979) and Angelier (1984, 1989). These methods have been applied at the regional

reliable way to reconstruct paleostress trajectories in a platform tectonics setting. The area investigated is located on the northeastern side of the Bresse graben (also called the Sai3ne graben), along a profile from the Jura “Avant-Monts” (the outermost units of the Jura mountains) to the southeastern portion of the Paris Basin. This profile crosses the transition zone between the Saline graben and the

scale (e.g., Letouzey, 1986), especially in the European platform (Bergerat, 1985, 1987). At the same time, specific methods for analysing calcite tectonic twins in order to reconstruct paleostress tensors have been improved by Laurent et al. (1981, 1990), Laurent (1984) and Etchecopar (1984). First applications to small areas in the Quercy (Laurent et al., 1990; Tourneret and Laurent, 1990) 0 1990 - Elsevier Science Publishers

by Etchecopar

suitable for the study of polyphase tectonics based on calcite twin analysis. The purpose of this study

paleostress reconstruction using fault slip data sets have been improved following the first successful attempt by Carey and Brunier (1974): Armijo and

0040-1951/90/$03.50

proposed

(Larroque and out, and suggest

B.V

280

,D

Fig. 1. Schematic 2 = sedimentary

map of the West European formations; molassic

platform.

3 = fold-and-thrust” basin;

6 = stratigraphic

The frame shows the location

system

of Jura

contact:

Fig. 2. Geological

map of the area studied.

Prauthoy, Jurassic

St Geosmes

limestones;

Black dots indicate

and Chaumont.

5 = Upper Jurassic

1 = Hercynian limestones:

of the area studied.

axes as dashed

7 = main faults;

Rhinegraben (Bergerat, 1977) (Fig. 1). Along this six sites have been systematically exprofile, amined, which are, from south to north, Taxenne,

Champlitte,

(fold

lines);

8 = thrust (barbs

the data collection basement;

on upthrust

6 = Cretaceous;

sites (from south

2 = Triassic;

I = Hercynian

4 = Oligocene

Montagney, Champlitte, and Chaumont (Fig. 2). This geological setting

rifts;

Prauthoy,

basement; 5 = Miocene

St. Geosmes

is very appropriate

to north),

formations;

PT AL

side).

3 = Lower Jurassic

7 = Cenozoic

LACOMBE

Taxenne, limestones;

8 = main faults.

for

Montagney, 4 = Middle

DECIPHERING

POLYPHASE

TECTONICS

such a study, because:

BY CALCITE-TWIN

(1) numerous

vide fresh outcrops

of limestones,

of striated

could

suitable pled;

faults

quarries

analysis

features

context,

which belong

be carried fected

be sam-

in a well-de-

close to major

of polyphase

the Mesozoic

by strike-slip

rocks

tectonic Rift,

system of the Jura (Figs. 1

a study

out in good

outcrops,

pro-

to the West European

and the fold-and-thrust and 2). Thus,

and

could

and (2) these sites are located

fined geodynamic

FAULT

so that large sets

be collected

for microscopic

AND

tectonics

conditions. formations

and

normal

can

In numerous have been faulting.

af-

In all

cases, sites of data collection display macroscopic evidence of polyphase tectonics, so that the identification of successive events requires sis and data separation.

cluster analy-

SLIP

281

ANALYSES

bears

a mirrored

the untwinned twin

crystallographic

portion

is consequently

amination

to

the twin plane.

The

easily

identifiable

in a polarizing

5). Limestones

which

calcite

with

thus

orientation

across

crystals allow

the

mechanical

random

spatial

sparitic orientation

of very

numerous

in fault

tectonics,

In contrast,

is often

ex-

(Figs. 4 and

abundant

development

twins.

the fault plane

microscope contain

by

imposed

in the rock (i.e.

inherited

surfaces)

direction

and sense are free (and thus depend

as for calcite twins, but the slip only

on the stress state). Experiments on carbonate rocks (Tullis, 1980) have pointed out that: (1) twinning seems to be quite independent of temperature, water and confining pressures,

and deformation

rate (as a conse-

Rock formations are limestones, middle to late Jurassic in age (Bathonian to Oxfordian), with

quence, the normal stress has little or no effect on twinning (Fig. 6) and can be neglected); and (2)

abundant

twinning

interstitial

calcite

and fossil calcitic

in-

fills. These formations have remained horizontal, except close to the Jura folds and thrusts (in the case of Taxenne, the southernmost site). Calcite twin analysis has been applied to samples from the same sites where fault slip data were collected. In addition, tectonic features such as stylolites and tension gashes have been systematically examined and compared with results of fault slip analysis. It is important to note that the paleostress tensors based on fault slip study and those based on calcite computed independently.

twin

Methods for reconstructing Microscopic

analysis

have

been

shear

is possible

stress

twinning “twinning

(the

provided

shear

that

stress

the resolved

projected

on

the

direction) exceeds a critical value, the threshold” of about 10 MPa. The ex-

istence and the significance of this critical shear stress for mechanical twinning, as well as its constant value, have been discussed in Cahn (1964) Tullis (1980) and Laurent (1984). In order to accurately deduce stresses, it is necessary to assume that (1) the crystallographic orientation of the studied sample is random (this point

is easily checked

through

an analysis

of the

paleostresses

scale: calcite twin analysis

The e twinning process in calcite crystals has been well documented in previous papers, so that there is no need to describe it in detail again (Handin

and

Griggs,

1951;

Turner

et al., 1954;

Friedman, 1964). It consists of an intracrystalline deformation that affects calcite crystals in a lowtemperature plasticity domain. This mechanical twinning occurs with a change of form of the crystal by an approximation to simple shear in a particular sense and direction on a given crystallographic plane (Tullis, 1980; Laurent, 1984) (Fig. 3). The resulting twinned portion of the crystal

Fig. 3. Schematic Twin plane

et. C, C’ = optical lamella, pendicular

sketch

of a twin lamella

(i.e. composition

respectively

plane)

axes of the host grain (the plane

to e,). The twinning

tion of motion

of the atoms

sense of shear is indicated

containing direction

located

in a calcite

horizontal,

referred

to as

and of the twinned C and [e,

C’ is per-

: e,] is the direc-

above the twin plane. The

by an arrow (and imposed

network

crystal.

orientation).

by crystal

282

0.

Fig. 4. Thin section (in natural lines which cross-cut

light) of a limestone

the host crystal.

collected

in Burgundy.

The two sets of twin lamellae thus appear

Twin lamellae

observable

as thin straight

appear

here are oblique

generally

as sharply

LACOMBE

ET AL.

definec I straight

at large angles to the thin set :tion and

lines.

stereographic projection of the c axes and of poles of twinned and untwinned e planes as in Fig. 7; the petrofabric is considered as random if their spatial distribution is uniform), and (2) twinning

in the Samples must be accurately oriented field. Three perpendicular thin sections are subsequently cut in each sample. This enables one to of twin obtain the most complete spatial coverage 1

is an irreversible

orientations, and also to make further transformations easier.

be discussed

process

(a characteristic

that will

later).

Fig. 5. Thin section

(crossed

polars)

of a limestone

angles to the thin section plane, are visible as broad

collected bands

in Burgundy.

Two observable

sets of twin lamellae

which differ from the host by abnormal

coloration

geon netrical

oblique

(compare

at small

with ! Fig. 4).

DECIPHERING

POLYPHASE

TECTONICS

BY CALCITE-TWIN

AND

FAULT

SLIP

283

ANALYSES

Ts l

that

0.5..

for a single

about

sample

270 twin planes

The basic

assumption

from stress related age

tensor,

with three

in

thin sections,

are taken into account. is that

twinning

results

to a single homogeneous

such

a way

that

the

aver-

difference

principal stress; 03 = (Jl - ~~(a, = maximum minimum stress) is larger than twice the twinning .

. 0

.

: .

.

.

0

G

+

.

0

0

.

.

.

I\. .”

.

2

.

. .

and

I

account

in order

. I\ :_. : ... ;! . -,;‘; _J Kv ..:0

.

0-

0

0-.

0.

.

. .

-.

.

.

(Laurent

twinned

.

0

threshold

c

.’

agram

of a normal

for a sample collected

T?= resolved represent

shear

stress

observed

stress/resolved at Montagney.

on each plane.

twinned

planes

the final reduced

stress tensor;

observed

planes

tensor. diagram, to +0.5

twinned

twinning dashed

threshold

(see Tourneret

planes are consistent

planes

does not depend agreement

with

represent

with the same planes.

7, ranges

On this

from

-0.5

see Etchecopar, and

separates

and where they are not (below). twinned

untwinned

considerations,

squares

1984).

case, the value of 0.165 is proportional

line (TV= 0.165)

where twinned

triangles

Laurent,

portions

di-

stress;

are consistent

which are inconsistent

from 0 to 1, whereas

(for theoretical

In the present

stress

Small black

which

small black

Small open circles represent on ranges

shear

0” = Normal

to the

1990).

The

of the inversion

of data obtained

through

scopic analysis, which accounts

in order to determine for the largest number

twin system, this tensor fulfils the rethat the value of the computed resolved

shear stress is equal to (or larger than) the threshold for twinned planes, and simultaneously smaller

than

the

threshold

for

the

untwinned

planes. In our study we have adopted the computer technique improved by Etchecopar (1984): the mathematical considerations are exposed in detail in Tourneret and Laurent (1990) and will not be described here. This process yields the directions of the three principal stresses u,, u2 and Us, as well as the value

with final solution

(above),

u3)/(ul - u3) between principal stress values. The separation of different stress tensors from calcite

Note that the distribution on the normal

stress

of

a,,, in

In the sparitic

crystals from both cement of oolitic and fossil infills in micritic limestones. are

of the ratio

twin data will be explained this paper.

with theory.

Rowe and Rutter, 1990). we have only examined

lamellae

micro-

a tensor of twins.

diagram

Several studies have shown that large crystals are twinned more easily than the small ones (Olsson,

twin

into

solution.

process consists

Macroscopic

Calcite

7), are taken the tensor

of the

The grain size of the studied sample is also an important parameter that must be considered.

1974; work,

(Fig.

to compute

As for fault slip data, the working

For each quirement Fig. 6. Example

et al., 1990). All the e planes,

untwinned

examined

present calcite

cipal stresses with

a

three-axis U-stage, (Figs. 4 and 5). The spatial positions of the c axes and poles of e twin planes of calcite crystals are then accurately determined (Fig. 7). Finally, the twinned or untwinned character of each potential twin plane is optically checked. Generally, about 30 crystals are examined, for each oriented thin section; this means

section

of

scale: fault slip analysis

Paleostress reconstruction in fault tectonics is based on the inversion of fault slip data collected in the field (Angelier, 1984, 1989). It consists of determining characterized corresponds

limestones

in another

Cp= (uz -

the average stress state that can be by a stress tensor. This stress tensor to the orientation of the three prin(I,, u, and

uj and to the ratio

@=

(ez - e3)/(e, - 03). The basic assumption is that the orientation of a given fault plane does not necessarily depend on the orientation of the principal stresses. For any inherited discontinuity, only the relative motion of the two blocks on both sides of the fault plane is significant. Moreover, this assumption of slip parallel to shear stress on a given plane is made

284

for both inherited

0

and newly

allows one to take into account well as conjugate

formed

faults.

This

interactions

non-conjugate

as

of average

shears. In such an analysis,

fault

rock

and more generally stress

axes and

are considered

LACOMBE

all the variations

of the ratio

to be negligible.

need to be controlled

ET AL.

@ inside Exceptions

by means of careful,

qualita-

tive observations, According

to the basic model,

tions should

correspond

of the greatest direction

shear

stress

(Bott,

and sense of motion

known,

computing

slickenside

to the direction

linea-

and sense

1959).

If the

on fault planes

a particular

solution

stress

tensor)

are

with four

unknowns

(the reduced

is possible.

If a tensor any tensor

T is a solution of the inverse problem, kT + II (with k > 0) is also a solution

(see Angelier, 1989); as a consequence, the final tensor will be an affine function of T. In order to compute

T and

determination, monly

to evaluate

a function

depends

the quality

F is defined;

on the angle

between

of the F com-

the actual

slip (observed from striations on the fault plane) and the computed shear stress. For the present study, a “direct inversion method” has been adopted with the function F referred to as S, in Angelier (1984). The optimization process consists of minimizing F according to the least-squares method, so that the minimal value of F corresponds to the best average stress tensor. One finally obtains T and thus the directions of the three principal stress axes ei, u2 and u3, as well as the value of the ratio @. Separation

of different paleostress

tensors

Using striated microfaults Despite the horizontal attitude of the rock formations we studied (showing that the total deformation remains very small), the most striking character of the area of interest is the predominance of polyphase tectonics. In fact, all examined sites displayed superposed structures, so that a

Fig. 7. Stereographic poles

of twinned

twinned

planes

collected The spatial

projection planes

(C, black

at Taxenne

close to uniform,

dots)

using

distribution

of C-axes (A, empty

(B, empty

circles),

for crystals

Schmidt’s

equal

of poles of e planes

showing orientation

triangles)

and poles of unfrom area

the sample projection.

and of c-axes is

that there is no preferential in the sample.

crystal

DECIPHERING

POLYPHASE

Fig. 8. Example different

TECTONICS

of paleostress

paleostress

slickenside

lineations

Paleostress

directions

states. as dots

BY CALCITE-TWIN

analysis Fault

with striated

slip data:

with double

(without

any selection).

consistent

with N-S

= minimal

arrows

and

Predominantly

strike-slip

separation of different quired (Fig. 8).

tectonic

normal E-W

SLIP

locality:

lower

faults

faults consistent

was

consistent

with WNW-ESE

A qualitative separation Qualitative separation. of successive tectonic events can be carried out according to three main criteria: (1) The relative chronology of tectonic features is usually based either on the identification of successive movements on a simple structure (such as superposed slickenside lineations of different

with

of data faults

ones (centrifugal-normal; compressive

shown

with NNE-SSW normal

of subsets

projection)

= maximal

or compression

D. Predominantly

re-

separation

hemisphere or simple

stars with five branchs of extension

285

ANALYSES

at Prauthoy

(Schmidt’s

extension.

events

FAULT

(left- or right-lateral)

empty

stress +. Direction

B. Predominantly compression

faults

diagrams

from fault slip analysis:

stress u2; three branchs

AND

extension.

compression

with

and NNE-SSW

and

centripetal-reverse). = middle

A. Entire data set

C. Predominantly

consistent

to

curves

stress 0,; four branchs

by large black arrows.

faults

corresponding

as thin

NW-SE

strike-slip

faults

extension.

E.

extension,

directions on a fault plane, tension gashes reactivated as normal faults, and so on), or on the intersections of distinct features (e.g., a tension gash cut and offset by a fault). (2) Normal faults, reverse faults and strike-slip faults are assumed to be systematically distributed in separate subsets, in order to distinguish stress states that correspond to different mechanisms of brittle deformation. (3) Previous

reconstructions

of regional

tectonic

0.

evolution

(Bergerat,

1985,

1987) are

account

in order

to propose

different

tectonic

tensors.

Quantitative

stress tensors

is made

using

and Manoussis

The

separation

and related

an algorithm

(1980) and Angelier

a preliminary

in Angelier

(1984). Where

the numerical

sis in terms of tensor computation unless

of the

classes of data

exposed

fault data sets are polyphase, results,

into of the

analy-

may yield poor

qualitative

selection

(commonly based on the use of relative chronology criteria) is carried out in order to identify the major tectonic events. However, more complex numerical polyphase neously

methods may also solve this problem of fault data by determining simultatwo or several

stress

tensors

data set (Angelier and Manoussis, copar et al., 1981; Angelier, 1984) tion of each datum of individual

misfits

that several twinned

depending

for a given 1980; Etchethe classifica-

on the comparison

planes

two or three different for this difficulty,

separation.

different

taken

a succession

solution

space

number

of random

termination consistent

LACOMBE

ET AL.

can be consistent

with

tensors.

To partially

a systematic

exploration

is undertaken, trials.

using This

of the maximum

of the

a very large

leads

to the de-

percentage

of twins

with each stress tensor. Thus, this meth-

odology

allows the determination

tensors,

but is less accurate

be expected nally,

account

with

striated

that at the present

of several stress

than that which could microfaults.

Note,

stage of our knowledge,

the optical analysis of calcite twins does not provide any evidence of twinning successions, and hence of the relative chronology between the different events. The only way to relate the different episodes

of twinning

to tectonic

events

known

from geological data still consists of comparing stress tensors obtained from calcite twin analysis with those independently

obtained

from fault slip

analysis.

with the tensor. Results: paleostress

evolution in Burgundy

Using culcite twins As mentioned in an earlier section, the inverse method proposed by Etchecopar (1984) and set

Four main been defined

regional paleostress systems in the area of Fig. 2. The

out in detail in Tourneret

characteristics

of the tensors

been

adopted

in our study.

and Laurent

(1990) has

As most of our sam-

ples are concerned with polyphase deformation, determination of a single stress tensor compatible with the entire data set (twinned and untwinned planes) is meaningless. For polyphase samples, the process consists of first searching for an initial solution by a random selection of many tensors which are applied to all the data, and then determining the optimal percentage planes consistent with the request This percentage is first arbitrarily then modified

fi-

of twinned stress tensor. chosen, and

to reach the best solution

according

to a given penalization function (referred to as F in Tourneret and Laurent, 1990). Second, the tensor solution is optimized with regard to the subset of data. When a first tensor is determined, the twinned planes consistent with it are withdrawn, and the process is repeated. We have to bear in mind that even if the existence of untwinned planes is an important constraint for the determination of all stress tensors, it is clear

computed

have main

from fault

slip analysis and from calcite twin analysis respectively are shown in Figs. 9-13, and in Tables 1 and 2.

NNE-SS

W extension

Tectonic analysis clearly the existence tion of extension, faults

which

trend

using both methods shows of a NNE-SSW first direc-

mostly

characterized

between

85”

and

by normal 115”

(in

some cases 60 “-115 o ). The trend of the principal minimal stress, u3, independently computed from faults and calcite twins, is nearly constant

at 20 O-

35” (60” in one case; Fig. 9). Scarce relative chronology criteria suggest that this extension event occurred prior to a N-S compression, for which we proposed a late Eocene age. But the age of this extension cannot be exactly defined in the area of interest from stratigraphic data because, in all sites studied, rock formations

mxx+muNG

FOLYPHASETECT~~~ICS

RY CAI.CITE-TWIN ANI) FALILT~LIPANALYSES

287

0. LACOMBE

288

P

r

ET AL

Fig. 11. NW-SE

extension

inferred

from microstructural

analysis

on the northeastern

w

side of the Bresse rift, using fault slips and calcite

\

twins.

I

Caption as for Fig. 9.

290

0. LACOMBE

***.**...**V)...t, I***.,..+*~

..***+A**,.~*