Magma flow in sub-aqueous rhyolitic dikes inferred from magnetic fabric analysis (Ponza Island, W. Italy

Magma flow in sub-aqueous rhyolitic dikes inferredfrom magnetic fabric analysis (Ponza Island, W. Italy)

Charles Aubourg a,*, Guido Giordano b, Massimo Mattei b, Fabio Speranza c

a Tectonique UMR 7072, Department of Earth Sciences, University of Cergy Pontoise, 8, Le Campus, 95031 Cergy Cedex, Franceb Dipartimento di Scienze Geologiche, University di Roma Tre, L. go S. Leonardo Murialdo 1, 00146 Roma, Italy

c INGV, Italy

Abstract

We have studied seven rhyolitic dikes from the �8 km2 Ponza island (Italy) in order to retrieve the sense of magma flow by using

magnetic fabric. These late Pliocene dikes were part of a sub-aqueous dome complex and were emplaced in wet and unconsolidated

rocks. Some dikes are inherited from normal faults. Dike thickness ranges from 3 to 100 m, where they are reintruded. We sampled

paired margins of four dikes and one margin in three dikes. In addition, we performed a profile in a 3 m thick dike. The magnetic

mineralogy of the rhyolite consists of magnetite and likely maghemite. Biotites also participate marginally to the magnetic fabric.

The magnetic fabric is generally well defined and flow-related as the magnetic foliation is close to the dike plane. The magnetic

foliation is generally imbricated respect to the dike plane. Conversely, in the center of the dike, the magnetic foliation is poorly

defined whereas the magnetic lineation is well defined. Optical inspection and textural analysis of seven representative thin sections

of rhyolite from the dikes demonstrates a good parallelism between opaque grains and magnetic lineation. We determine the sense of

flow by using the geometry of both magnetic lineation and foliation. The magnetic lineation in six out of seven dikes is sub-vertical,

suggesting vertical flow. However, a main result of our study is that the lineation by itself is not sufficient to image the sense of flow,

being the lineation also related to the intersection of magnetic foliations. The analysis of the imbrication of magnetic foliations is

instead a much better tool to image the sense of flow. The sense of flow inferred with imbrication of magnetic foliation shows a good

agreement with the geology. The dikes feeding small cryptodomes have shown downward flow that is interpreted as related to

deflation after emplacement. Horizontal flow has been imaged in the thickest of the sampled dikes suggesting the presence of domes

at shallow depth feeding an array of concentric dikes.

� 2002 Elsevier Science Ltd. All rights reserved.

Keywords: Magnetic fabric; Dike; Rhyolite; Ponza; Flow; Imbrication

1. Introduction

Imaging the sense of magma flow in dikes is an im-

portant task for volcanology. There are different man-

ners to find the magma flow direction, depending on

field conditions. Dike epiphyses or dike segmentation

are often encountered in the field. However, these fea-tures are also related to the crack propagation. In ad-

dition, they are observed within 2D exposures, which

limits the accuracy to resolve the flow direction. Gas

bubbles, elongated vesicles or mineral lineation, are

good indicators, but their occurrence is rather rare along

the margins. Finger groove or strieas are also used (e.g.

Varga et al., 1998) but their meaning have been ques-

tioned (Baer, 1995). It appears that textural analysis of

grains remains one of the best techniques to infer the

flow direction. Thin section analysis is a classical way to

do this but it is rather time-consuming (Varga et al.,

1998). By contrast, the magnetic fabric provides a fast

and accurate 3D image of the texture of 103 up to 1012

magnetic grains for a standard 10 cc samples (Hrouda,1982). Statistical processing allows gathering about 100

cc of oriented material within a representative section of

the dike. Magnetic fabric is characterized qualitatively

by a magnetic foliation and a magnetic lineation. In

addition, the anisotropy parameters quantify the degree

of anisotropy and the shape (oblate to prolate) of the

magnetic fabric ellipsoid. The magnetic fabric is gener-

ally related to the magmatic flow when the magneticfoliation is close to the dike plane (Hrouda, 1982;

Rochette et al., 1991). Rochette et al. (1999) reviewed

the occurrence of inverse magnetic fabrics, which are

Physics and Chemistry of the Earth 27 (2002) 1263–1272

www.elsevier.com/locate/pce

* Corresponding author. Tel.: +33-1-3425-4981; fax: +33-1-3425-

4904.

E-mail address: [email protected] (C. Aubourg).

1474-7065/02/$ - see front matter � 2002 Elsevier Science Ltd. All rights reserved.

PII: S1474-7065 (02 )00113-4

mail to: [email protected]

characterized by magnetic foliations perpendicular to

the dike plane. In basaltic dikes, the proportion of

normal magnetic fabric ranges from 40% to 90% and the

cause of inverse magnetic fabric is still debated. The

magnetic lineation is generally used as a reliable indi-

cator of the magma flow direction (e.g. Tauxe et al.,

1998).A breakthrough in the dike analysis came from

Knight and Walker (1988) when they proposed to study

the imbrication of magnetic fabrics along the dike

margins (Blanchard et al., 1979). The imbrication of

phenocrystals as well as magnetic fabric provides the

sense of flow, whereas magnetic fabric from the center of

the dike gives only the direction of flow. It should be

noted that since this publication, few studies dealt withthe imbrication of magnetic fabrics along the margins

(Tauxe et al., 1998; Varga et al., 1998; Rochette et al.,

1999; Bervera et al., 2001). Tauxe et al. (1998) discussed

the advantages of using the imbrication of magnetic

fabric. They proposed a method of sampling, and some

statistical criteria to reject or not a magnetic fabric.

Their approach is entirely based upon the mirror geo-

metry of magnetic lineation along the margins. Bycontrast Moreira et al. (1999) proposed to use also the

imbrication of magnetic foliations along the margin in

order to infer a direction and a sense of flow. In lava

flow, the imbrication of magnetic foliation has been also

successfully applied in hyaloclastics to determine the

sense of flow (Hillhouse and Wells, 1991). Magnetic

lineation, indeed, can be meaningless as it can result

from various mechanisms such as stretching and cren-ulation of magnetic foliation.

In this study, we report a magnetic fabric investiga-

tion in rhyolitic dikes from Ponza that intruded wet and

unconsolidated clastic hyaloclastite rocks in sub-aque-

ous environment during the late Pliocene (De Rita et al.,

2001). Ponza Island offers a wide variety of rhyolitic

dikes and continuous outcrops along its coastal cliffs. De

Rita et al. (2001) have interpreted the Ponza dikes aspart of sub-aqueous domes, controlled by the local

stress field, rather than regional, implying intrusion

from a local source, that is the core of the dome. We

sampled margins and center from seven dikes (114

samples) and studied both magnetic foliation imbrica-

tion and lineation to infer the sense of magma flow and

to test the De Rita et al. (2001) interpretation.

2. Geological frame work and sampling

The island of Ponza (�1 km wide, �7 km long) is

part of the Pontine Archipelago, which consists of four

islands, aligned along a regional NE-trend. The volcanic

and tectonic framework of Ponza has been discussed by

De Rita et al. (2001). The islands are mostly made of latePliocene (4.7–1.9 Ma), high-K, calc-alkaline rhyolithic

lavas outpoured during submarine vents. The most

common rocks (80%) present at Ponza are co-genetic

hyaloclastite and coherent dikes. Hyaloclastites and

dikes are believed to be cogenetic. Dikes are broadly

distributed concentrically around three main domes

(Fig. 1). Several lines of evidence indicate that dikes

emplaced into a loose water saturated water environ-ment: the 3D irregular shape of dikes, the presence at

the dike margin of zones of thin obsidian perlite grading

outward into hyaloclastic breccia. Dikes are not recti-

linear but irregular and concentrically arranged around

the dome centres. Dikes are ubiquitously associated to

extensional deformation, which generally does not cut

the rhyolitic dikes. This extensional deformation has

been related to local stress induced by magma intrusion,because it swings around following the local dike trend.

Faults appear as narrow anastomosed shear zones ra-

ther than discrete planes, suggesting that this deforma-

tion developed in still unconsolidated sediments (cf.

Maltman, 1994). Several of the dikes are set in small

grabbens with extensional faults converging toward the

dike (Fig. 2).

We sampled seven dikes trending NE–SW to E–W,with variable thickness from 3 to �100 m (Fig. 1, Table

1). We drilled paired margins at dikes A, C, and E in

order to check imbrication. The dikes A and C feed

upward two small cryptodomes. Although a 2D expo-

sure along a cliff poorly constrains the flow direction,

cryptodome suggests vertical flow at dikes A and C.

From Fig. 2, we see a set of normal faults that converge

toward dike C. This pattern is observed frequently atPonza (De Rita et al., 2001). Close to dikes A and C, we

sampled dikes B and D (Fig. 1). Only the southern

margin has been drilled at dike B. At dike D, we drilled

the center of the dike (Fig. 2), where steeply dipping flow

bending with steep dipping striaes is visible, suggesting

vertical flow. The NE-trending dike E is the thickest one

(�100 m) drilled in this study, and it has been reintruded

by a E–W 30 m thick dike F. At dike E, steep striaes andapparent bending of the magmatic foliation along the

thick hyaloclastic southern margin suggest upward flow.

A profile (36 cores) along the 3 m thick dike G (Table 1)

has been performed in order to determine the magnetic

fabric zonation of the dike. In addition, we measured a

profile of magnetic susceptibility in situ using a Kapp-

ameter Kly5 (Agico Ltd.). This well exposed dike is

parallel to a set of normal faults, and for that reason, webelieve that dike G is inherited from those faults.

3. Magnetic fabric

To describe the magnetic fabric, we measured the

anisotropy of low-field magnetic susceptibility (AMS) of

114 magnetically oriented cores using a spinner Kly3(Agico Ltd.). When AMS axes are statistically defined,

1264 C. Aubourg et al. / Physics and Chemistry of the Earth 27 (2002) 1263–1272

the magnetic lineation corresponds to the maximum axis

while the minimum axis is the pole of magnetic foliation.

We applied mean tensorial analysis (Lienert, 1991) in

order to derive magnetic lineation and foliation at the

scale of the margin or the core of the dike. In Ponza

rhyolites, we observed quartz, felspars, biotites and

unidentified opaques. Biotites can have significant

paramagnetic susceptibility. However, given the strongmagnetic susceptibility observed in the dikes (10�4 to

10�2 SI), one can expect little contribution of biotite to

AMS (Rochette et al., 1992). To identify the ferromag-

netic grains, we monitored the evolution of low-field

magnetic susceptibility versus temperature (K–T curves),

from room temperature to 700 �C, under air. We do not

observe significant decrease of magnetic susceptibility in

the low temperature range. This means that paramag-

netic contribution is negligible. Besides, a Curie tem-perature close to 580 �C characterizes the main mineral.

Fig. 1. Geologic map of Ponza (from De Rita et al., 2001). Location of sampled dikes is annotated with letters from A to G. Cross sections illustrate

the deep geometry of dikes.

C. Aubourg et al. / Physics and Chemistry of the Earth 27 (2002) 1263–1272 1265

This supports the occurrence of poorly substituted

magnetites in rhyolitic dikes (Fig. 3A). A secondary

phase is commonly observed, with descending slopes

between 200 and 400 �C (Fig. 3A). We believe that thismineral is titanomaghemites because it is destroyed

during heating. Note that we cannot rule out the occur-

rence of titanomagnetites. The magnetic susceptibility

profile measured in situ at dike G reveals a characteristic

V-shape, with a maximum value in the center of the dike

(Fig. 3B).

Two common parameters describe the shape of AMS

ellipsoid; the corrected degree of anisotropy P 0 and theshape parameter T (Jelinek, 1981; see Table 1 for defi-

nition). Individual P 0 are never higher than 1.06 and the

mean P 0 that results from tensorial analysis is lower than

1.04 (Fig. 3C, Table 1). The anisotropy is therefore

weak. The shape of AMS ellipsoid is triaxial in average

(T close to 0, Fig. 3C).

The orientation of AMS axes with their confidence

ellipse at 95% is shown (Fig. 4), together with the margin

orientation, and striaes directions when there are docu-

mented. At dike G, we grouped data in five subsets:

along the paired margins (two sets), from the center,

and between the center and margins. We consider first

the magnetic foliation and then the magnetic lineation.Among the 10 margins sampled, eight provide well-

defined magnetic foliation with tight clustering of mini-

mum axes of AMS. Only the two southern margins of

dikes A and C are poorly constrained, although the ten-

sorial magnetic foliation is roughly close to the margin.

By contrast, the magnetic foliation from the center of the

dikes is poorly constrained. This is observed at dikes D

and G (Fig. 4). One striking feature observed throughoutall dikes is that magnetic foliation is generally parallel to

its respective margin. Only the magnetic foliation along

the northern margin of dike G has an obliquity of about

50� (Fig. 4). The magnetic fabric is thus normal, likely

magmatic in origin, and no inverse magnetic fabric is

readily observed. We had the opportunity to measure

with a home-designed goniometer the magmatic foliation

visible at the core scale along the northern margin of dikeA (Fig. 4). It appears that magmatic and AMS foliations

are similarly imbricated with respect to the dike margin.

Magnetic foliations of paired margins of dikes A, C and E

are symmetrically imbricated (Fig. 4). We consider now

the magnetic lineation, which is the most common feature

that is used to infer the flow direction. The magnetic

lineation is well defined in all dikes except along southern

margins of dikes B and C. When striaes are present,magnetic lineation is parallel (dikes D and F). The center

of the dike shows a well-defined magnetic lineation

whereas magnetic foliation is poorly defined. In order to

compare the magnetic lineation to the texture of opaque

grains, we analyzed eight thin sections from dikes A

(north margin), C (south margin) and G.

We used a similar approach to that proposed by

Callot et al. (2001). All steps of the treatment are sum-marized in Fig. 5. We cut thin sections parallel to the

magnetic foliation. From each thin section, we have

selected three representative circular views. We filtered

the image in order to enhanced dark mineral (biotites)

and opaque grains. We processed each view using IN-

TERCEPT software (Launeau and Robin, 1996). IN-

TERCEPT determines the orientation of the contacts

between different phases. In this study, we analyzed onlythe black level of black and white photo of the thin

section. Through a Fourier analysis, INTERCEPT de-

termines a rose of direction and a characteristic shape.

The long axis of equivalent shape can be compared di-

rectly to the magnetic lineation. We found a good par-

allelism between the long axis of the equivalent shape of

opaques phases and magnetic lineation at dikes 5P and

7P. The angles are lower than 12� (7� in average) and 14�(9� in average) for respectively dikes A and G. The

Fig. 2. The photo shows dikes D, C and B, and contemporane-

ous hysaloclastics exposed at Punta Bianca. White circles point (1P–

3P) are sampled area. The thinnest dike C is topped by a small

cryptodome, suggesting vertical flow. There is a gradation from mar-

gins of rhyolitic dikes and hyaloclastics at dikes D and B. Normal

faults are typical in this island and there are contemporaneous of dike

injection.


agreement is not as good for dike C. The angles oscillate

between 6� and 59�, depending on thin section, and also

selected circular view. We observe discrepancies when

confidence angles (angle E12) provided during AMS

measurement are greater than 30�. This optical ap-

proach suggests that magnetic lineation truly reflects a

preferred orientation of opaque phases (biotite,

magnetites, maghemites). In summary, the magneticfabric in rhyolitic dikes from Ponza is well defined, with

no inverse magnetic fabric. This contrasts with a higher

proportion of inverse fabrics in basaltic dikes.

4. Discussion

Our study constitutes, to the best of our knowledge,

the first attempt to retrieve magma flow by means ofAMS in rhyolitic dikes. Note however that Rochette

Table 1

Tensorial AMS results

Dike

site

Margin N km L F P 0 T D, I E1–2 E2–3 E3–1

kmax kint kmin

A

5P North 9 628� 1438 1.008 1.010 1.012 0.110 74,50 206,30 310,25 10 6 6

6P South 6 6319� 601 1.009 1.006 1.010 �0.199 157,69 257,4 348,21 32 48 25

B

4P North 9 7008� 1291 1.003 1.019 1.016 0.725 223,60 41,31 311,0 37 9 12

C

1P Norh 7 2590� 646 1.009 1.013 1.015 0.180 112,57 220,11 316,30 18 5 13

2P South 8 1129� 1548 1.004 1.014 1.013 0.553 68,10 312,68 161,19 34 39 8

D

3P Center 9 4349� 2431 1.029 1.005 1.025 �0.702 1,79 104,2 195,10 6 42 10

E

8P North 12 240� 64 1.010 1.007 1.012 �0.175 216,82 37,8 307,0 8 13 7

9P South 10 309� 228 1.013 1.008 1.015 �0.236 92,67 263,23 354,3 9 12 8

F

10P North 11 549� 503 1.012 1.011 1.016 �0.043 163,58 266,8 1,31 29 23 12

G

7Pn North 8 1589� 898 1.006 1.003 1.006 �0.332 321,63 217,7 124,26 12 13 11

7Pnc N–C 4 2766� 318 1.007 1.002 1.006 �0.554 21,48 210,43 116,4 27 67 24

7Pc Center 8 3302� 777 1.005 1.003 1.005 �0.249 281,88 12,0 102,3 14 35 9

7Psc S–C 4 1642� 626 1.006 1.003 1.006 �0.332 17,61 225,26 129,12 48 79 25

7Ps South 9 957� 450 1.011 1.005 1.011 �0.373 313,64 60,8 154,25 11 17 10

N¼ number of specimens.

km ¼ ðkmax þ kint þ kminÞ=3 (mean susceptibility, in 10�6 SI).

L ¼ kmax=kint.

F ¼ kint=kmin.

P 0 ¼ expf2½ðg1 � gÞ2 þ ðg2 � gÞ2 þ ðg3 � gÞ2�g1=2(corrected anisotropy degree; Jelinek, 1981).

Tj ¼ 2ðg2 � g3Þ=ðg1 � g3Þ � 1 (shape factor; Jelinek, 1981).

g1 ¼ ln kmax; g2 ¼ ln kint; g3 ¼ ln kmin; g ¼ ðg1 þ g2 þ g3Þ=3.

E1–3, E2–3, E1–2: semi-angles of the 95% confidence ellipses around the principal susceptibility axes.

Fig. 3. (A) Thermomagnetic curve (K–T: magnetic susceptibility versus temperature) under air. The main magnetic carrier is poorly substituted

magnetite with a Curie temperature close to 580 �C. (B) secondary magnetic carrier is commonly detected, and it characterized by a Curie tem-

perature in the range 200–400 �C. We interpret this mineral as titanomaghemite. Note that no paramagnetic behavior is suggested at low tem-

perature. (C) Magnetic susceptibility profile at dike G measured in situ with a Kappameter Kly5. (D) Jelinek T–P 0 type diagram (Tarling and

Hrouda, 1993) with average anisotropy parameters P 0 and T resulting from arithmetical mean.


et al. (1999) reported magnetic fabrics of two rhyolitic

dikes from Yemen. We found well-defined AMS, with

no inverse magnetic fabric. Thus, the magmatic origin ofAMS makes it possible to infer magma flow direction,

and also the sense of flow in rhyolitic dikes. Magnetic

lineation is often taken as flow direction indicator (e.g.

Tauxe et al., 1998). In our study, the magnetic lineations

are steep in average, and parallel to striaes. When we

drilled the center of the dike (dikes D and G), the

magnetic lineation is steeply dipping. All these obser-

vations plead for general steep flow in good accordance

with field observation (cryptodome, striaes). However,

we have information about direction, but not about thesense of flow. We now consider imbrication of magnetic

foliations. Magnetic foliations along margins are also

well defined, and apparently imbricated. To better un-

derstand how work imbrication, we present the rela-

tionship between an idealized imbricated magnetic

fabric and upward flow (Fig. 6). When paired margins

are drilled, and magnetic fabrics are fairly defined, we

Fig. 4. Magnetic fabrics of the seven dikes are shown in situ (Equal-area projection in a lower hemisphere). The dike margin is plotted as well as the

mean magnetic foliation. Striaes, when present are also indicated.


propose to rank the geometry of imbricated magnetic

fabric in three types. Type I (Fig. 6A) is the expected

geometry when magnetic foliation and lineation are

flow-parallel. Note that the sense of flow is normal to

the intersection of magnetic foliations. Type II (Fig. 6B)

is characterized by a magnetic lineation perpendicular to

the flow direction. Magnetic lineation is here parallel to

the intersection of magnetic foliations. Types III (Fig.

Fig. 5. Textural analysis of opaques grains (biotite, magnetite, and probably maghemite using INTERCEPT software (Launeau and Robin, 1996).

See Callot et al. (2001) for description of the method. Two representative examples are presented. (A) The magnetic lineation is poorly defined (30�)and the opaque grains are distinctly oblique to the magnetic lineation. (B) The opaque grains aligned here fairly with the well-defined magnetic

lineation.

Fig. 6. Idealized imbricated magnetic fabric in case of upward and vertical flow. Type I. Magnetic fabric geometry when magnetic lineation and

imbricated foliations are parallel to the flow. Type II. Magnetic fabric geometry when magnetic lineation is perpendicular to flow. Type III. Magnetic

foliations are imbricated but dike trend is not bisecting the magnetic foliations.


6C) shows imbricated magnetic foliation, but the dike

trend is not bisecting magnetic foliations. Such a scheme

can result from various mechanisms such as poorly de-

fined magnetic fabric, poor orientation of dike margin,

or simple shear (Rochette et al., 1991). We present now

our magnetic fabric data in dike coordinates (Fig. 7) asproposed by Rochette et al. (1991). This consists to tilt

margins to the vertical, and rotate paired margins to fit

the same direction. Dikes A and C exhibit similar im-

brication of magnetic foliation (Fig. 7). Magnetic lin-

eations are sub-vertical. However, in both cases the

geometry of the paired magnetic foliations suggests

downward vertical flow, as the two foliations point

downward. This rather unexpected flow direction can beexplained taking into account that dikes A and C are

feeder to small cryptodomes (Fig. 2). The presence of

cryptodomes indicates that magma buoyancy was at the

point of annihilation. Furthermore, once the crypt-

odome was filled, i.e. the feeding of magma stopped, a

partial emptying would have been possible, causing a

limited and late stage re-flux within the dike. The syn-

emplacement normal faults developed in the hyaloclas-tite surrounding dikes A and C (Fig. 2) might also be

related to such a collapse. We suggest therefore that the

downward flow in the case of dikes A and C is likely,

and recording the late stage deflation of the crypto-

domes at the end of magma feeding. Dike E (Fig. 7)

shows a type II geometry, with a SW horizontal flow

suggested by the geometry of the magnetic foliations.

The magnetic lineation is sub-vertical, i.e. perpendicularto the magma flow. The occurrence of sub-horizontal

flow in dikes is not uncommon and can be related to the

relative location of the dike respect to the feeding system

(magma chamber or cryptodome) and also relate to the

buoyancy of magma. Dike G shows a type III geometry.

The imbrication is well defined, but it does not make a

V-shape within the dike plane. At dike G, a set of

conjugate extensional faults is present. The trend of thedike is about N260-70 and parallel to the north-

ward dipping set of faults. The southward dipping set

of faults displaced the southern margin of the dike, but

not its northern margin, indicating deformation is syn-

emplacement and that the faults acted accommodating

the emplacement of magma. In addition, the magnetic

fabric is very well defined thus the obliquity is well

constrained. We believe therefore that the obliquity ofthe magnetic foliation may result from a simple shear.

To account for the large angle of obliquity at dike G,

one can expect that simple shear acted when magma was

still flowing. Therefore we propose that the magnetic

foliation in inherited dikes cannot be used to determine

the sense of flow, but it gives an important information

on the kinematic of the intrusion. At dike B, imbrication

of magnetic foliation along southern margin suggestseastward horizontal flow, in contrast with steep flow

suggested by magnetic lineation. At dikes D and F,

magnetic foliations cannot be used to infer flow.

5. Conclusion

Our investigation of rhyolitic dikes, emplaced in sub-aqueous environment, shows convincing results. We

summarized here the key points:

Fig. 7. Magnetic fabric of dikes where the paired margins were drilled are shown in dike coordinates (see text). A cartoon illustrates the flow inferred

in the dike.


• Magnetite is the main magnetic carriers, together

with titanomaghemite or titanomagnetite, in lower

proportion. The paramagnetic contribution of biotite

is not significant in these rhyolitic dikes;

• A nice V-shape of magnetic susceptibility is observed

through a 3 m thick dike;

• The degree of anisotropy is rather low (P 0 < 1:06) andthe shape of AMS ellipsoid is triaxial in average

(T � 0). These characteristics are also encountered

in basaltic dikes.

• Magnetic fabric is well defined in about 80% of the

samples. We do not see any evidence of inverse mag-

netic fabric.

• Along the margins, magnetic foliation is always well

defined and imbricated while within the center, mag-netic foliation is poorly defined. We found a good

correspondence between magnetic foliation and mag-

matic foliation measured from the core at one mar-

gin.

• By contrast, magnetic lineation is well defined in the

center of dike, but can be scattered along margin.

Magnetic lineations are parallel to striaes. Optical in-

vestigation shows that magnetic lineation are fairlyparallel to alignments of biotites and opaques grains.

• When taken only magnetic lineation, we found steep

flow direction in all dikes.

• When taken into consideration the imbrication of

magnetic foliation, we found debatable results.

Downward flow is found in small dikes A and C

topped by small cryptodomes. We interpret the

downward flow as related to a late-stage re-flux with-in the dike, related to magma buoyancy annihilation

and the end of magma feeding. Horizontal flow is

suggested at the thickest dike E, in contradiction with

the steep magnetic lineation. Horizontal flow is in

agreement with a poorly buoyant magma fed laterally

by a shallow source as suggested by the field volca-

nology of Ponza.

• Simple shear is expected in the inherited dike G be-cause imbricated magnetic foliations do not show

mirror symmetry in respect to the dike trend.

Our study finally suggests that magnetic fabric

method is a very promising method to study magma

flow in rhyolitic dikes. We propose that magnetic lin-

eation, although parallel to the linear preferred orien-

tation of opaques grains, cannot be in all circumstancesparallel to the flow.

Acknowledgements

We acknowledge Prof. D. De Rita for useful discus-

sions. P. Rochette and E. Pueyo-Morer provided helpful

reviews and constructive comments. C.A. thanks the

technical and scientific staff from the Earth Science

Department at the University of Roma Tre for their

kind help. This work was funded by Bilateral program

between the University Roma Tre and the University of

Cergy Pontoise.

References

Baer, G., 1995. Fracture propagation and magma flow in segmented

dykes: Field evidence and fabric analyses, Makhtesh Ramon,

Israel. In: Baer, G., Heimann, M. (Eds.), Physics and Chemistry of

Dykes. Balkema, Rotterdam, pp. 125–140.

Bervera, E., Walker, G.P.L., Canon-Tapia, E., Garcia, M.O., 2001.

Magnetic fabric and inferred flow direction of dikes, conesheets

and sill swarms, Isle of Skye, Scotland. Journal of Volcanology and

Geothermal Research 106, 195–210.

Blanchard, J.P., Boyer, P., Gagny, C., 1979. Un nouveau crit€EEre de

sens de mise en place dans une caisse filonienne: le ‘‘pincement’’ des

min�EEreaux aux �EEpontes. Tectonophysics 53, 1–25.

Callot, J.P., Geoffroy, L., Aubourg, C., Pozzi, J.P., Mege, D., 2001.

Magma flow of shallow dykes from the East Greenland volcanic

margin inferred from magnetic studies. Tectonophysics 335 (3–4),

313–329.

De Rita, D., Giordano, G., Cecili, A., 2001. A model for sub-

marine rhyolite dome growth: Ponza Island (central Italy). Jour-

nal of Volcanology and Geothermal Research 107, 221–

239.

Hillhouse, J.W., Wells, R.E., 1991. Magnetic fabric, flow direction,

and source area of the lower Miocene Peach Spring Tuff in

Arizona, California and Nevada. Journal of Geophysical Research

96 (B7), 12443–12460.

Hrouda, F., 1982. Magnetic anisotropy of rocks and its application in

geology and geophysics. Geophysical Surveys 5, 37–82.

Jelinek, V., 1981. Characterization of the magnetic fabric of the rocks.

Tectonophysics 79, 63–67.

Knight, D., Walker, M., 1988. Magma flow direction in dykes of

Koolau Complex, Oahu, determined from magnetic fabric studies.

Journal of Geophysical Research (B5), 4301–4319.

Launeau, P., Robin, P.Y.F., 1996. Fabric analyses using the intercept

method. Tectonophysics 267, 91–119.

Lienert, B.R., 1991. Monte carlo simulation of errors in the anisotropy

of magnetic susceptibility: a second-rank symmetric tensor. Journal

of Geophysical Research 96 (B12), 19539–19544.

Maltman, A.J., 1994. Prelithification deformation. In: Hancock, P.

(Ed.), Continental Deformation. Pergamon Press, Oxford, pp. 144–

158.

Moreira, M., Geoffroy, L., Pozzi, J.-P., 1999. Magmatic flow in Azores

hot spot dykes: preliminary results with anisotropy of magnetic

susceptibility (AMS) in San Jorge Island. Comptes Rendus de

l�Academie des Sciences––Series IIA––Earth and Planetary Science

329 (1), 15–22.

Rochette, P., Jenatton, L., Dupuy, C., Boudier, F., Reuber, I., 1991.

Diabase dykes emplacements in Oman ophiolite a magnetic fabric

study with reference to the geochemistry. In: Ophiolite Genesis and

Evolution of the Oceanic Lithosphere. Kluwer, Dordrecht, pp. 55–

82.

Rochette, P., Aubourg, C., Perrin, M., 1999. Is this magnetic fabric

normal? A review and case studies in volcanic formation.

Tectonophysics 307, 219–234.

Rochette, P., Jackson, M., Aubourg, C., 1992. Rock magnetism and

the interpretation of ansisotropy of magnetic susceptibility. Review

of Geophysics 30 (3), 209–226.

Tarling, D.H., Hrouda, F., 1993. The Magnetic Anisotropy of Rocks.

Chapman & Hall, New York, p. 212.


Tauxe, L., Gee, J.S., Staudigel, H., 1998. Flow direction from

anisotropy of magnetic susceptibility data: The bootstrap

way. Journal of Geophysical Research 103 (B8), 17775–

17790.

Varga, R.J., Gee, J.S., Staudigel, H., Tauxe, L., 1998. Dike surface

lineations as magma flow indicators within the sheeted dike

complex of the Troodos ophiolite, Cyprus. Journal of Geophysical

Research 103 (B3), 5241–5256.


Magma flow in sub-aqueous rhyolitic dikes inferred from magnetic fabric analysis (Ponza Island, W. Italy

Documents