Injection mechanism of clay-rich sediments into dikes during earthquakes

21 Pages·2017·1.77 MB·English

Injection mechanism of clay-rich sediments into dikes
during earthquakes
Tsafir Lévi, Ram Weinberger, Tahar Aifa, Yehuda Eyal, Schmuel Marco
To cite this version:
Tsafir Lévi, Ram Weinberger, Tahar Aifa, Yehuda Eyal, Schmuel Marco. Injection mechanism of clay-
rich sediments into dikes during earthquakes. Geochemistry, Geophysics, Geosystems, 2006, 7 (12),
pp.Q12009. ￿10.1029/2006GC001410￿. ￿hal-00128401￿
HAL Id: hal-00128401
https://hal.archives-ouvertes.fr/hal-00128401
Submitted on 9 Jun 2017
HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est
archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents
entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non,
lished or not. The documents may come from émanant des établissements d’enseignement et de
teaching and research institutions in France or recherche français ou étrangers, des laboratoires
abroad, or from public or private research centers. publics ou privés.
33
Article
GG
GGeeoocchheemmiissttrryy Volume 7, Number12
GGeeoopphhyyssiiccss 29December 2006
Q12009,doi:10.1029/2006GC001410
GGeeoossyysstteemmss
ISSN: 1525-2027
AN ELECTRONIC JOURNAL OF THE EARTH SCIENCES
Published by AGU and the Geochemical Society
Injection mechanism of clay-rich sediments into dikes during
earthquakes
Tsafrir Levi
DepartmentofGeologicalandEnvironmentalSciences,BenGurionUniversityoftheNegev,P.O.Box653,BeerSheva,
Negev 84105, Israel
Geological Survey of Israel, 30 Malkhei Israel Street, 95501, Jerusalem, Israel
Ramon Science Center, Ben Gurion University of the Negev, P.O. Box 194, 80600, Mizpe Ramon, Israel
([email protected])
Ram Weinberger
Geological Survey of Israel, 30 Malkhei Israel Street, 95501, Jerusalem, Israel
Tahar A¨ıfa
Ge´osciences-Rennes,CNRSUMR6118, Universite´ deRennes l,CampusdeBeaulieu,F-35042RennesCedex, France
Yehuda Eyal
DepartmentofGeologicalandEnvironmentalSciences,BenGurionUniversityoftheNegev,P.O.Box653,BeerSheva,
Negev 84105, Israel
Shmuel Marco
Department of Geophysics and Planetary Sciences, Tel Aviv University, 69978, Tel Aviv, Israel
[1] Clastic dikes may form by simultaneous fracture propagation in rocks and injection of clastic material
into the fractures resulting from strong seismic shaking. We studied the mechanisms of clastic-dike
formation within the seismically active Dead Sea basin, where hundreds of clastic dikes cross-cut the soft
rock of the late Pleistocene lacustrine Lisan Formation. We analyzed the anisotropy of magnetic
susceptibility (AMS) of dikes with known formation mechanisms and defined the characteristic AMS
signatures, mainly of dikes developed by injection process. Most of the dikes were emplaced due to
fluidization of clay-rich sediment and are characterized by triaxial AMS ellipsoids. The dominant triaxial
AMS ellipsoids along the dike widths suggest that the fluidization mechanism of clay-rich sediment is
different from the known liquefaction process of sand. The AMS analysis supported by field evidence
indicates that the injection of clay-rich sediment is characterized by two main regimes: (1) Vertical flow
characterizedbysubverticalV axesandsubhorizontalV andV axes.TheV axesmayindicatetheflow
2 1 3 2
directions during fast flow. (2) Horizontal slow flow characterized by subvertical V axes and
3
subhorizontal V and V axes. A streaked AMS pattern mainly composed of V and V axes represents a
1 2 2 3
turbulent flow that generated local eddies simultaneously with the clastic transport. The AMS parameters
alongthedikesandpossible grainimbrications alongdikewalls supportorganizationofgrainsunderhigh
strain rates. This application of the AMS method provides a petrofabric tool for identifying seismites and
inferring their flow kinematics in complex geologic areas.
Copyright 2006bythe American GeophysicalUnion 1of 20
33
GeochemistryGG levi et al.: clastic-dike formation 10.1029/2006GC001410
Geophysics
Geosystems
Components: 9713words, 13figures.
Keywords: clastic dikes;fluid flow;anisotropyof magneticsusceptibility;Dead SeaTransform.
Index Terms: 1518 Geomagnetism and Paleomagnetism: Magnetic fabrics and anisotropy; 7221 Seismology:
Paleoseismology(8036);3653 MineralogyandPetrology:Fluidflow.
Received 4July2006; Revised27 August2006;Accepted18September2006;Published 29December2006.
Levi,T.,R.Weinberger,T.A¨ıfa,Y.Eyal,andS.Marco(2006),Injectionmechanismofclay-richsedimentsintodikesduring
earthquakes,Geochem. Geophys.Geosyst., 7, Q12009,doi:10.1029/2006GC001410.
1. Introduction especially during an earthquake shaking. Al-
though unconsolidated sandy soils are considered
[ ] Paleoseismic records commonly document to be materials that are most sensitive to lique-
2
strong to very strong (magnitude, M > 6) earth- faction, clay-rich soils can also be fluidized dur-
quakes, whereas geologic evidence of small and ing an earthquake. Several geological studies
moderate-sized earthquakes is rarely preserved at [e.g., Mohindra and Bagati, 1996; Rodr´ıguez-
the surface. Recognition of seismic events in the Pascua et al., 2000; Moretti, 2000] described
stratigraphic record is important in order to char- soft-sediment structures and suggested that these
acterize their frequency-size relations and com- structures were formed during earthquakes.
plete the paleoseismic record [McCalpin, 1996;
[ ] In this study we apply anisotropy of magnetic
5
Rodr´ıguez-Pascua et al., 2000].
susceptibility (AMS) analysis to characterize the
[ ] Seismites are deformational structures attribut- seismic origin of injection dikes. Foliation and
3
able to seismic activity. Our study deals with lineation of the magnetic fabric may form during
injected clastic dikes which are discordant sheets transport, deposition and deformation of rocks
ofclasticsedimentsformedbyforcefulintrusionof [Borradaile and Henry, 1997]. The fabric is com-
fluidizedclasticmaterialintothesurroundingstrata. monlyassociatedwithAMS,andcanbeusedasan
Injection clastic dikes are one form of seismites, indicator for flow directions in sediments [Tarling
and their emplacement corresponds to episodic andHrouda,1993;Liuetal.,2001]andinmagmas
pulses of increasing hydraulic pressure generated [Baer, 1995; A¨ıfa and Lefort, 2001; Abelson et al.,
by seismic loading [McCalpin, 1996]. The pattern 2001; Poland et al., 2004]. AMS has also been
of injection dikes hasbeen used for locating paleo- correlated with strain in rocks [Borradaile, 1991;
epicenters [Galli, 2000, and references therein]. Pare´s et al., 1999; Pare´s and Van der Pluijm,
Noteworthy, determining a seismic origin of injec- 2003], and has been used to characterize soft-
tionclasticdikesisnottrivialbecausetheirappear- sediment deformation [Schwehr and Tauxe, 2003].
ance may be similar to that of depositional dikes,
[ ] Although magmatic dikes and injection clastic
6
formed by passive infilling of fissures from above.
dikes are both related to hydrofracturing and fluid
[ ] Injection clastic dikes are referred to as lique- flow, they differ in several aspects. (1) The time of
4
faction structures. Engineering studies of ground AMS acquisition of magmatic dikes may be very
deformation associated with historical earthquakes long due to the slow rate of mineral growth
have shown that near-surface water-saturated sedi- [Fe´me´nias et al., 2004, and references therein].
mentsbecomeliquefiedasaresultofcyclicalshear On the other hand, because the magnetic minerals
stress [McCalpin, 1996, and references therein]. already exist, the origin AMS in clastic dikes is
The liquefaction occurs as a consequence of the acquired rapidly, immediately after the emplace-
increased pore water pressure whereby the granu- ment process. (2) Natural fluidized sediments are
lar porous material is transformed from a solid characterized by turbulent flow in contrast to
state into a liquefied state. Soft-sediment deforma- magma, which is more viscous [Turcotte and
tion is referred to as flowage or fluidization of Schubert, 1982]. Little is known about the kine-
cohesionless clay-rich sediments [Mohindra and matics of the flowage of sediments within dikes
Bagati, 1996]. Pore pressure may be involved as during earthquakes, especially if the injected sedi-
well, but little is known about its mechanism, ments are clay-rich. Nor is there much information
about the acquisition of AMS fabric during sedi-
2 of 20
33
GeochemistryGG lleevvii eett aall..:: ccllaassttiicc--ddiikkee ffoorrmmaattiioonn 1100..11002299//22000066GGCC000011441100
Geophysics
Geosystems
clastic transportation at fast flow, and improve the
understanding of the clay-rich sediment fluidiza-
tion process.
2. Geologic Setting
[ ] The Ami’az Plain study area (Figure 1) is
8
located west of the Mount Sedom salt diapir [Zak,
1967; Weinberger et al., 2006a, 2006b] near the
southwestern margin of the Dead Sea basin, along
the segmented Dead Sea fault (transform) [e.g.,
Quennell, 1959; Freund et al., 1968; Garfunkel,
1981]. The bedrock of the Ami’az Plain is the
(cid:1)40 m thick Late Pleistocene lacustrine Lisan
Formation consisting mostly of laminae of authi-
genic aragonite andgypsum layers alternating with
fine detritus layers [Begin et al., 1980]. A thin
veneer of eolian and fluvial sediments covers large
parts of the plain. The incision of Nahal (Wadi)
Perazim in the Ami’az Plain exposed the entire
Lisansectionandabout250clasticdikes,whichare
embedded within this section. The U-Th age of the
Lisan Formation is between (cid:1)70,000 and 15,000
years B.P [Haase-Schramm et al., 2004].
[ ] The Dead Sea basin is a continental depression
9
located within the rift valley that extends along the
Dead Sea fault. The basin is bounded on the east
and west by a series of oblique-normal step-faults.
The Ami’az Plain is oneofthe downfaulted blocks
developed within the rift valley. Paleoseismic
records from the Dead Sea basin based on breccia
Figure 1. Location maps of the study area. The layers reveal numerous M > 5.5–6 earthquake
regional setting of the Dead Sea Transform (inset) and events during the last 70,000 years [e.g., Marco
the Ami’az Plain with the clastic dikes marked and Agnon, 1995; Enzel et al., 2000; Ken-Tor et
schematically by broken lines. DST, Dead Sea Trans- al., 2001], as well as several M > 7 earthquake
form; SD, Sedom Diapir [after Levi et al., 2006]. events [Begin et al., 2005, and references therein].
Recent seismicity in the Dead Sea basin is pre-
sented by Shapira [1997]. The recorded strongest
event in the Dead Sea basin was the M = 6.2
ment flow through a channel, especially under fast
earthquake of 11 July 1927; its source mechanism
flow conditions.
was of a left-lateral motion [Ben-Menahem et al.,
[7] In a previous study [Levi et al., 2006] we used 1976; Shapira et al., 1993].
theAMSfabrictodistinguishbetweendepositional
[ ] The injection dikes in the Ami’az Plain
and injection clastic dikes emplaced along the 10
(Figure2)are composedofgreen clay,silty quartz,
margins of the seismically active Dead Sea Trans-
and some aragonite fragments. These dikes, up to
form. We showed that the AMS application pro-
hundreds of meters long, 30 m high and up to
vides a petrofabric tool to differentiate between
0.4 m wide, most probably originated in the lower
clastic dikes of different origins [Levi et al., 2006].
layers of the Lisan Formation [Levi et al., 2006].
In the present study we use the AMS fabric to
They are arranged mainly in radial and tangential
explore the emplacement mechanism and fluid
geometry. The connection of a green clay-rich
flow of dozens of Holocene injection dikes that
layer of the Lisan Formation to the dike-fill ob-
cross-cut late Pleistocene lacustrine soft rocks
servedinseveraldikesunequivocallyindicatesthat
exposed in the southwestern margin of the Dead
they were formed by injection of material from the
Sea basin. The results extend our knowledge on
3 of 20
33
GeochemistryGG levi et al.: clastic-dike formation 10.1029/2006GC001410
Geophysics
Geosystems
Figure2. (a) Clastic dike, filled with clay-rich sediment, cross-cutting the LisanFormation about 12–18mabove
itssourcelayer.Thedikebranchesupwardintoseveralstrands(seearrows)whichpartlycoalesceatdifferentlevels.
(b) Two disconnected, partially overlapping dike segments in the upper section of the Lisan Formation. Similar to
magmaticdikes[e.g.,Weinbergeretal.,1995],thisgeometryhintsattheroleplayedbyinternalpressureduringdike
emplacement and horizontal transport of clastic material into the evolving dikes. Lisan laminae are not displaced
along the dike walls, indicating that clastic dikes are extensional fractures. Note that because the dike segments are
not physically connected to the surface, the flow within it had a lateral component.
clay-rich layer. The majority of dikes terminate (Figure 2a) resembling dynamic fractures that
againsta0.5-m-thickgypsumlayeratthetopofthe bifurcate during upward propagation [e.g., Bahat
Lisan section. The lower ends of the dikes are et al., 2004]. The large strands are typically seg-
withinalternatinglaminaeofaragoniteandgypsum mented, forming numerous small-scale segments
layers of the lower Lisan Formation. At least five ((cid:1)0.15 m height) about 13 m above the source
of these dikes are composed of several (up to 12) layer. The architecture and discontinuity in the
distinct vertical sheets of sediments, 0.02–0.05 m vertical section of these dike segments is compat-
wide each, which we interpret as evidence of ible with a lateral propagation. The overlapping
multiple injections. Occasionally, the injection geometry between two segments implies opening
dikes are wider in their lower part than in their of a fracture under internal pressure [e.g., Delaney
upper part. Several injection dikes branch toward and Pollard, 1981; Weinberger et al., 1995] (see
the surface and split into 3–5 large strands Figure 2b). Levi et al. [2006] concluded that the
4 of 20
33
GeochemistryGG levi et al.: clastic-dike formation 10.1029/2006GC001410
Geophysics
Geosystems
formation of these dikes resulted due to seismic mayalsobeindicatedbytheinclinationsofV axes
3
loading.Localverticalstaticpressureasatriggeris as well [Rees and Woodall, 1975; Cagnoli and
excluded mainly because the dike phenomenon is Tarling, 1997; Liu et al., 2001; and references
local and very close to the surface. therein]. In high-energy currents with particles
entrained, V axes are perpendicular to the flow
[ ] Lesscommonaredepositionaldikescomposed 1
11 direction and V axes are commonly streaked,
of brownish silt, in places showing horizontal bed- 3
resulting in prolate or triaxial AMS ellipsoids
ding planes, which resemble the veneer of surface
[Tauxe,1998,andreferencestherein](seeFigure3:
sediments.Thesedikesalwaysintersectthepresent
E1–E3).Inearthquake-inducedinjectiondikesflow
surfaceandarecommonlywiderintheirupperpart.
velocityisexpectedtobehigh,henceV isexpected
1
to be perpendicular to the flow direction and
3. AMS Application for Clastic Dikes
streakedV –V distributionsmayevolve(Figure3:
3 2
E1–E3). Inthe lattercase, the flow direction would
[12] AMS is a second-rank tensor which is de- be indicated by either the V or V axes [Tauxe,
2 3
scribed by its principal values and principal axes 1998]. In both cases, the principal axes should
[Borradaile and Jackson, 2004, p. 300]. The prin- be well grouped, characterizing a ‘‘flow fabric.’’
cipal values, t1, t2, and t3, correspond to the Consequentdeformationduetoclosureofporespace
maximum, intermediate and minimum magnetic and expulsion of pore water contained in the sedi-
susceptibility magnitudes respectively, and the ment is not expected to modify the shapes of the
principalAMSaxesareV1,V2andV3,respectively AMSellipsoids[Pare´setal.,1999].
[Tauxe,1998].
[ ] In our previous study we showed that deposi- 4. Sampling Strategy and Methods
13
tional dikes display a sedimentary AMS fabric in
which V3 axes are vertical and well grouped, [15] A total of 312 samples were recovered from
whereas V and V axes are dispersed within the nine clastic dikes and country rocks. We carved
1 2
horizontal plane (Figure 3: A1). The values of the 2.5-cm-cylinder pedestals using a sharp knife, and
associated t and t are indistinguishable and placed on them plastic araldite-glue-coated cylin-
1 2
characterized by oblate AMS ellipsoids [Tarling ders with no AMS signal. The dikes were sampled
and Hrouda, 1993; Borradaile and Henry, 1997; across their width and along their height, 8–35
Liu etal., 2001]. The injection dikes, characterized specimens in each. On the basis of field observa-
byprolateortriaxialAMSellipsoids,displayaflow tions,twoofthedikesaredepositionaldikes(Tand
AMSfabric(Figure3:B2–E2),suggestingthatthe To) and seven are injection dikes (Tg, Tk, Tn, Tp,
flowdirectionisreflectedbyV axes.Inthepresent Q, Sb and TR). Twelve to twenty specimens were
2
study,weelaboratevariousaspectsofflowandflow collected from two Lisan layers, one of which is
fabric developed within injection dikes. We com- theclay-richsourcelayerofDikeQ(Figure4),and
pare it with the AMS fabric that is formed by a the other is an undisturbed layer.
viscous Newtonian flow in magmatic dikes, and
[ ] The sampling scheme of the seven injection
16
with the flow fabric developed in low- to high-
dikes was as follows: Two dikes (Tg and Q) were
energy currents in sedimentary environments.
sampled at different localities below the upper
[ ] In Newtonian flow, particle imbrication along gypsum layer, along their height and length. An
14
the dike walls and prolate/triaxial AMS ellipsoids intensive sampling of Dike Q was carried out
areexpected[Fe´me´niasetal.,2004,andreferences because it comprises several sheets (Qa, Qb, Qc,
therein] (Figure 3: B1–B3). The imbrication helps Qd; see Figure 4) that are connected to a Lisan
toresolvetheflowdirection,withV axesbeingthe sourcelayer(SLQ).DikeTRisamultipleinjection
1
common flow indicator. In the dike core inverse dike composed of 12 vertical bands of sediments.
fabric with oblate AMS ellipsoids are expected Threeofthesebandsweresampledinseveralplaces
(Figure 3: B1.1, B2.1, and B3). In low-energy along their heights. Dikelet Sb, one of the small-
currents, particle imbrication results in slightly scale dikes extensively developed in the upper part
off-vertical V axes [Rees, 1979] with oblate to of the Lisan section, was sampled along it length.
3
very weak triaxial AMS ellipsoids (Figure 3: C1–
[ ] Three dikes: Tk, Tp, and Tn, were sampled in
17
C3).Inmoderate-energycurrents,grainimbrication
the lower Lisan section near the level of the source
resultsinslightlyoff-verticalV axes[Tauxe,1998]
3 layer (i.e., (cid:1)18 m below the upper gypsum layer).
and V axes are inclined to the opposite flow
1 In each dike between 12 and 20 specimens were
direction (Figure 3: D1–D3). The flow directions
5 of 20
33
GeochemistryGG levi et al.: clastic-dike formation 10.1029/2006GC001410
Geophysics
Geosystems
Figure 3. Anisotropy of magnetic susceptibility (AMS) for clastic dikes that were emplaced under different
conditions.A1–E1:equal-areaprojections(lowerhemisphere)ofsyntheticAMSprincipalaxesandtheirconfidence
zonesbasedonthebootstrapmethod:whiteregion,distributionofV axes;grayregion,distributionofV axes;black
1 2
region, distribution of V axes. Dike strike is marked by adashed line. A2–E2: distribution of principal values (t ,
3 1
t , and t ) with 95% confidence bounds, using the bootstrap method [Tauxe,1998]. A3–E3: Schematic illustration
2 3
of various possibilities of particle distribution within clastic dikes. Arrows mark the direction of particle
transportation. (A) Sedimentary (oblate) AMS fabric in depositional dikes. (B) Flow AMS fabric developed during
injectionof viscous Newtonian fluid and characterized by imbrications along the dike’s walls. The flow direction is
inferredonthebasisoftheimbricationsofV axes.Theprincipalaxesinthehorizontalplaneareeithergrouped,e.g.,
1
along the dike walls (B1), or dispersed, e.g., in the dike core (B1.1). The values of the AMS ellipsoids are either
triaxial (B2) or oblate (B2.1). (C) Low-energy flow fabric of oblate to weak triaxial AMS ellipsoids. (D) Moderate-
energy flow fabric of triaxial (D2) or weak triaxial (D2.1) AMS ellipsoids. The principal axes are either grouped in
thehorizontalplane(D1)ormoredispersed(D1.1),dependingontheflowrate.Interpretationoftheflowdirectionis
basedonV inclinations[e.g.,Liuetal.,2001]andisinoppositedirectiontoV axesinclinations[Rees,1979;Tauxe,
3 1
1998]. (E) High-energy flow fabric of triaxial (E2) or weak triaxial (E2.1) AMS ellipsoids. The principal axes are
eithergrouped(E1)orstreaked(E1.1)duetorotationofparticlesduringturbulentflow.V axesmightbeorthogonal
1
to the flow direction and either V or V axes indicate the flow direction.
2 3
evenly distributed across the dike width and be- [ ] The AMS was measured with a KLY-3S
18
tween 5 and 9 specimens were sampled along the Kappabridge at the Geosciences Laboratory, Uni-
dikemargins.Thissamplingstrategyformsan‘‘H- versity of Rennes 1, France. The principal suscep-
like shape’’ designed to detect possible particle tibility axes with the 95% confidence ellipses and
imbrication and effect of syn- and post-shearing the bootstrapped axes were analyzed with the
along the margins. program ‘‘programs magnetic anisotropy analysis’’
6 of 20
33
GeochemistryGG levi et al.: clastic-dike formation 10.1029/2006GC001410
Geophysics
Geosystems
Figure 4. Schematic illustration of the Q dike system composed of the source layer Q (SLQ), Dike (a), and the
sheets system (SS: b–d). The dike system is bounded below by a series of thin alternating gypsum layers (GL) and
detritus(DL)andbyathickgypsumlayer(GL)above.Cylindersmarksamplingpointsandthestereogramsshowing
schematic fabric variations along the dike height (see Figure 8).
written by Henry. B, Lienert. B, and Le Goff. M. Thermomagnetic curves of brownish clay-rich
The shapes of the AMS ellipsoids were also sediments taken from the dike infill exposed in
analyzed by the Bootstrapping method [Tauxe, the uppersection oftheLisan Formation showthat
1998]PMAGsoftware (L.Tauxe, 2002), assuming the magnetic carriers are titanomagnetite and
that the samples represent the whole population. maghemite (Figure 5b). The lack of hematite in
thedikeinfillimpliesthatoxidationplayedaminor
[ ] Flow directions were analyzed following
19 role after its emplacement [A¨ıfa and Lefort, 2001].
Moreira et al.’s [1999] procedure. To characterize
the magnetic carriers of the dike’s infill and the [ ] Hysteresis loops of clay-rich sediments taken
21
Lisan source sediments, we used 20 thermomag- from the Lisan source sediments and dike infill
netic curves and 20 hysteresis loops of specimens show that the ratio J /J is near 0.1 and the H /H
rs r cr c
from the dikes. isaround6.5(Figure5c).Thesevaluesarestrongly
favored for multidomain (MD) grain size [Day et
5. Results al., 1977]. Comparison between the present results
and published hysteresis loops of the Lisan sedi-
ments [Marco et al., 1998] and clay sediments
5.1. Rock Magnetism
[e.g., Schwehr and Tauxe, 2003; Cifelli et al.,
[20] Thermomagnetic curves of green clay-rich 2004] shows that the clay-rich sediments are of
sediments taken from the Lisan source sediments relatively large grain size.
and dike infill exposed in the central and lower
[ ] The mean magnetic susceptibility values
sections of the Lisan Formation show that the 22
(K = [k + k + k /3] [Nagata, 1961]) of all
magnetic carrier is titanomagnetite (Figure 5a). mean 1 2 3
7 of 20
33
GeochemistryGG levi et al.: clastic-dike formation 10.1029/2006GC001410
Geophysics
Geosystems
Figure 5. Rock magnetism of clastic dikes. (a) Representative thermomagnetic curve from Dike Q. The steep
gradient of the susceptibility around 510(cid:1)C is attributed to titanomagnetite [e.g., Ferre´ et al., 2002]. (b)
ThermomagneticcurvefromDikeSb.Theinflectionpointaround350(cid:1)Cisattributedtomaghemite[e.g.,Archanjoet
al., 2000], and that around 510(cid:1)C is attributed to titanomagnetite. (c) Day diagram [Day et al., 1977] showing the
results of20hysteresis loopsofspecimens fromthemarginsandthecenters of thedikes.Triangles, injectiondikes;
squares, depositional dikes. Jrs is the saturation remanence, and Jr is the saturation magnetization. Hcr is the
coercivityofremanence,andHcisthecoerciveforce.TheaverageratioJrs/Jrisnear0.1,andtheaverageratioHcr/
Hc is around (cid:1)6.5, suggesting multidomain grain size.
samples recorded from the clastic dikes and the
Lisan source layer range between 40 and 360 (cid:2)
10(cid:3)6 SI. This range of values is typical for clay
sediments [Cifelli et al., 2004]. The average K
mean
of the depositional dikes is (cid:1)225 (cid:2) 10(cid:3)6 SI
whereas that of the injection dikes and their source
layer is (cid:1)70 (cid:2) 10(cid:3)6 SI (Figure 6). This difference
may reflect different sedimentary origins [e.g., Liu
et al., 2001]. The K of all the sampled dikes
mean
decrease toward the bottom from 150 (cid:2) 10(cid:3)6 SI at
the surface to 60 (cid:2) 10(cid:3)6 SI (cid:1)18 m below the
upper gypsum layer (Figure 6).
5.2. AMS
5.2.1. General Figure 6. Variations of the mean susceptibility values
withdepthbelowtheuppergypsumlayer.Susceptibility
[23] The AMS fabrics of eight representative clas- values increase unmonotonously upward along Dike Q
ticdikesandtheLisansourcelayerarepresentedin and all others dikes. Diamonds, Dike Q; circles,
Figures 7 and 8. The depositional dikes, repre- injection dikes; triangles, sedimentary dikes.
8 of 20
33
GeochemistryGG levi et al.: clastic-dike formation 10.1029/2006GC001410
Geophysics
Geosystems
Figure7. AMS of the depositional dike (Dike To,A) and six injection dikes (Dikes Sb, Tg, Tk, Tp, Tn, andTR).
A1–J1: Lower-hemisphere, equal-area projections of AMS principal axes and the 95% confidence ellipses; squares
representV axes,trianglesrepresent V axes,andcircles represent V axes.A2–J2:Lower-hemisphere,equal-area
1 2 3
projectionsofAMSprincipalaxesanalyzedbythebootstrappingmethod.A3–J3:Principalvaluesdistributionwith
95%confidencebounds.Dashedlinesmarkthedikestrike.A2–A3showasedimentaryfabric.B2–C2andB3–C3
showgroupedsubvertical V axeswithweaktriaxialflowfabric.D2–J2andD3–J3showsubverticalV axeswith
3 2
weak triaxial to triaxial flow fabric. Streaked V –V fabric is seen in E2, G2, H2, I2, and J2. Notice that the 95%
2 3
confidence ellipses look similar to the confidence zones analyzed by the bootstrap method (see also Figure 8).
sented here by Dike To (Figure 7: A), exhibit a havedifferentanddistinctiveAMSfabrics.Inthese
sedimentary fabric (Figure 3: A) similar to that of dikes the three principal axes are well grouped,
the Lisan layer (Figure 8: A). This fabric also whereastheV axesaresubvertical(Figure7:D1–
2
resembles that of a low velocity current (Figure J1 and Figure 8: C1–E1) or subhorizontal (Figure
3: C). 7:B–CandFigure8:B–D,F)andtheV axesare
1
subhorizontal, and parallel to the dike strike. In
[ ] The injection dikes (Sb, Tg, Tk, Tp, Tn, TR)
24 addition, the bootstrap statistics principal values
(Figure 7: B–J) and the Q system (Figure 8: B–F)
9 of 20