Journal of Conference Abstracts

Hans Ramberg Symposium ­
Analogue Modelling of Tectonics

EUG 10

Hemin Koyi

N. Mancktelow

Hans Ramberg Symposium ­ Analogue Modelling of Tectonics

Session L01:3A

0277

KEYNOTE

L01 : 3A/01 : F4

Sand as an Analogue Material in Tectonics: Advantages, Limitations and Future Developments

Peter Cobbold (cobbold@univ-rennes1.fr)

Géosciences-Rennes, Campus de Beaulieu, 35042 Rennes, France

Sands of various kinds have been used for over a century for modelling tectonic processes. M. King Hubbert was one of the first to realize that dry sand is an excellent analogue for brittle deformation of the upper crust in the Earth's field of gravity. In recent years, dry quartz sand has become very popular as a model material, especially in multilayered models, where ductile materials represent the lower crust, mantle lithosphere or asthenosphere.

Advantages. When sand undergoes large strains, the applied stresses satisfy a yield criterion of frictional (Navier-Coulomb) type. The angle of internal friction is about 40° and the cohesion is of the order of 100 Pa. This means that a layer several cm thick can deform easily under its own weight, satisfying the scaling laws for body forces and surface forces. Physically, permanent deformation in sand is achieved by faulting. Faults form by progressive dilation. They can be imaged in an X-ray scanner without the model being destroyed. New layers of sand can easily be added to the free surface of a model, or removed from it, during ongoing deformation. This is invaluable in modelling the development of sedimentary basins and mountain belts. Last but not least, sand is clean, inert and relatively cheap. So far, computers have not been able to compete with sand for efficient modelling of progressive three-dimensional deformation in the upper crust.

Limitations. The elastic strain range of sand is much smaller than that of brittle rock. This makes sand inadequate for modelling seismic events. Another disadvantage of sand is that dilations in fault zones tend to be large (up to 30%), in comparison with fault zones in nature.

Future developments. An entirely new development is the use of pore fluids in sandpacks. Compressed air is perhaps the simplest fluid to use. Its viscosity and the intrinsic permeability of sand (about 1 darcy for a grain size of 0.3 mm) yield a convenient time scale for experiments. Initial tests have shown that sand obeys Terzaghi's law of effective stress. As pore fluid pressure approaches lithostatic pressure, the shear strength of sand decreases to a value little greater than its cohesion. By deforming models with layers of differing frictional properties and permeabilities, it is possible to study interactions between faulting and fluid flow. Future applications will probably include detachments in sedimentary basins and the migration and trapping of hydrocarbons.

3277

L01 : 3A/03 : F4

Critical Evaluation of Tectonic Model Designs

Bruno Vendeville (vendevillb@begv.beg.utexas.edu)

Bureau of Economic Geology, University Station, Box X, Austin, TX 78713, USA

Following Hans Ramberg's pioneer work,there has been a flurry of experimental tectonic models simulating various geological settings at various scales. These models are often used by non-modellers as guidelines for interpreting field or seismic data. However, non-modellers cannot confidently assess the validity of specific model results because model designs, based on assumptions and simplifications (e.g., rheological properties, boundary conditions), modeling materials (e.g., viscous, brittle, elastic), and experimental setups (e.g., initial geometries, imposed strain rates are rarely clearly discussed in published articles. Also, specific design characteristics that may appear as unimportant to the reader can exert a drastic control on the final result, and hence its geological applicability and validity. Moreover, some of these effects are often overlooked by modellers themselves when they do not critically evaluate their own experimental designs.

I will use my own models, as well as models from the literature, to illustrate some of the main design pitfalls that are frequently found in tectonic model. These pitfalls include the following:

- The myth of absolute scaling: we commonly read articles on models presented as analogs for a particular natural example, having very specific properties. However, model scaling cannot be more accurate than the data from the natural prototype. For example, estimates of salt viscosity vary across two to three orders of magnitude. Therefore the accuracy of model scaling (whether for time, deposition and strain rates, and for viscosities) varies by three orders of magnitude.

- The importance of proper scaling for viscous forces and strain rates: any model involving viscous material, hence time-dependent rheology, must be properly scaled according to the range of values for viscosity and rates in nature (e.g., strain or sedimentation). I will illustrate how models based on improper values for the rate of basement-fault slip and for salt viscosities have mislead us about deformation styles above basement faults.

- The role of lateral friction: modellers commonly assume that lateral friction effectively cancels out at or near the center of models. On the contrary, lateral friction influences deformation of the entire model by decreasing the amount and rate of extension or contraction and by controlling the preferential fault and fold vergence. In fold-and-thrust-belt models, lateral friction affects the angle of the critical taper and can even reverse the orientation and sense of propagation of the wedge.

- Proper and improper experimental analogs for a low-friction slip plane in nature: do sheets of thin, unstretchable, flexible plastic film (as commonly used in modelling) constitute a proper kinematic and dynamic analog for slip planes?

- Proper and improper natural analogs for basal discontinuities commonly used in modelling to trigger complex 3-D faulting, such as pull-apart basins.

0668

L01 : 3A/04 : F4

The Role of Polyphase Deformation on the Growth of Evaporitic Diapirs: Experimental Modelling of the Bicorp-Quesa Diapir (Eastern Betics, Spain)

Eduard Roca (eduard@natura.geo.ub.es)¹,

Maura Sans (maura@natura.geo.ub.es)¹ &

Hemin Koyi (hemin.koyi@geo.uu.se)²

¹ Dept. Geodinamica i Geofisica, Universitat de Barcelona, 08071 Barcelona, Spain

² Hans Ramberg Tectonic Laboratory, Department of Earth Sciences, Villavägen 16, S-752 36 Uppsala, Sweden

The Bicorb-Quesa diapir, located in the eastern Prebetics, is an excellent example to study the effect of later compression on an outcropping diapir. This structure belongs to a complex system of evaporitic diapirs developed in a deformed wedge of Mesozoic and Cenozoic sedimentary rocks, detached along Triassic evaporites. The Neogene sediments located in the Bicorb-Quesa diapir rim synclines document a regional polyphase deformation history. In this area, an initial extension phase formed a graben system in the overburden along which diapirs rose locally. A later compression phase caused regional shortening which necked the pre-existing diapirs. During a final extension phase, a new set of diapirs were initiated.

In order to investigate the key factors of this dynamic evolution, scaled sand-box models were designed. These models were allowed to undergo a similar kinematics as the eastern Prebetics area: a first extension phase which triggered the diapirs, a later compression and a final stage of extension. The results of these models show that, during the first extension phase, the overburden was divided into separated undeformed blocks that rotated along the bounding extensional faults, which accommodate most of the deformation. During this stage, the ductile layer accumulated beneath the grabens, where they locally pierced the thinned overburden. Subsequent compression closed the pre-existing diapirs forming secondary welds along their previous stems. It should be emphasized here that this later compression reactivated the pre-existing normal faults as high angle reverse faults accommodating the shortening. Shortening also redistributed the ductile layer thickness, acumulating it in the cores of the anticlines and displacing it beneath the synclines. This process compartmentalized the ductile layer to form isolate bodies in the cores of the anticlines. This new redistribution played a significant role in localizing the new diapirs which formed in the second extensional stage. These second generation diapirs pierced through the hinge areas of the pre-existing compressive anticlines. As a result, the second generation diapirs rose through a different place than the squeezed diapirs formed during the first extensional stage.

0693

L01 : 3A/05 : F4

Influence of Backstop Geometry on Fold Style and Salt Response During Experimental Shortening

Elisabetta Costa (costae@ipruniv.cce.unipr.it)¹ &

Bruno Vendeville (vendevillb@begv.beg.utexas.edu)²

¹ Dipatimento di Scienze della Terra, Parco Area delle Scienze 157/A, 43100 Parma, Italy

² Bureau of Economic Geology, The University of Texas at Austin, University Station, Box X, Austin, Texas USA

We use a series of systematic experiments to illustrate the effect of slow shortening onto a brittle cover overlying a viscous shaly or evaporitic decollement. All experiments were run under identical conditions except two varying parameters: (1) the cover/decollement thickness ratio and (2) the presence or absence of a deformable backstop. Models without a backstop comprised a continuous basal layer of viscous polymer and were deformed by moving a rigid plexiglass wall. In the other models, a narrow, deformable backstop, made of dry sand overlying glass-beads, was intercalated between the moving wall and the model itself. The glass-bead/sand interface acted as a low-angle detachment plane, in contrast with the polymer layer that deformed as a decollement layer. In models without a deformable backstop, the brittle layer was entirely detached from its base and deformed by box-folds bounded by a pair of conjugate kink bands. The underlying viscous decollement layer thickened but never rose diapirically. Instead, the viscous polymer was expelled downward from the core of the anticlines, where the horizontal stress was highest. In models with a deformable backstop, a fault-propagation fold formed in front of the backstop. There the viscous layer was detached at or near its base and was entrained along and incorporated into the hangingwall of the fold. As the fold grew and tightened, the viscous layer was able to forcefully pierce the streched fold crest and emerge. Experiments also show that a minimum thickness is required for the viscous layer to rise diapirically. We will compare our models results with data from Eastern Mediterranean, where shale diapirs are present. They occur south of Crete in the crest of the Mediterranean Ridge. There diapirs form elongate trends parallel to structures within the ridge, in front to the continental backstop. This situation is comparable to that of models having deformable backstop. On the contrary, at the outer deformation front of the accretionary wedge no mud diapirs and mud volcanoes emerge, although in this area collision is occurring and salt is very thick. This area is comparable to models without a deformable backstop.

0671

L01 : 3A/06 : F4

Modeling the Role of Erosion in the Formation of Salt Diapirs in Contractional Settings

Maura Sans (maura@natura.geo.ub.es)¹ &

Hemin Koyi (hemin.koyi@geo.uu.se)²

¹ Dept. Geodinamica i Geofisica, Universitat de Barcelona, 08071 Barcelona, Spain

² Hans Ramberg Tectonic Laboratory, Department of Earth Sciences, Villavägen 16, S-752 36 Uppsala, Sweden

Shortening inhibits diapirism by thickening overburden units and pinching the salt out of cores of anticlines. Nevertheless, salt diapirs exist in many fold-and-thrust belts. In the south Pyrenees, many diapirs of Triassic salt developed prior to compression (Upper Cretaceous-Miocene) and were later transported in the thrust sheets. However, the Cardona diapir of Eocene salt (Cardona salt formation) crops out in the deformed foreland. This diapir developed superimposed on an anticline related to a NE-SW trending fold system, which is detached above the Eocene salt. The diapir, which pierces 300 m of sediments on the southern limb of the Cardona-Pinós fold close to the hinge area, is 2 km long and 0,7 km wide. The internal structures of Cardona diapir are well studied, but its origin and age of piercement is less well understood.

Dynamically scaled models were shortened in a centrifuge in order to investigate the triggering mechanism for the Cardona diapir and to study the role of erosion in the development of diapirs in a previously shortened area. The model consisted of a microlaminate overburden of alternating laminae (1 mm thick each) of plastilina and Dow Corning silicone simulating a 2,5 km thick overburden of fluvial and alluvial sandstone with some inter-bedded thin lacustrine limestone and marls. The overburden units were placed over a layer of silicone putty simulating a 300 m thick layer of Cardona salt. The model was laterally shortened to 12% in the centrifuge. In the model, the folded overburden units compartmentalized the underlying ductile layer, which flowed from under the sinking syncline troughs to accumulate in the cores of the anticlines. Later erosion across the crests of the anticlines thinned the overburden units above the salt pillows in the cores of the anticlines and aided piercement of two dimensional diapiric structures (salt walls) along the crests of the anticlines. This erosion caused a differential loading, which was 8 times larger in the synclinal areas than in the cores of the anticlines.

Model results suggest that salt diapirs in areas of moderate compression, are likely to rise from the salt in the cores of the anticlines if later erosion thins overburden units above the topographically high anticlinal to initiate sufficient differential loading to trigger piercement. Our models show that the anticlines need to be open structures in order to allow enough accumulation of salt, which can later rise diapirically when the overburden is thinned by erosion.

Ascent velocities, which were calculated using the deformed Pliocene-Pleistocene deposits around the Cardona diapir, support a post-shortening age for the piercement after removal of 2 km of its overburden 3 Ma ago.

3539

L01 : 3A/09 : F4

Field Constraint of Rock Viscosities

C. J. Talbot (Christopher.Talbot@geo.uu.se)

Hans Ramberg Tectonic Laboratory, Department of Earth Sciences, Uppsala University, SE-752 36 Uppsala, Sweden

The strain rates currently used for modelling ductile rocks are extrapolated several orders of magnitude from lab. experiments, calculated from mineral contents, or simple but arbitrary ratios. However, structural geologists could potentially follow meta-morphic mineralogists and distinguish strain (as opposed to metamorphic) facies and map the paths of viscosities in P-T space-time of "pelites, psammites, carbonates, basalts, granites and vein quartz".

Attention will be focused on direct retrospective measurements of rock rheologies during natural deformations by reviewing classical field tools for their constraint. The potential of these tools will be illustrated using viscosity ratios measured from structures and fabrics in pairs of rocks that have deformed together in a limited range of tectonic settings.

0966

L01 : 3A/10 : F4

Mechanical Instabilities Associated to Rod Development: Preliminary Results from Analogue Modelling

Albert Griera (geotec@geologia.uab.es) &

Jordi Carreras

Dept. Geologia, Universitat Autonoma de Barcelona, 08193 Bellatera, Spain

Quartz rods include a variety of structures characterized by their linearity, commonly formed by highly elongate quartz bodies included in a less competent matrix. There are several lithological and deformational settings which can lead to rod-like structures (e.g. stretched fold hinges or pebbles). A particular setting of quartz rods is observed in the Culip area of the Cap de Creus peninsula, eastern Pyrenees. These quartz rods derive from D1 boudinaged quartz veins which become subsequently folded by D2 event. Due to the fact that boudined veins become extended parallel to the D2 fold hinges, the resulting structures consist on arrays of elongated quartz bodies, each being a rotated quartz boudin. Field evidence suggest that during D2 deformation preexisting boudins are the locus of strain perturbations which are responsible for the complex structure observed in the enclosing host rock made of an alternace of metagreywackes and metapelites.

The study of the development of these structures is being performed by means of analogical modelling with the BCN-deformation apparatus at the Universitat Autonoma de Barcelona The undeformed samples consist on a competent boudin-shaped paraffin layer in a weaker paraffin matrix either homogeneous or layered. The experimental variables are: i) the competence contrast between boudin-layer and the isotropic or multilayered nature of the matrix, ii) the length/width ratio of the initial boudins (R) and the ratio boudin length/ interboudin length (d), iii) the initial orientation and dip of the layer and iv) the type of flow field varying from Wk = 0 to Wk = 1, imposed by the cell boundary conditions.

Preliminary experiments has shown (1) the aspect ratio of the boudins (R), the interboudin ratio (d) and the initial orientation respect instantaneous stretching axes (ISA) have a strong influence on the generation and evolution of the structures. Small d values generated an imbricate boudin system; (2) at the beginning, the bulk flow is strongly partitioned and the strain is preferentially concentrated in the interboudin regions; (3) the rotation of boudins produced the spinning of ISA in the interboudin regions; (4) domains with vorticity opposite to that the bulk flow and to the sense of rotation of boudins occurred in the matrix close to two edges of each boudin; (5) rotation of boudins is more effective if slip occurs along the interface.

The final pattern of the experiments consists on trains of asymmetric folds affecting also the adjacent matrix layers, with boudins localized in the short limbs, showing different degrees of rotation.

1162

L01 : 3A/11 : F4

Viana do Alentejo-Alvito's Antiform Macrostructure (SW Iberia-Osso Morena Zone): Experimental Modelling of Multilayer Sheath Fold Generation by Homogeneous Simple Shear

Filipe Rosas (frosas@fc.ul.pt) &

Fernando Marques (fmarques@fc.ul.pt)

Dep. Geologia-Faculdade de Ciências, Universidade de Lisboa, Campo Grande, Bloco C2, 5º, Portugal

Based on the observation of 'mushroom'-like geometric pattern of sheath folds in transverse (approximately W-E) cross sections in Perdizes' outcrop (Viana do Alentejo-Alvito's sector, Ossa Morena Zone, Southern Iberian Variscan Fold Belt, e. g. Fonseca 1995) several experiments were carried out to exemplify (after Cobbold & Quinquis 1980, Ghosh & Gengupta 1984, Brun & Merle 1988, Marques & Cobbold 1995) multilayer generation of such structures under simple shear deformation regime. In an horizontal simple shear box a multilayer set was conceived: over a thin pink silicone layer a concave deflection was imprinted on top of a thicker transparent silicone plate and then covered with another similar transparent silicone layer. Imposed top to the left simple shear led to the accumulation of deformation around such a deflection giving rise to a well developed sinform sheath fold. When deformation stopped, several transverse cuts were made, from the apex of the sheath fold's 'nose' to its root zone, in order to observe its geometric pattern through this direction. A series of transverse cross sections of the sheath fold's geometric patterns were obtained, all comprised between two end terms: the 'eye'-like geometric pattern (corresponding to a cut made in the apex of the sheath fold's nose) and the, wide open, 'mushroom'-like geometric pattern (resulting from a cut near the root zone of the sheath fold.Comparison of the geometric patterns of transverse cross sections of sheath folds observed in the field (Perdizes' outcrop) and in the laboratory, confirms that this type of structures can form by accumulation of deformation in non-cylindrical anisotropies, under an homogeneous simple shear deformation regime.

Brun J & Merle O, Tectonophysics, 145, 129-139, (1988).

Cobbold PR & Quinquis H, J Struct Geol, 2, 119-126, (1980).

Fonseca PE, PhD Thesis Univ Lisboa, 325, (1995).

Ghosh SK & Sengupta S, J Struct Geol, 6, 703-709, (1984).

Marques FG & Cobbold PR, J Struct Geol, 4, 589-602, (1995).

L01 : 3A/12 : F4

Behaviour of the Seismogenic Crust Experimentally Simulated with Clay and Plasticine in Pure Shear

Luísa Pinto Ribeiro &

Fernando Ornelas Marques

Dep. Geologia, Fac. Ciências, Univ. Lisboa, Edifício C2, Piso 5, 1700 Lisboa, Portugal

Except for most of the Cainozoic Alpine fracturing, late orogenic deformation (fracturing mostly), presently exposed at the surface as a result of denudation, has developed some km deeper in the crust. Thus, instead of using sand (or similar) to model fracturing (brittle behaviour), we have used a more plastic material, clay, sometimes with thin plasticine markers. We have submitted the clay to pure shear, with a variable side stress, at room temperature, and up to 40% shortening. In these conditions, the used clay reacts by flattening and fracturing (controlled by a grid carved on top of the model), being thus a good analogue for the seismogenic continental crust. Formation of faults is usually preceded by, and results from, en échelon fracturing and their coalescence. In the early stages of faulting, two sets of conjugate brittle shear zones form at an angle around 60º. With proceeding deformation, two situations can develop: 1. the initial geometry is kept throughout the experiment, with virtual no rotation of the conjugate faults (rare); 2. the earlier shear zones rotate significantly, they show a sigmoidal shape and strike-slip duplexes develop (common). Because the experiments take place at constant volume, no hollows are allowed, and space problems must be solved by alternating intersection and kinematics of the conjugates. Only two to three major faults develop, and, although at the early stages of the experiments both conjugates show identical development, one of them will dominate over the other and prevail throughout the rest of the experiment. In most cases the faults are discrete surfaces, but sometimes shear zones develop with a significant thickness and intricate duplex geometry. In the early stages of conjugate development, the intersection is always a conflict zone, with development of many minor faults. In many cases, sets of minor faults develop that do not cross the main conjugate, but, nevertheless, show a significant amount of displacement. This can be analysed in terms of the rotation of faults and flattening of clay in the surroundings. Flattening is far from being homogeneous across the model, especially when there is significant rotation of the faults; then, significant domains show virtually no strain, despite the final high percentage of shortening (40%). The analysis of the graphic - compressive stress x% shortening - suggests that the early stages of deformation of the clay are accompanied by strain hardening, and, after the maximum of <sigma> 1 is reached, strain softening slowly develops. Seismicity can be deduced from the <sigma> 1 x time graphic.

1513

L01 : 3A/13 : F4

Modification of Early Lineations in Simple Shear

Sudipta Sengupta (sudipta@jugeo.clib0.ernet.in) &

Hemin Koyi

Dept. f Geological Science, Jadavpur university, Calcutta 700 032, India

A series of experiments were carried out with analogue models to study the deformation of early lineations on layers undergoing buckling in a non coaxial bulk strain. Model results show that the eometry of the deformed lineations is dependent mainly on four factors, viz., the initial angle between the lineation and the fold axis, orientation of the folded layer in relation to the bulk strain ellipsoid, intensity of deformation and the competence contrast between the layer and the embedding medium. The results conform well with the theoretical predictions. A U-shaped pattern of deformed lineation forms only when the orientation of the lineation is at right angle to the fold axis during the initiation of the fold. When the angle is lower, the deformed lineation shows an S or Z geometry, where it makes a lower angle with the fold axis in the hinge area than at the limbs. It is emphasized here that, in addition to the homogeneous strain, the variation of tangential longitudinal strain contributes to the reorientation of the pre-existing lineation during later deformation.

0473

L01 : 3A/14 : F4

Analogue and Finite-Element Modelling of Elasto-Viscous Single-Layer Folding

Neil S. Mancktelow (neil@erdw.ethz.ch)

Geologisches Institut, ETH-Zentrum, CH-8092 Zürich, Switzerland

As experimental techniques, analogue and numerical models have both inherent strengths and weaknesses. Analogue scale-models are limited in the range of rheological parameters and applied boundary conditions by the materials available and the deformation rig geometry. Practical limitations are often encountered, such as achieving perfect adherence between layer and matrix, determining displacement and strain fields with high resolution, and measuring internal stress distributions. The major advantage is that results are physically realistic for the particular model parameters employed. Numerical modelling provides the ideal complement. The material properties, imposed boundary conditions and resolution of stress and strain determination across the model can be easily varied (within limits) and mathematically exact shapes (e.g. bell-shaped perturbations or other irregularities with known spectral distributions) used as initial irregularities. However, this freedom in the choice of parameters, together with the ever-present possibility of mistakes, limitations or hidden assumptions in the computer code, means that there is no guarantee that the final results are realistic.

Combined finite-element and analogue modelling was used to study single-layer folding of materials with elasto-viscous rheology (both linear and power-law viscosity) during shortening under pure shear, constant natural strain rate conditions. The experiments establish that the shape of the finite-amplitude folds is controlled both by material properties (in particular Deborah number and effective viscosity contrast) and the shape of initial irregularities. Isolated perturbations initially amplify locally and propagation sideways along the layer is gradual. The sudden and rapid development of a quasi-periodic wave-train along the length of the layer above a certain shortening threshold, as reported from earlier finite-difference models, was not observed in either analogue or finite-element models. An ultimate amplitude to individual folds is attained, so that although folds are added serially, the final fold packet is regularly periodic. The only indication of the sequential development is a preserved variation in layer thickness. Increasingly elastic rheology and power-law viscosity both enhance the growth of shorter wavelengths and thus greater localisation of folding around the isolated perturbation. Folding experiments with an initial random distribution of irregularities produce a quasi-periodic final fold shape, in which the influence of initial perturbation shape is still clearly discernible. Random segments "cut" from one layer distribution and "pasted" into another still develop the same characteristic final fold shape, irrespective of the shapes to either side. Increased elasticity and power-law viscosity again promote faster growth of shorter wavelengths but there is no fundamental change in folding style, for example a change to greater localisation. The study demonstrates that the initial perturbation shape always maintains a strong but localised influence and explains the rather irregular, quasi-periodic form of natural single-layer folds.

Session L01:3B

2106

L01 : 3B/25 : F4

Analogue Models of Ductile Transpression Zones

Alexander Cruden (cruden@credit.erin.utoronto.ca) &

Pierre-Yves Robin (probin@credit.erin.utoronto.ca)

Dept. of Geology, University of Toronto, Mississauga, Ontario, Canada

Finite strains and particle paths have been investigated in a series of analogue "dual viscosity" transpression models with a thin, planar low viscosity inner zone and an outer higher viscosity zone. The models consist of a 12 cm high, 35 cm long planar zone of variable width squeezed between Plexiglas plates. A piston drives the plates towards each other at a velocity of about 1 cm/hr. The angle, alpha, between the piston and plates can be varied from 12 to 90°, simulating a wide range of transpressive conditions. Transparent PDMS is used to model the outer, high viscosity parts of the zones, and transparent polybutenes (Hyvis 150 & 2000), dyed for visualization purposes, are placed in the centre to model the inner, low viscosity zone. Synthetic sponge stoppers placed at the ends of the model prevent lateral escape of material. Material in the zone is therefore simultaneously extruded upwards and sheared, simulating the conditions of transcurrent transpression.

Strain and particle path visualization is achieved using vertical passive marker grids imprinted at various locations within the models. Results are presented for model runs with alpha = 12 and 45° and viscosity contrasts, m = 4 and 67. The results are in good agreement with theoretical predictions and highlight the importance of velocity gradients across the models due to the vertical extrusion component and the presence of the low viscosity inner zone. For various combinations of alpha and m, strongly contrasted strain geometries are developed in the inner and outer zones. These include the possibilities of forming a vertical lineation in the outer zone and a horizontal lineation in the inner zone, or a pure flattening fabric in the inner zone and a vertical lineation in the outer zone.

2096

L01 : 3B/26 : F4

Evolution and Abandonment of Major Faults During a Progressive Deformation: A Physical Analogue Modelling Approach

Veronique Meyer (veroniq@fag.esc.liv.ac.uk)

Fault Analysis Group, Dept. of Earth Sciences, The University of Liverpool, England

The purpose of this talk is to show how analogue modelling can contribute to understanding mechanisms of a progressive brittle deformation. Scaled sand-box experiments combining pure and simple shearing allow us to highlight the role of some early faults accommodating substantial part of the whole area finite strain. The place of major faults localisation appears to strongly depend on the deformed area shape which could before the faults initiation, distribute and concentrate stresses and deformation relatively to this shape. From this initiation, the geometrical characteristics of major faults are always directly related to the finite strain field magnitude and axes. Displacement gradients on major faults are much higher than those of secondary neighbouring faults that accommodate a less important component of the bulk strain and have a shorter evolution. Experiments have shown that propagation of these secondary faults is stopped or delayed by the major faults. An early dominant fault could actually only be crosscut by a recent fault when its displacement rate and energetic state would be much lower. We have analysed the progressive geometric evolution of some experimental major faults in terms of fault displacements, length, fractal dimension, rigid rotation and orientations. This study has revealed that fault evolution takes place in 3 successive frames. After fault initiation, the first stage rapidly involves a large increasing of the fault length without significant displacement rising. The second stage displays an important amplification of fault displacement rates and finite strain. Finally, according to strain magnitude and the vorticity number, the progressive and non coaxial strain behaviour results in a certain rotation of early major faults towards the principal lengthening axis. Consequently on the latest stage of growth, a major fault is likely to lie along a critical orientation relatively to the stationary boundary stress field and so experience a decrease in applied shear stress magnitude. As a consequence, displacement rates of dominant faults will decrease and the system energy will be reduced to a temporary minimum. At this point, the decelerating principal faults are progressively crosscut and connected by newly formed faults gradually linking the fault system. These new faults occur in the same orientation that the major faults initially had, illustrating the beginning of a new deformation cycle with another increasing energetic state. The frequently applied crosscutting criteria should therefore be used with caution.

1292

L01 : 3B/27 : F4

Experiments on Granite Intrusion in Transtension

Teresa Román-Berdiel (acasas@posta.unizar.es)¹,

Aitor Aranguren (goparira@lg.ehu.es),

Julia Cuevas (gppcuurj@lg.ehu.es),

José M. Tubía,

Denis Gapais (Denis.Gapais@univ-rennes1.fr)² &

Jean-P. Brun

¹ Departamento de Geodinámica, Universidad del Pais Vasco, Apartado 644, 48080 Bilbao, Spain

² Géosciences Rennes, Université de Rennes 1, 35042 Rennes Cédex, France

Granite intrusion in transtensional regime was modelled by injecting a Newtonian fluid (low-viscosity silicone putty) into a sand pack containing a ductile layer of silicone putty acting as a weak level along which the injected material could spread. The transtensional regime was obtained using two plastic sheets attached to two vertical end walls, sliding along two rigid horizontal plates, and diverging from two narrow spaces (two fixed velocity discontinuities, V.D.). The injection tube is located in a central space between these plates. Both symmetrical (when the two sheets are displaced with equal and opposite velocity vectors) and asymmetrical (in which only one sheet was displaced) V.D. experiments were performed. Transtension is applied with a systematic variation (every 15°) of the divergence angle (<alpha>), between 15 and 90. Experiments showed that (1) Intrusions localise strain from the first stages of deformation. (2) Intrusions result in partially conformable laccoliths with bowler hat geometry in cross-section. (3) Intrusions show an important offset towards the mobile basal plate in the case of asymmetric regime vs. more symmetric intrusions centered on the injection point for symmetric transtensional regime. (4) For low <alpha>-values (15) faults localise in the central part of the model, above the V.D.; intrusions are elongated in the principal extension direction determined by the bulk strain field and show a shape of asymmetric drop in horizontal section. (5) For <alpha> = 45 and <alpha> = 60 the main structure developed in the overburden is a listric fault above the intrusion. The injected silicone intrudes in the footwall of this structure spreading in the soft level. (6) For <alpha>-values of 75 and 90 the main extensional structures developed in the overburden are two graben localised on both sides of the intrusion. The injected material is affected by normal faults, and spreads in the silicone layer on both walls of the normal faults.

2828

L01 : 3B/28 : F4

Brittle Viscous Coupling in Extensional Regime

Thomas P-O. Mauduit (t.p.o.mauduit@siep.shell.com)¹ &

Jean Pierre Brun (brun@univ-rennes1.fr)²

¹ EPT-SG, Volmerlaan, PO Box 60, 2280 AB Rijswijk, The Netherlands

² Geosciences Rennes, UPR 4661, campus de Beaulieu, 35042 Rennes Cedex, France

At a crustal scale, either continental or oceanic, brittle upper crust overlays a lower ductile crust. At a basin scale, brittle overburden may overlay viscous layers (salt or undercompacted shales). On a mechanical point of view, understanding lithosphere deformation requires therefore a clear understanding of the effects of the mechanical coupling between ductile and brittle layers. Using laboratory experiments on brittle-ductile models, we have studied the effects of the deformation rate for different ratio of brittle viscous thickness, and different boundary conditions in extensional regimes. Brittle layer is represented with sand and underlying ductile layer with silicone putty. The experiments are compared to natural examples from both academic and industrial world (seismic section) and the kinematics significance and development of all type of extensional structures is studied. It is shown that coupling controls:

* the localisation of deformation,

* the width of the deformed zone,

* the amount of translation

* the amount of rotation.

Because these experiments are dimensionless the results can be apply whatever the scale. It becomes therefore possible to determine the depth of the viscous layer interface or to predict invisible structures on seismic sections.

2820

L01 : 3B/29 : F4

Analysis of Fault Zone Width and Splays in Extensional Faults by the use of Analogue Plaster Experiments

Roy H. Gabrielsen (roy.gabrielsen@geol.uib.no) &

Jill A. Clausen (jill.clausen@geol.uib.no)

Geological Institute, Univ. of Bergen, Ellégt.41, N-5007 Bergen, Norway

Because of its very fine-grained nature and hence, ability to preserve very delicate details, and because it solidifies after deformation, plaster-of-Paris is very well suited for detailed fault analysis. In the present study, a series of experiments have been performed to study the development of fault zone width and footwall and hangingwall splaying in extensional faults.

A mechanically homogeneous plaster sequence with passive colour markings was extended until gravitational collapse occurred. Approximately 30 experiments were used in the present series, and each experiment was documented by the use of still photographs and video recordings. In turn, these recordings were used in measuring layer thickness in the hangingwall and footwall during deformation, thickness of layers dragged into the fault zone as a function of fault displacement, time for disruption of such layers, vertical and horizontal displacements, fault dip angle, and time of initiation of splay faults. The continued development of single splay faults were also recorded in detail.

In these experiments, the initiation of the first master fault usually occurs after approximately 10-12% of extension. Both planar and listric faults are produced, depending upon the nature of the basement (hard or soft) of the experiment. After 20-30% of extension, the first splay fault commonly is initiated, nucleating at the master fault surface. In some experiments up to four splay faults are produced during the experiment, which ends after 40-80% extension.

For the behaviour of the layers dragged into the fault zone, it is concluded that the layers withstand ductile deformation up to considerable amount of displacement and thinning before disruption occurs. In this homogeneous seguence, the layers maintain a continuous, smooth surface during the intra-faultzone shearing. Thus the disrupture of the layers is ruled basically by the development of hangingwall splay faults

Several types of splay faults were produced:

Hangingwall splays are most common in cases where a smooth fault plane is produced. Three types are seen, namely those which are nucleated at the master fault propagating upsection (partly antithetic), those nucleated at the master fault surface propagating down-section (synthetic), and those nucleating above the fault plane, travelling up-section or down-section, finally coalescing with the master fault.

Footwall splays are less common, but may follow the same alternative patterns as those described for the hangingwall splays. The footwall splays seem to be more common in cases where the master fault surface is uneven, so that development of cut-off-faults will contribute to smoothen the fault surface.

By continued displacement, the splay faults are commonly seen to meet with other splays to define horses, and in some cases even extensional duplexes. The final result of this process is in most cases geometrically similar, so that it is not possible to distinguish between the horses that developed either from footwall or hangingwall cut-offs.

0047

L01 : 3B/30 : F4

Physical Modelling of the Relation between Displacement and Structural Trend at Corners of Master Structures

Peter Zweigel (peter.zweigel@iku.sintef.no)

IKU Petroleum Research, S.P. Andersens vei 15 B, N-7034 Trondheim, Norway

Displacement vectors in convergent or extensional settings are usually, as a first approximation, assumed to be either roughly parallel to or normal to local structural trend. However, application of this assumption in structural reconstruction leads to space problems in non-cylindrical settings, in which master structures such as plate margins or major faults are not straight but exhibit corners or bends.

Physical deformation models allow for observation of structural evolution and for quantification of strain in complex settings. Such models were carried out for the case of convergence (regional shortening). In these models, a Coulomb-type material (sand) was accreted at the corner of a box, simulating a plate margin that exhibited a corner of 90°, being convex towards the foreland. The main variable in these experiments was the angle of convergence obliquity (<alpha>), i.e. the angle between the regional shortening axis and the normal to the main segment of the plate margin, which was varied from 0 degrees ('frontal case') to 45° ('symmetric case'). All these models generated curved accretionary wedges which exibit, however, systematic differences of the arc-shape in map view, the amount of orogen-parallel extension, the pattern of map-view structures, and the pattern of displacement vectors (connecting initial with final positions of markers), depending on the angle <alpha>.

Thrust traces and fold axes are curved in the corner region but trend parallel to the plate margins elsewhere. Displacement vectors of accreted material exhibit fanning around the corner in all cases, diverging towards the foreland. In oblique indentation (<alpha> > 0°) the spread of their orientation is smaller than the change of structural trend (nappe traces and fold axes). This means that they have orientations intermediate between the normal to local structural trend and the regional shortening direction. This deviation from orthogonality increases with increasing <alpha>.

The results of these experiments (A) can most likely be extrapolated to other settings of non-straight master structures such as convergence across a plate margin with a concave corner towards the foreland (B), or extension at a master fault with a concave (C) or convex (D) corner towards the hangingwall. In all these cases, local structural trend might be strongly controlled by the strike of master structures (plate margin, major faults). However, mass transfer paths and local strain are expected to be in directions intermediate between the normal to local structural trend and the regional strain axis. Displacement vectors should thus converge towards the foreland in the convergent case (B), and converge or diverge towards the hangingwall in extensional cases (C) and (D), respectively.

2156

L01 : 3B/33 : F4

Abnormal Reverse Faulting Above a Depleting Reservoir

Francis Odonne (odonne@cict.fr)¹,

Isabelle Ménard,

Jean-Luc Bouchez,

Gérard J. Massonnat² &

Jean-Paul Rolando¹

¹ Université Paul Sabatier, UMR 5563, Pétrophysique, 38 rue des 36 Ponts, 31400 Toulouse, France

² Elf Exploration Production, CSTJF, avenue Larribau, 64018 Pau cedex

Subsurface deformation is observed during pumping of some hydrocarbon fields. Deformation features include subsidence centered on the field and subsidence-related centripetal horizontal displacements and faulting. Focal mechanisms yield reverse movements on steeply dipping faults. Similarly, focal mechanisms around magma chambers also show reverse displacements on steep and cone-shaped faults (Ekström, 1994). In our sand-silicone analogue model, the reservoir is represented by a latex balloon or by undercompacted ground sand. Deflation of the reservoir results in formation of steeply dipping reverse faults bounding a downward-opened cone. The cone moves downward to follow the reservoir contraction. Faults along the cone are straight beneath a thick reservoir cover and tend to curve upwards with decreasing cover. Calculations of Prucha et al. (1965) show that maximum principal stress steeply dips at depth where the void created by the depletion imposes a downward displacement due to gravity. As a result of the subsiding structure, sand is attracted to the center of the subsided area, as shown by centripetal displacements at surface. At depth, an oblique orientation of the maximum principal stress is no longer possible close to the surface of the model. Because sand follows the Mohr-Coulomb criterion for failure with a coefficient of friction close to 0.60 (Krantz, 1991), fault surfaces are oriented about 30° to the maximum principal stress in the models. This is why fault traces are curved, with steep dips being imposed by downward displacements at depth and, when they cut the surface of the model, low dips imposed by centripetal displacement and horizontal orientation of maximum principal stress at top. Segall (1989) proposed a numerical model in which the pore-fluid extraction results in a volumetric contraction of the reservoir rocks and triggers earthquake formation. His results give simulated focal mechanisms that are consistent with most of the observations made in oil and gas fields during fluid extraction. Calculated focal mechanisms give two equivalent planes but cannot, however, determine which is the actual one. We propose a new interpretation of these results by choosing the steepest nodal plane as being the active one above the reservoir. Numerical and analogue models lead us to conclude that volumetric contraction of a reservoir causes three-dimensional reverse coned-shaped faults to form in isotropic brittle material. These coned-shaped faults are steeply dipping which is abnormal for reverse faults (Anderson, 1951).

Anderson EM, Oliver & Boyd eds Edimburgh, 206 p, (1951).

Ekström G, E. P. S. L, 128, 707-712, (1994).

Krantz RW, Tectonophysics, 188, 203-207, (1991).

Prucha JJ, Graham JA & Nickelsen RP, A. A. P. G. Bulletin, 49, 966-992, (1965).

Segall P, Geology, 17, 942-946, (1989).

1339

L01 : 3B/34 : F4

Cracks in Starch: Simulation of Basalt Columns and Rock Joints

Gerhard Mueller

(gmueller@geophysik.uni-frankfurt.de)

Institute of Meteorology and Geophysics, Feldbergstr. 47, 60323 Frankfurt/Main, Germany

Desiccation of starch-water mixtures produces contraction cracks of various forms. A spectacular example is an irregular crack pattern which grows from the surface of a lamp-dried specimen into the interior, becoming polygonal and forming columns similar to basalt columns (Mueller, 1998). The reason for the similarity is that starch desiccation and basalt cooling are diffusion processes, with similar time-depth variations of water concentration and temperature. The scales of the processes are different, because their diffusion constants differ by a factor of about 100 (10^-8 m²/s in starch, 10^-6 m²/s in basalt): crack-front propagation in starch is about 10 times slower than in basalt, with the consequence that the depth gradient of (dimensionless) water concentration in starch is much larger than the depth gradient of (dimensionless) temperature in basalt. This in turn explains the large difference of the column diameters in starch (millimeters) and basalt (typically decimeters).

In an earlier stage of desiccation a few larger cracks occur which rupture the specimen from top to bottom. Rupture velocities cover the range from dynamic or critical crack growth (100-200 mm/s) to subcritical growth (< 1 mm/s); they were estimated from photos or measured from videos. Inspection of the crack surfaces reveals a broad spectrum of morphologies, in particular plumose structures (steps, crests, grooves) radiating from the nucleation point. They are similar to structures on joints in rocks. In starch layers with predominantly horizontal crack propagation the downward curved form of the plumose lines implies a general decrease of rupture velocity from top to bottom by a factor of 2-5. This is due to a decrease of tensile stress which, in turn, is due to the increase in water concentration. The horizontal variation of rupture velocity is less easily inferred from the plumose lines; we are currently developing an inverse method which is based on the approximate analogy of plumose lines and seismic rays. The topography amplitudes of the plumose structures are largest for the slowest, subcritical parts of a crack. This observation is opposite to the preferred assumption in tectonofractography (Bahat, 1991).

Bahat D, Tectonofractography, Springer, (1991).

Mueller G, J. Geophys. Res, 103, 15239-15253, (1998).

1L01 : 3B/35 : F4

Thermomechanical Analogue Modelling of Crustal Processes: First Experiences

Djordje Grujic (grujic@sun2.ruf.uni-freiburg.de),

Elmar Wosnitza (wosnitza@ruf.uni-freiburg.de) &

Jan Behrmann (behrmann@sun2.ruf.uni-freiburg)

Geologisches Institut, Albert-Ludwigs-Universität Freiburg, D-79104 Freiburg, Germany

Hereby we introduce a new investigative technique for thermomechanical modelling of crustal processes supported by a case study in Himalayas. To simulate crustal rheological stratification in physical models, it is necessary to take into account the variations in mechanical properties induced by temperature. So far this has been done by using different materials, such as sand and silicone putty, for brittle and ductile behaviour respectively (e.g. Davy and Cobbold, 1991). Major advances have been made in understanding the crustal and lithospheric processes in this way (e.g. Chemenda et al., 1995). However, the major drawback with such models is that the material points within the model crust retain their physical properties throughout the experiment regardless of their changing position within the crust. The second approach to model the crustal layering is to use a single material with temperature-sensitive viscosity. This design has the advantage to reproduce the mechanical consequences of thermal readjustment during the model run. To our opinion this is the most important direction in improvement of modelling technique for plate tectonic processes. This is particularly significant in the experiments investigating lithosphere stretching (Brune and Ellis, 1997) and the stability of mountain belts. Furthermore, it has been shown that the changing rheology of subducted sediments is an important phenomenon governing the material balance in the subduction zones (e.g. Mancktelow, 1995). This concept has been also applied to the exhumation in collision zones (Grujic et al., 1996). On order to investigate the convergence processes, in particular the role of rheology in distribution and propagation of deformation, we have constructed a deformation rig specially designed for thermomechanical modelling (Wosnitza et al., this volume). We use analogue materials with temperature-sensitive viscosity to properly simulate the change of mechanical properties with depth. Correlating the thermal with physical field in the model one can deduce its rheological structure. Using all three informations together the dynamics of the structures formed can be investigated.

Brune JN & Ellis MA, Nature, 387, 67-70, (1997).

Chemenda, AI et al, Earth. Planet. Sci. Lett, 132, 225-232, (1995).

Davy P & Cobbold PR, Tectonophysics, 188, 1-25, (1991).

Grujic Det al, Tectonophysics, 260, 21-43, (1996).

Mancktelow NS, J. Geophys. Res, 100, 571-581, (1995).

Wosnitza Eet al, J. Conf. Abs., 4, (1999).

0833

L01 : 3B/36 : F4

Kinematic Analysis of Rock Deformation Fabrics with the Coordinates of the Particle Path Field

Frank Schrader (schrader@geologie.uni-halle.de)

Institut für Geologische Wissenschaften, Martin-Luther-Universität Halle-Wittenberg, Domstraße 5, D-06108 Halle/Saale, Germany

In many analyses of rock deformation finite strain (Ramsay and Huber, 1983) is used as a standard coordinate system. It has three principal axes X, Y, and Z, which are orthogonal to one another, and which may rotate during progressive deformation. They may be determined with 2D sections parallel to the principal planes of the ellipsoids of deformed markers.

However, the particles of a deformed material may have been transposed to their configuration in finite strain by path fields of various geometry (e.g. Hoeppener, 1964; Ramberg, 1975). Such path fields may be determined from lineation arrays on the surfaces of rigid rock inclusions (e.g. pebbles, garnet nodules), which were displaced as undeformed particles relatively to their neighbours and to their matrix (Schrader, 1988). Each inclusion may be used as the reference point of the coordinates, and usually every inclusion in a limited volume of rock shows a similarly oriented lineation array. The inclusion surfaces are examined directly in 3D, the lineation arrays are interpolated with trajectories plotted with a permanent marker, and the arrays are recorded in stereoplots.

The arrays often show three principal axes (AD - lineation divergence / material contraction, AI - intermediate, AC - convergence/extension). Each pair of axes may make orthogonal or non-orthogonal angles. In cases of coaxial deformation, all three angles are orthogonal. In conventional cases of non-coaxial deformation (e.g. oblique shear or simple shear), the angle between AD and AC is non-orthogonal. In non-conventional cases, however, the angle between AD and AI may be non-orthogonal (divergence shear, similar to transpression sensu Sanderson and Marchini, 1984), or the angle between AI and AC (convergence shear, similar to transtension).

So the various types of geometry of progressive deformation may be defined in the particle path field by the various angles between its principal axes AD, AI, and AC much more precisely and graphically than in finite strain, the axes X, Y, and Z of which can only be orthogonal. Particle path field axes may also define the axes of (rotation or torsion) spirals (Schrader, 1997), the geometric effects of which cannot be represented with finite strain. So the ellipsoid of a deformed marker may have been produced by simple (as commonly supposed) or by rather complicated deformation geometry. Even the more complicated geometry results from single deformation events, as visible by the absence of lineation intersections.

Many deformation fabrics (e.g. winged porphyroclasts, pyrite shadows, schistosity surfaces and their lineations, etc.) may be understood much better in the coordinates of particle paths than in finite strain. Also mathematical and experimental modelling of rock deformation may be done much more effectively.

Hoeppener R, Felsmechanik und Ingenieurgeologie, 2, 22-44, (1964).

Ramberg H, Tectonophysics, 28, 1-37, (1975).

Ramsay JG & Huber MI, The Techniques of Modern Structural Geology, Acad. Press, London, 1, (1983).

Sanderson DJ & Marchini WRD, Journal of Structural Geology, 6, 449-458, (1984).

Schrader F, Journal of Structural Geology, 10, 41-52, (1988).

Schrader F, Freiberger Forschungshefte, C471, 201-203, (1997).

0606

L01 : 3B/37 : F4

Shape Preferred Orientation and Matrix Interaction of Three-Dimensional Particles in Analogue Simple Shear Flow

Laurent Arbaret (arbaret@erdw.ethz.ch)¹,

Neil Mancktelow &

Jean-Pierre Burg

Geologisches Institut ETH- Zentrum, Sonneggstrasse, 5, 8092 Zurich, Switezerland

The particle shape and symmetry influence the preferred orientation of rigid particles in a deformed linear-viscous matrix. We investigate this influence through a series of three-dimensional analogue experiments. The matrix is composed of a transparent polydimethyl-siloxane polymer (PDMS SGM36) characterised by a Newtonian behaviour with a viscosity of ca. 5 x 10⁴ Pa.s at constant shear strain rates <gamma>< 9.10^-4 s^-1. The ring apparatus used allows to carry out unlimited shear strain. Trajectories of the initially randomly oriented particles of polyethylene are compared with theoretical periodic paths for rigid ellipsoids.Quadratic, slightly orthorhombic and monoclinic particles have trajectory paths and rotation rates closest to theory. Orthorhombic and monoclinic particles, with a minimum aspect ratio > 1.5, develop non-periodic trajectories. The measured paths show that the simplified theoretical model for ellipsoidal forms is a very satisfactory approximation for a wide range of natural particle shapes.The relationship between rotation of rigid particles and the three-dimensional geometry of the adjacent passive matrix was also studied. The results are illustrated by the behaviour of the plane initially parallel to the shear plane and passing through the centre of the particle. This plane was constructed by combining surface profiles from a series of experiments for which the only variable was the depth of the particle below the surface grid. Two deformation domains are recognised in the matrix affected by particle rotation. In the domain close to the particle the matrix rotates along with the particle, both forming together an approximately ellipsoidal object. The second domain, slightly further away from the particle, is dominated by asymmetric folds verging towards the rotation direction. These folds have a limited axial dimension in the direction of the rotation axis and a large amplitude in the principal direction of maximum finite stretch. Hinges are increasingly curved with increasing shear strain and develop a sheath shape at very high shear strain.

L01 : 3B/38 : F4

Palaeopiezometry Using Steady State Microstructures

Paul D. Bons (paul@earth.monash.edu.au),

Mark W. Jessell &

Rosalind Klein

Epsilon Laboratory, Dept. Earth Sciences, Monash University, Clayton, VIC 3168, Australia

Grain size is an establish microstructural parameter for the determination of palaeo stresses. Grain size is one of the easiest parameters to measure and a theoretical framework and a large body of experimental data are available for this palaeopiezometric technique. Yet, deformation affects the microstructure as a whole and it would be advantageous if more parameters than only the grain size could be used.

To investigate the relationship between deformation rate and microstructure, we deformed the crystal-plastic organic rock analogue octachloropropane (OCP) in a transparent deformation cell with a cylindrical shear zone, capable of achieving arbitrary high shear strains. Microstructural developments during deformation were also modelled with a modelling package ELLE. The program ELLE is capable of simulating the operation and interaction of several concurrent processes acting on a microstructure.

As has been observed before, OCP exhibits a relationship between grain size and deformation rate and hence stress. Other parameters are also affected by the deformation rate in a systematic way:

- average grain shape (elongation)

- average grain orientation (grain shape preferred orientation)

- grain boundary characteristics (serrate, smooth, etc.)

The computer simulations, which involved grain boundary energy and internal energy driven grain boundary migration and subgrain formation, gave comparable results.

Our experiments and simulations are also useful to study the robustness of microstructural palaeopiezometric indicators. Each has a different reaction time to changes in deformation rate and has a different sensitivity to post-tectonic recrystallisation.

Session L01:3P

0207

L01 : 3P/01 : PO

Modelling the Evolution of Viscous Thrust Wedges Using Paraffin as an Analogue Material

Federico Rossetti (rossetti@uniroma3.it)¹,

Claudio Faccenna (faccenna@uniroma3.it)¹,

Giorgio Ranalli (granalli.ccs.carleton.ca)²,

Fabrizio Storti (storti@uniroma3.it)¹ &

Renato Funiciello (funiciel@uniroma3.it)¹

¹Dip. Scienze Geologiche, Università Roma TRE, Largo S. L. Murialdo 1 Roma, Italy

² Department of Earth Sciences, Carleton University, Ottawa, Ont. k1S 5B6, Canada

Orogenic wedges have been adequately described as a mechanical wedge-shaped entity with a rigid buttress at the back and resting upon a basal décollement Coulomb rheology sandbox models have been successfully used to model accretion at upper crustal levels. On the other hand, steady-state thermally activated ductile flow is dominant at depth in mature orogenic wedges. We modelled the evolution of thick orogenic wedges by using commercial macrocrystalline paraffin 52/54 EN type as a temperature-sensitive viscous analogue material, deformed in a specifically designed thermomechanical shortening machine. The used paraffin, for homologous temperature T/Tm (Tm is the melting temperature) higher than 0.69 (T > 37°C), behaves as an approximately Newtonian fluid (n varies between 1.0 and 1.3). In the Newtonian field and under natural gravity conditions, this material fulfils similarity criteria to model lower crustal rock flow, if stress is scaled by a factor 10⁶ and time by a factor 10¹¹, for an experimental strain rate of 10^-4-10^-5 s^-1. Temperature is not considered in the scaling procedure: it constitutes a controlling factor to obtain mechanical similarity between model and nature. Experiments were performed in an insulated room at constant temperature of 37°C. To obtain the required temperature profile inside the paraffin model, the base is constantly maintained at a temperature of 40°C by an heating system. The temperature gradient within the model produces a  of two orders of magnitude, from 10⁷ to 10⁵ Pa s. In this work we present results of an experimental research programme specifically designed to investigate the influence of shortening rate variations on the evolution of orogenic wedges accreted against a vertical, rigid backstop. The undeformed experimental configuration was wedge-shaped, with a basal taper ß = 5° and a surface taper <alpha> = 0°. Model wedges evolved with different geometries and kinematics, at different shortening rates. Lowering the shortening rate favoured the outward migration of the deformation front and a decrease of the uplift rate in the axial zone. The aspect ratio of the wedge (i.e. the ratio between the maximum height of the wedge and the distance of the deformation front from the backstop) has been found useful for describing the evolution of model wedges. Plots of the aspect ratio versus shortening show that the experimental wedges, for a given shortening rate, attained a characteristic critical stable profile, that is: the wedge grew in a self-similar mode, with a constant aspect ratio when the load of the accreted material reached a critical value. This in turn indicates that increasing convergence, the mechanical balance between vertical and horizontal stress components was attained during wedge accretion.

0851

L01 : 3P/02 : PO

Studying the Geodynamics of the Tyrrhenian-Apennines System in the Lab: The Analogue Modelling Strategy of the Experimental Tectonics Laboratory at the Roma Tre University

Renato Funiciello (funiciel@uniroma3.it),

Valerio Acocella (acocella@uniroma3.it),

Claudio Faccenna (faccenna@uniroma3.it),

Federico Rossetti (rossetti@uniroma3.it),

Francesco Salvini (salvini@uniroma3.it) &

Fabrizio Storti (storti@uniroma3.it)

Dipartimento di Scienze Geologiche, Università Roma Tre, Largo S. L. Murialdo 1, 00146 Roma, Italy

The geodynamic framework of the Tyrrhenian Sea-Apennines system results from the complex interaction between the European and African plates. A characteristic feature is the eastward, space-time migration of the compressional-extensional pair from the hinterland outwards. In the inner sector, Tertiary extensional tectonics overprinted the early crustal thickening phase, accompanying the unroofing of deep-seated units and leading to the opening of the Tyrrhenian Sea basin. Magmatic activity developed during extension, with a similar eastward space-time migration. At present, the topographic divide area of the Apennines corresponds to the outer boundary of the extensional domain. Moving eastward, compressional tectonics is still active, particularly in the frontal region of the Apenninic wedge. The coexistence of both the extensional and compressional regimes makes the understanding of their genetic mechanism crucial to unravel the evolution of the western Mediterranean area. In the last two years, in our Department analogue modelling techniques have been flanked to traditional geological investigations for providing additional tools to approach this geodynamic complexity. Three main modelling techniques are used, depending on the scale of the natural processes investigated. Lithospheric-scaled modelling has been used to simulate the overall geodynamics of the region and, particularly, to investigate the dynamics of subduction. Sandbox, crustal-scaled modelling has been used to study upper crustal processes, in both extensional and compressional regimes. The main topics consist of the evolution of thrust wedges in a dynamic environment, where tectonic and surface processes interact, on the extensional re-activation of pre-existing thrust ramps, and on the development of transfer faults in extensional domains. Thermomechanical modelling using commercial paraffin as analogue material has been used to study the evolution of thick orogenic wedges and their gravitational instability. At present, two main additional research programmes are flanking the above mentioned topics. They consist of the thermomechanical modelling of extensional processes, and of the modes of emplacement of intrusive bodies in the upper crust.

1032

L01 : 3P/03 : PO

Compressional Structures in A Multi-Layered Mechanical Stratigraphy: Insights from the Sand-Box Modelling

Claudio Turrini (claudio.turrini@fina.be)¹ &

Antonio Ravaglia (aravaglia@manhattan.unipv.it)²

¹ FINA Italiana S.p.A., viale Premuda, 27, 20129 Milano, Italy

² Dip. Sc. Terra, Università di Pavia, via Ferrata, 1, 27100 Pavia, Italy

To investigate the influence which a vertically non-homogeneous stratigraphy plays on the deformation kinematics developing in a thin-skin type compressional regime a set of experiments were performed in a simple glass-side sand-box machine. Different types of dry, granular, non-cohesive materials were used to produce the initial mechanical undeformed package: strong layers alternated vertically with weak ones, the latter proved to provide a good mechanical analogy to localise disharmonic deformation and modelize decoupling surfaces. In order to force variability of the deformation along the strike of the developing structures different solution were adopted: on experiments 1, 3, 4, 5 & 9 a rigid, undeformable ramp, alternatively normal or oblique with respect to the shortening direction, was positioned to occupy a part of the basal surface of the apparatus, on the far away side from the advancing piston; experiments 2, 6, 7 & 12 were run making the model to squeeze on a basal surface composed of two different pieces of plastic sheets the two being characterised by a strong friction coefficient contrast; on experiments 10 & 11 a part of the basal layer was attached to the sand-box surface, while on experiments 7 & 8 the external sector of the model was impregnated with water before deformation started. Deposition of new material during experiments 1, 2 & 9 was performed to simulate sin-tectonic sedimentation occurring at the front of the evolving synthetic fold and thrust belt. On the other hand erosion of the growing structures at the top of the innermost part (near the advancing piston) of the deforming package was arbitrary acquired during experiment 1.

The final models resulted in the description of a wide range of deformation geometries, mechanisms and kinematics. Different order of fold-fault pairs, widespread intra-stratal decoupling, sub-thrust geometries and sin-sedimentary embricates contribute to the deformation evolution and they demonstrate the critical role that the initial mechanical stratigraphy, the deposition of sin-tectonic material, the erosion of the deformed structures and the form/position of the rigid ramp feature plays in constructing the related thrust belt architecture.

The comparison of the modelled geometries with selected structural associations interpreted and described by the literature from the Southern Alps and the Apennines suggests that strong similarity exists between the two set of data in terms of both deformation geometries and kinematics.

3564

L01 : 3P/04 : PO

Analogue Modelling of the Significance of Erosion and Redeposition During the Rise of Orogenic Wedges

Katarina S. Persson (Katarina.Persson@geo.uu.se) &

Dimitrios Sokoutis (Dimitrios.Sokoutis@geo.uu.se)

Hans Ramberg Tectonic Laboratory, Department of Earth Sciences, Uppsala University, Uppsala, Sweden.

Analogue sand models were used to analyse the effects on the development of orogenic wedges of syn-shortening erosion with local redeposition. Cool strong continental crust was represented by a rigid indenter with a frontal dip of 75° which was driven laterally into sand representing a hanging wall of brittle crust. Previous experiments without erosion showed that an effective indenter with a front face dipping at 30° accreted in front of the rigid indenter. The present experiments explore the effect of erosion and redeposition on the growth of effective indenters. Different proportions of the extruding wedges were eroded and redeposited forwards onto the footwall and backwards onto the hangingwalls.

The model results indicate that surficial redistribution can significantly change the geometry and rates of development of orogenic wedges. Erosion decreases vertical loading above the active thrusts while deposition loads the margins. As a result, fewer thrusts remained active for longer periods than equivalent models without erosion. Moderate redistribution (as well as total erosion) narrowed and steepened subsurface structures without the development of an effective indenter. The effective indenter deformed internally during limited erosion.

Models with Initial Overburden Thickness of 1.0cm and Rigid Indenter Dip of 75° Illustrating the Effect of (A) no Erosion, (B) Limited Erosion and (C) Moderate Erosion. Shaded is the Area of the Internally Deformed Effective Indeter in (B)

3717

L01 : 3P/05 : PO

Material Transfer within Active Margins Associated with Seamount Subduction ­ Insights from Sandbox Experiments

S. Dominguez (stephdom@dstu.univ-montp2.fr),

S. E. Lallemand (lallem@dstu.univ-montp2.fr) &

J. Malavieille (malavie@dstu.univ-montp2.fr)

UMR 5573, Université Montpellier 2 - CNRS, Case 60, 34095 Montpellier, France

Large domains of oceanic plates are characterized by a very rough seafloor, covered by seamounts, aseismic ridges and volcanic plateaus. As revealed by geophysical data, proceeding from oceanographic cruises, the subduction of these topographic highs, carried by the subducting oceanic crust, strongly interacts with the structural evolution of the overriding plate. With the aim to improve our understanding of these highly deformed region of active margins, we have performed a series of sandbox experiments including the analysis of structures in 3-D. The results of these experiments show that, in a such tectonic context, seamount subduction generate significant material transfer within the margin. The subducting high deviates upward the basal décollement in its wake. This mechanisms generates a shadow zone behind the trailing flank of the seamount (Figure 1). As a result, part of the frontal margin, located in this area, is dragged into subduction. The shadow zone is active as long as the décollement remains in an upper position which depends on the basal friction. Finally, the décollement returns to its basal level which greatly reduced the volume of the shadow zone and caused underplating of the sediments subducting in the wake of the seamount. This mechanism induces the subduction of frontal margin materials to greatest depths and generates a frontal erosion of the margin. The underplating of large volumes of relatively undeformed, water laden sediments, beneath the rear part of the margin could also result in large variation of pore fluid pressure and thus the effective basal friction.

Sandbox experiment showing the shadow zone which is generated in the wake of the subducting seamount.

L01 : 3P/06 : PO

Evolution of Slabs in the Upper Mantle: Results from a Comparison between Laboratory and Numerical Experiments

Claudio Faccenna (faccenna@uniroma3.it)¹,

Thorsten W. Becker (becker@eps.harvard.edu)² &

Richard J. O'Connell (oconnell@geophysics.harvard.edu)

¹ Dipartimento di Scienze Geologiche, Universita di Roma TRE, Largo S. L. Murialdo-00146- Rome, Italy

² Harvard University, EPS, 20 Oxford St., Cambridge MA 02138, USA

We conducted laboratory and numerical experiments to model the formation and evolution of the subduction of oceanic lithosphere in the upper mantle underneath an overriding cratonic plate. This joint approach allows us to estimate the importance of the simplifications made in each case and verify the individual results.

In the laboratory experiments, we employ a stratified rheological profile that is scaled to the Earth. A sand layer represents the brittle behaviour of the upper crust, silicone putty with Newtonian viscosity and variable densitymodels the creeping behaviour of the lower crust and mantle lithosphere, and glucose syrup simulates the weaker asthenosphere. In the numerical experiments, we use ConMan, a two-dimensional finite element code for convection modeling a Newtonian fluid with lateral viscosity and density heterogeneities.

We evaluate the influence of the following parameters on the style of subduction: velocity of convergence, density and viscosity contrast between the oceanic plate and the upper mantle and slab anchoring at depth. For each set of experiments, we analyze the time-behaviour of the flow-field in the mantle, the deformation state of the slab and of the upper plate, the velocity and the dip of slab and the trench migration. These effects depend on the dynamical equilibrium between acting and resisting forces which can be expressed as the dimensionless buoyancy number F.

The style of subduction is similar in the two model approaches although their rheology differs in complexity. Slabs nucleate in the form of a broad instability, first growing at a low rate and then sinking into the mantle obtaining the form of a slab. Their velocity and their dip increases steeply in a time-span of Myr before reaching the 670 km discontinuity. We observe an upper limit for the viscosity contrast between oceanic plate and mantle of about 500, for stiffer plates the tectonic style changes from subduction to underthrusting. The negative buoyancy of the oceanic plate represents the main acting force in the system, as expected, whereas the viscous bending of the oceanic plate at thes ubduction zone represents the major resisting one. The competition between those two forces exerts the major control on the style of subduction, the mode of trench migration and the state of stress in the upper plate.

L01 : 3P/07 : PO

Influence of Subducting Angle and Fluid on the Patterns of Wedge Flow

Ryo Anma (anma@arsia.geo.tsukuba.ac.jp)¹ &

Dimitrios Sokoutis (dimitrios.sokoutis@geo.uu.se)²

¹ Institute of Geoscience, Tsukuba Universiyu, Ten-no dai, Tukuba, Ibaraki, Japan

² Hans Ramberg Tectonic Lab., Uppsala University, Villavagen 16, 75236 Uppsala, Sweden

We present analogue models of the patterns of wedge flow induced by subduction of a rigid plate sinking obliquely along its length. To simplify, we use a transparent viscous fluid (SGM₃₆: µ = 5 x 10⁴Pa s) to represent ductile mantle wedge beneath the hot and weak crust of continental margins and a plexiglass plate to mimic stiff subducting oceanic slab. Patterns of wedge flow induced by the subducting slab were visualized by printing a passive carbon grid along the vertical midplane of SGM₃₆. Influence of subduction angle and buoyant fluid inserted along the top boundary of the sinking slab on the pattern of wedge flow were studied.

The SGM₃₆ bound to the non-slip boundary along the top of the sinking slab was carried downward with it due to model subduction. Vertical marker lines show horizontal flow toward the trench beneath the top free surface implying that the model 'back-arc basin' undergoes horizontal extension normal to the trench axis. The most intense horizontal surface extension overlies the upwelling above the leading edge of the subducting plate and migrates away from the trench as the wedge flow evolves. Steep oblique subduction results in pervasive back-arc extension that starts close to the trench, whereas gently-dipping subduction results in more localized and more intense back-arc spreading further from the trench. Arc-normal extension was previously attributed to suction induced by the roll-back of the subducting slab. We propose that both back-arc extension, plus migration of the extension zone, and upwelling of mantle material can be attributed to subduction of the oceanic lithosphere alone.

A thin layer of buoyant fluid (Hyvis 2000, µ = 9.2 x 10³ Pa s) inserted along the top boundary of the subducting plate decouples the mantle wedge from sinking slab by lowering friction along the top of the slab. Wedge flow in our lubricated model has similar form but evolves more slowly and is therefore less intense than a dry wedge with similar dip. A natural example is seen along the Ryukyu trench-arc system where such magma lubricated subduction zone with intense volcanic activity in the northern domain lacks in active back-arc extension, whereas dry subduction zone in the central and southern domains involves intense back-arc extension.

1721

L01 : 3P/08 : PO

Single and Multi-Layer Folding in Non-Linear Material Models

Tatiana Tentler (tatiana.tentler@geo.uu.se) &

Genene Mulugeta (genene.mulugeta@geo.uu.se)

Department of Earth Sciences, Mineralogy-petrology, Uppsala University, Uppsala, Sweden

Experimental data on creep flow of natural rocks indicate that they have a non-linear rheology. Here, we use non-linear materials compressed parallel to the layering to study single and multi layer folding in laboratory experiments.

In single layers, we study how the strain-hardening and/or softening rheology of the competent material influences the growth rate of folds as they mature to finite amplitudes. The growth rate of (brittle) strain softening single layer folds is twice that for (ductile) strain-hardening folds, for the same initial competence contrast between layer and matrix. Fold shape in the strain hardening single layer is sinusoidal and exhibits hinge-thickening and limb-thinning, as the folds grow to finite amplitude. Folding in the strain-softening single layer initiates with a sinusoidal wave form and becomes chevron-like when faulting nucleates at the hinges.

In multilayer folds, we investigate the role of viscosity contrast and ratio of incompetent to competent layer thickness in controlling fold geometry and growth rate. For a high viscosity ratio of competent to incompetent layer and low to moderate ratio of incompetent to competent layer thickness, the multi layer folds develop a combination of concentric and regular chevron fold forms. The wavelength thickness ratio in a competent layer is large. With time, the incompetent material migrates from the limbs to the hinges, resulting in thickening of hinges and thinning of fold limbs. Decrease in ratio of incompetent to competent layer thickness (for the same viscosity contrast) results in enhanced growth rate of the fold perturbations.

By studying folding in non-linear materials it is possible to study fold formation in rocks and it may be possible to infer the rheological properties of natural folds during their growth to finite amplitudes.

1010

L01 : 3P/09 : PO

Topographic Analyses of (Analogue Models on) Superposed Folds

Thomas Walter (thwalter@ruf.uni-freiburg.de)¹,

Djordje Grujic

grujic@sun2.ruf.uni-freiburg.de)¹,

Hansjörg Gärtner

(gartner@ito311.ito.uni-stuttgart.de)²,

Jean-Marc Nivet

(nivet@ito311.ito.uni-stuttgart.de)²,

Christoph Voland (voland@ito.uni-stuttgart.de)² &

David Tanner (dtanner@ugcvax.dnet.gwdg.de)³

¹ Geologisches Institut, Albertstr. 23-B, Albert-Ludwigs-Universität Freiburg, Germany

² Institut für Technische Optik, Universität Stuttgart, D-70569 Stuttgart, Germany

³ Geochemische Institut, Universität Götingen, D-37073 Göttingen, Germany

To form a mental picture of intricate 3D shapes of superposed folds is still a problem in structural geology meeting the limits of human imagination. Geological phenomena can be simulated in a simplified form by analogue models to ease the analysis of possible solutions. The complex 3D problems, however, allow a lot of space for subjective interpretation. We present a new method for quantitative analysis of complex forms, like surfaces of multiply folded layers. In an earlier performance of analogue experiments with paraffin wax a series of pre-formed cylindrical folds were deformed to produce Type 1 and Type 2 interference patterns (Grujic, 1993). Roundness, tightness and amplitude were varied to investigate the influence of first-fold geometry on the interference patterns. After experimental deformation, the matrix on one side of the competent layer was removed to allow the observation of the refolded surface. Originally, this analysis was only descriptive. To make the analysis accurate the multiply deformed surface should be statistically appraised. The topographic database fur such an analysis was in our procedure obtained with a newly developed optical triangulation system. The high-resolution (ca. 0.2 mm) surface digitising is based on the concept of structured light measurement by "double-scan" technique (Gärtner et al., 1996). The so acquired database consisting of xyz co-ordinates allows a complex surface to be represented and analysed in 3D. Three neighbouring surface points define a triangular sub-surface. Assigning virtual geographic co-ordinates to the digital model allows the orientation measurement of sub-surfaces and application of standard analytical methods (e.g. Stereographic projection-techniques, Eigenvector analysis). It is also possible to study linear structural elements (fold axes), geometric planes (axial surfaces), or selected parts of the model. The stereographic diagrams of digital models with fully known geometry (and deformation conditions) can thus be compared to natural examples and used for their interpretation. The accuracy of analysis of analogue models opens new ways of understanding the nature. In example of superposed folds was hereby discernible that the finite geometry of a multiply deformed surface is primarily controlled by the geometry of the initial folds, to a lesser extent by the strain geometry, and only in rare cases can this geometry be ascribed to a single type of interference pattern.

Gärtner H, Lehle P, Tiziani HJ & Voland C, SPIE, 2784, 21, (1996).

Grujic D, J. Struct. Geol, 15, 293-307, (1993).

1507

L01 : 3P/10 : PO

An Experimental Approach of Fold Development and Progressive Strain Distribution in Single Layers Under Non-Coaxial Flows

David Soler (geotec@geologia.uab.es) &

Albert Griera

Dept. Geologia, Universitat Autonoma de Barcelona, 08193 Bellaterra, Spain

A series of experiments involving compression of a single layer rock analogue has been carried out in the BCN deformation laboratory at the Universitat Autonòma de Barcelona. The models have been performed under pure shear and non-orthogonal pure shear flows with constant area and vorticity values. The used materials consisted in paraffin waxes with power-law rheology: a soft matrix with steady state flow, and a stiff layer showing strain softening, both at the experimental conditions. The bulk vorticity of the flow is determined by the imposed boundary movements of the experimental cell. Two main different sets of experiments have been studied, (i) one with the layer oriented parallel to the shortening apophyse or eigenvector of the bulk flow and (ii) another with the layer oriented at different degrees with respect to the shortening eigenvector flow.

Some analysis and results concerning the obtained different fold structures are presented. A great amount of finite and incremental deformation parameters has been calculated for every progressive deformation stage. The treatment of these parameters contributes to the knowledge of different aspects related to fold development. 1) Elaboration of distribution patterns of strain parameters around folding structures, as the intensity and orientation of the strain ellipse, vorticity and area change. This approach provides data on the role of the folds asymmetry, either as a function of the vorticity flow or as a function of the layer orientation with respect to the kinematic axes. 2) Variation of the dominal wavelengths, growth rates and fold limbs rotation. This approach provides new information for the understanding of the buckling process in competent layers with parallel shear strain.

This type of experimental approach attempts to establish a link for the best understanding of similar structures developed in natural rocks under the widely common general shear deformation regimes.

0353

L01 : 3P/11 : PO

Physical Modelling of Cenozoic Structures in Huanghua Basin, North China

Jianxun Zhou (jxzhou@www.bjpeu.edu.cn)

Department of Earth Science, University of Petroleum-Beijing, Changping, Beijing, P. R. China, P. R. China, 102200

Huanghua rift basin of Cenozoic age is located in the centre of Bohai Bay Basin, which is one of the most important oil-and-gas bearing basins in Eastern China. The basin, which generally trends NNE, possesses a kinked outline. The strike of the basin boundary changes from NNE in the southern part to near E-W in the middle and northern parts. Strike of normal the faults in the basin varies significantly form NNE in the south to near E-W in the middle and to NE in the north. The change of fault strike from NNE to near E-W takes place along an approximately N-S discontinuous zone, which is more or less parallel to the shoreline of the basin. In addition, E-W striking faults are common over the whole basin. There are different interpretations regarding the mechanism responsible for the formation of faults with different orientation in this basin. In this presentation, sandbox models are used to model the Huanghua Basin in order to put some light on the evolution history of the basin. A series of sandbox experiments were conducted, using loose quartz sand of 0.3-0.6 mm in grain size. Models, in accordance with the actual geometry of the basin scaled at 1:270,000, were extended by 20% at a rate of 5x10^-3s^-1 to 8 x10^-3s^-1. Model results suggest that a single episode of N-S extension could be responsible for the rifting and the formation of faults with different orientations. In the models, variation in strike of the faults was caused by the irregular orientation of the basin boundary. Model results also suggest that the discontinuous N-S trending zone in the Huanghua Basin is an accommodation zone caused by the irregular geometry of the basin boundary, rather than being a shear zone outlining a major tectonic boundary. It is emphasised here that this study does not agree with previous interpretations of the basin where extension direction of the basin is considered to be NW-SE, and the faults with different orientation are interpreted to have formed due to multi-phase deformations. This study is also in disagreement with previous works interpreting the N-S trending discontinuous zone to be a major shear zone. Based on the model results, a similar mechanism may be responsible for the formation of and the tectonic evolution in the whole Bohai Bay Basin.

0890

L01 : 3P/12 : PO

Modelling of Sedimentation as a Response to Extensional Tectonics in the Bohai Bay Basin Province

Jiafu Qi (jqi@bjpu.bjpeu.edu.cn),

Qiao Yang,

Hengmao Tong &

Kezheng Lu

Dept. of Earth Science, University of Petroleum-Beijing, Changping, Beijing Beijing, China 102200, China

Balance sections and sandbox experiments reveal that the morphology of a basin and its sedimentary fill are influenced by the geometry of major extensional faults and deformation pattern of the hanging wall. Results of sand-box models show that simple half-grabens form on the hanging wall of a simple listric fault, whereas on the hanging wall of a ramp-flat fault (complicated listric fault) a comparatively complicated half-graben forms which embodies secondary anticlinal uplift and synclinal sag. The width of such a half-graben basin is directly proportional to the detachment depth of the major fault. Results of extended sand-box models show similar features as those seen in the Bohai Bay basin province. This basin province includes 6 basin or over 60 extensional sags filled by Eocene and Oligocene sedimentary units. Most of the sags are styled in half-grabens. Within the sags, the widths of Oligocene half-grabens overlapped on Eocene half-grabens are almost equal to the widths of Eocene half-grabens, but the sequence construction of "apparent regressive overlap" is developed in the slope of the half-graben basin. Within some basin which embodies secondary anticlinal uplift and synclinal sag, the crest line of the Oligocene uplift migrated away from the crest line of the Eocene uplift toward the major fault. The results of sand-box models help us to suggest most of the sags in this basin were developed on the hanging wall of either simple listric or ramp-flat normal faults. As the results of extended sand-box models the sedimentation in the Bohai Bay basin province is a response to extensional tectonics. With progressive extension, an "apparent regression" and a "lateral accretion" sequences were formed in the Eocene-Oligocene sags in the basin.

1880

L01 : 3P/13 : PO

The Role of Relief in Extensional Settings: the Contribution of Analogue Modelling

Roberta Giuliani (giuliani@ssn.dstn.pcm.it)¹,

Valerio Acocella (acocella@uniroma3.it)²,

Claudio Faccenna (faccenna@uniroma3.it)² &

Renato Funiciello (funiciel@uniroma3.it)²

¹ Servizio Sismico Nazionale, Via Curtatone, 3, Roma, Italy

² Dipartimento di Scienze Geologiche, Università di Roma Tre, Roma, Italy

The dynamic role of a pre-existing relief has been widely recognized in several orogenic belts subject to extensional collapse as well as in different volcanic environments. Analogue models, simulating the stress field induced by large volcanoes, have previously provided some insights on the dynamic role of a pre-existing relief. A sequence of analogue models has been now performed at the Experimental Laboratories of Roma TRE to study the role of a pre-existing relief in the development of the deformative pattern during extension. In our experiments, the rheological behaviour of the brittle, upper crust has been modeled by using a pack, with variable thickness, of dry quarz sand; newtonian silicone putty has been used to reproduce the rheological behaviour of the ductile lowercrust. Two different initial configurations of the apparatus were performed. The initial configuration of the first set was characterized by a velocity discontinuity at the base of the sand and silicone layers, in order to focus the deformation; the initial configuration of the second set consisted of a basal rubber sheet, in order to induce a widespread deformation. Both set of models were characterized by different geometries and thickness values of the topographical relief as well as by different values of the brittle-ductile thickness ratio. The results of both sets show that the models undergo a similar progressive deformation, characterized by the development of a rift, bordered by sinthetic and antithetic normal faults; the width of the rift depends upon the initialconfiguration of the apparatus, being larger in the case of a basal rubbersheet. A shift of the rift axis towards the relief systematically occurred in both sets of experiments. The results show a proportional correlation between the thickness of the topography and the amount of shifting of the rift axis, allowing to define a critical thickness value below which no shifting is observed.

1735

L01 : 3P/14 : PO

Experimental Modelling of Strike-Slip Faults, Results and Constraints

Hans Peter Steyrer

(hans-peter.steyrer@sbg.ac.at) &

Martin Schöpfer

Institut für Geologie und Paläontologie, Universität Salzburg, Hellbrunner Strasse 34, A-5020 Salzburg, Austria

Analogue modelling has provided new insights into the progressive evolution of fault systems but few analogue model studies have adressed the progressive evolution of strike-slip faults. We ran a comprehensive suite of sandbox models with a straight offset fault at the base and step by step increasing thickness of the homogeneous sand layer. We ran the experiments with dry quartz sand (125-200 µm), which is widely used to simulate brittle deformation in the upper crust. With increasing offset (0-5 cm), a set of Riedel shears appears and propagates away from the fault zone. On both sides of the master fault an area is influenced by these propagating Riedels, which is ca. 1.5 times the thickness of the sand layer. Parallelogram-shaped lenses, confined by these Riedel-shears and by high-angle reverse faults rise, the elevations reaching 10% of the thickness of the sand layer. At shallow levels the reverse faults are overprinted by step faults. With increasing offset the initial lenses are first rotated, then cut by new synthetic Riedel shears until finally the central fault cuts through all previous structures. With increasing thickness (from 1 to 5 cm) the diagonal sizes of the firstly formed lozenge shear lenses also increase, their length being 2-3 times the thickness of the sand layer. Experiments appear to be reproducable only if the sand is applied very homogeneously with a special strewing apparatus providing constant height of fall and thus constant rate of compaction over the whole experimental area. Even slight inhomogenities in compaction of the sand layer cause distinct changes in the sizes of the initial parallelograms.

1664

L01 : 3P/15 : PO

Analogue Models of Calderas and Resurgent Domes

Valerio Acocella (acocella@uniroma3.it),

Francesca Cifelli (cifelli@uniroma3.it) &

Renato Funiciello (funiciel@uniroma3.it)

Dipartimento Scienze Geologiche. Università Roma TRE. Largo S.L. Murialdo, 1. Roma, Italy

Calderas and resurgent domes constitute a common association in volcanic areas and are usually considered to be related to inflation-deflation processes inside magma chambers. The deformation pattern associated to calderas and resurgent domes is the result of repeated, superimposed vertical movements; these are responsible for the development of inward and outward dipping ring fractures and radial fractures; the presence of a regional stress field may further complicate such pattern. In order to study the mechanism of dvelopment of calderas and resurgent domes, we have performed a sequence of analogue sand-box models at the Experimental Laboratories of Roma TRE. We used two newtonian silicone putties, with different viscosities, to simulate the rheological behaviour of magma and of the ductile crust overlying the magma chamber; well-sorted, dry quartz sand was used to simulate the rheological behaviour of the upper crust. The crustal profile was simulated by means of a 3 to 5 cm thick horizontal layer of sand and by a 1 to 2 cm basal thick layer of silicone putty. Different volumes of silicone putty with lower viscosity were injected at the base of the horizontal silicone and sand layers at different velocities, comprised between 3 and 10 mm/h, imposed by the scaling. Three sets of models have been performed: First, we simulated the formation of calderas as a consequence of a decrease in the pressure injection, studying the related deformation pattern. Second, we reproduced the formation of resurgent domes as due to an increase in pressure injection. Third, we superimposed the development of resurgent domes on pre-existing calderas and vice-versa, studying the deformation pattern resulting from such interaction. In these set of models the influence of a regional stress field was not taken into account: such contribute will be analyzed in a future set of experiments. Finally, the experimental results have been semi-quantitatively compared with several examples of calderas and resurgent domes worldwide.

0797

L01 : 3P/16 : PO

Thermomechanical Modeling: Soon Reality?

Elmar M. Wosnitza (wosnitza@uni-freiburg.de),

Djordje Grujic (grujic@uni-freiburg.de) &

Jan H. Behrmann (behrmann@uni-freiburg.de)

Geologisches Institut, Universität Freiburg, Albertstraße 23 B, 79104 Freiburg, Germany

We present a newly designed apparatus for thermomechanical analogue modeling. The deformation rig offers the ability to generate a vertical thermal gradient: The lower plate can be heated up to 150^oC while the surface is maintained between 20^oC and 80^oC. The temperatures are recorded at three positions with an accuracy of 0.01^oC and in the steady state are constant within 0.02^oC. The machine offers a volume of 35 l, allowing to scale models to gravity. A maximum of around 60% shortening in horizontal direction is possible.

The practicable temperature gradients can be used to control the viscosity of suitable analogue materials. We are using substances with temperature-sensitive viscosity varying by a factor of 2.5 up to 9 orders of magnitude in the temperature range. Polydimethylsiloxane (PDMS) offers a range of viscosities between 2x10⁴ and 5x10⁴ Pa s, gum rosin (colophony) offers 4x10³ to 10⁷ Pa s and Parafin wax 10 to 10¹⁰ Pa s.

Since the confining glass side walls have the same thermal diffusivity as the analogue materials they map the temperature distribution of the model side to the outside of the glass. This 2D temperature field can be measured in real time using an IR-camera with a resolution of 0.03^oC.

The horizontal deformation as well as the fluid level at two positions at the surface is measured with an accuracy of 1 µm. By measuring the deforming force with an accuracy of 0.1% the applied stress can be calculated. Together with strain rates between 10^-6 and 4x10^-5s^-1 (constant within 5%) this allows to estimate the bulk rheology.

During the test phase the 3D temperature field has been measured in PDMS and colophony using a PT100 probe. The results shows that the IR images faithfully represent the temperature distribution within the model because of the homogeneity of the field in the line of view of the camera parallel to the bulk y strain axis.

Thus scaling down linear dimensions by a factor of 10⁶, time by 10¹¹, viscosity by 10¹⁷ and temperature by 10¹ we can obtain a consistent set of scaling factors to reach geometrical, kinematic, dynamic and rheological similarity necessary for true thermomechanical modeling of lithospheric processes.

3008

L01 : 3P/17 : PO

Cataclastic Solution Creep of wet Polycrystalline Sodium Chlorate Aggregates

Mohsine Zahid (zahid@mail.uni-mainz.de)¹ &

Bas den Brok (denbrok@erdw.ethz.ch)²

¹ Johannes Gutenberg Universitaet Mainz, FB 22, Institut fuer Geowissenschaften, Tektonophysik, Becherweg 21, D-55099 Mainz, Germany

² Geologisches Institut ETH, Sonneggstr 5, CH-8092 Zürich, Switzerland

Uniaxial dead-weight creep experiments were carried out on wet synthetic polycrystalline aggregates of sodium chlorate (NaClO₃) at room temperature and atmospheric pressure. Aim of the experiments was to study the deformation mechanism and resulting microstructures in wet rock analogues that are loaded below the plastic limit within the pressure solution deformation regime. NaClO₃ has a solubility and dissolution/precipitation rate comparable to that of NaCl, but it is an elastic/brittle salt. At room temperature and atmospheric pressure dry single crystals of this material fail catastrophically at differential stresses in the range 15-25 MPa. Below this stress it only deforms elastically (strain resolution ~0.05%). Cylindrical samples (5 mm in diameter and 10 mm long) of polycrystalline NaClO₃ were loaded axially up to a differential stress of 2 MPa. The samples have a grain size in the range 180-212 µm and a porosity of 1-2%. There is about 1-2 Vol% saturated NaClO₃-solution present at grain boundaries and in the pores. Samples were sealed-off with latex jackets and were deformed undrained. The mechanical data show rate-dependent steady-state creep at a strain rate of ~10^-7/s up to finite bulk strains of ~5%. The data are in good agreement with mechanical data obtained in compaction experiments on porous NaClO₃ aggregates where pressure solution was inferred to be the dominant compaction mechanism. At axial strains higher than about 6%, samples drastically weaken and strain rates increase up to ~2.10^-6/s within 6 hrs (and strains of about 18%). The optical deformation microstructures indicate that this weakening effect is associated with a transition from pressure solution creep to bulk (i.e., non-localised) cataclasic creep. Original grains show large numbers of healed micro fractures, mostly oriented subparallel to the {100} crystallographic planes. Our poster documents the mechanical data and resulting deformation microstructures.

Session L01:4A

3408

L01 : 4A/01 : F4

Centrifuge-Model Investigation of Spatial Variation of Bedding-Plane Strain, Partitioning of Shortening, and Displacement Transfer in Fold-Thrust Structures

Evsen O. Aydemir (Evsen.O.Aydemir@can.conoco.com)¹ &

John M. Dixon (dixonj@post.queensu.ca)²

¹ Conoco Canada Ltd.,, 205-5th Ave SW, Calgary, Alberta T2P 2V7, Canada

² Experimental Tectonics Laboratory, Dept. of Geological Sciences, Queen's University, Kingston, Ontario K7L 3N6, Canada

We examine strain partitioning and displacement transfer in fold-thrust structures using physical analogue models constructed of interlayered Plasticine (competent units) and silicone putty (incompetent units) and deformed at 2500-4000 g in a centrifuge. Shortening occurs by buckling, thrusting (primarily on fore-limb thrusts but also on smaller back-thrusts) and distributed ductile strain.

Spatial variation of 2-d strain within the plane of bedding of a competent stratigraphic unit is monitored using a new technique. Two sets of markers are embedded on what will be the horizontal mid-plane of the unit during its assembly from thin laminae of contrasting colour. First, narrow Plasticine marker strips of contrasting colour are aligned at ~2 mm spacing, parallel to strike and normal to the shortening direction. Second, galena (PbS) powder is used to mark a rectangular array (~2x3 mm) of points on the marker strips. In the assembled model, the galena markers are situated at a known height and can be located in plan view by X-ray. After deformation of the model, the galena marker points can again be located in plan view by X-ray, and their heights interpolated from the positions of the respective coloured marker strips as seen in serial transverse sections cut through the model.

To compute the variation of two-dimensional strain in the plane of the competent stratigraphic unit, computer representations of the marker grid before and after deformation are constructed using GOCAD 3-d visualization software. The shapes of corresponding triangular elements (each defined by three adjacent marker points) in the two grids are compared to calculate 2-d strain associated with the center point of each element. GOCAD is used to create colour-contoured maps of strain-intensity variation in the marker surface.

The models show that transverse shortening is constant along strike, but there are systematic covariations in the proportions of shortening accommodated by folding, thrusting and distributed strain. An individual fold-thrust pair can change along strike from a tight fold with overturned forelimb riding on a small forethrust to an open fold with upright forelimb riding on a large forethrust. Distributed bedding-plane strain also varies with fold tightness: it is most intense in a steep but upright forelimb but is lower where the forelimb is overturned, possibly because of down-dip stretching in the overturned limb. In a fold backlimb, bedding-plane strain intensity is lower in the vicinity of a backthrust because the backthrust takes up a higher proportion of the local shortening. Bedding-plane strain intensity can be used to predict variations in orientation and intensity of fractures in prototype structures.

0937

L01 : 4A/02 : F4

Modelling Thrust Wedges with Anisotropic Stratigraphy

Antoni Teixell (geotec@geologia.uab.es)¹ &

Hemin Koyi²

¹ Dept. Geologia, Universitat Autonoma de Barcelona, 08193 Bellatera, Spain

² Hans Ramberg Tectonic Laboratory, Department of Earth Sciences, Uppsala, Sweden

With the exception of the decollement zone, thrust systems are usually modelled using homogeneous stratigraphies, both in kinematic and physical modelling. However, natural thrust ramps often cross layers of contrasting rheological properties, which cannot behave as passive markers during deformation. We conducted field and sand-box model studies to investigate how stratigraphic variations influence the styles of deformation associated with thrusting at the scale of individual layers and the entire wedge. A particularly instructive natural example of the Spanish Pyrenees (the Aragüés thrust system) was investigated in detail, and model analogues were prepared with alternating layers of granular materials exhibiting different mechanical properties (sand, glass beads and combinations of the two).

Field and model results show that deformation is decoupled between mechanically stronger layers (quartz sandstone at Aragüés, pure sand in models) and weaker ones (limestones and turbiditic rocks at Aragüés, glass beads and sand/glass beads combinations). In both model and nature, this decoupling is displayed by change in orientation and splay of thrust surfaces as they propagate from a stiffer to a weaker unit. Clean-cut single thrusts pass to more distributed zones of deformation (including ramification into closely spaced imbricates). Also noticeable are changes in fault-related folding styles, i.e. transitions from fault-bend to fault-propagation folding.

Modelling also reveals that the effect of stratigraphic variations is also dependent on the basal friction at the decollement. Low-friction experiments enhance lithological contrasts, whereas high-friction experiments induce internal deformation to be distributed throughout the succession. In models with weak surface layers, efficient slope regularization processes by slumping permit close stacking of underlying stiffer units, where the slope of the fold envelope of the imbricate system and the surface slope are dissimilar.

0179

L01 : 4A/03 : F4

4-D Analysis of Analogue Models: Examples of Transfer Zones in Thrust Belts

Guido Schreurs (schreurs@geo.unibe.ch),

Reto Hänni (haenni@geo.unibe.ch) &

Bernard Dräyer

Geological Institute, Baltzerstrasse 1, CH-3012 Bern, Switzerland

About a decade ago the non-destructive X-ray computerized tomography (CT) technique was introduced in geology to analyze analogue models (Mandl, 1988; Colletta et al., 1991). On the basis of attenuation of X-rays by materials, early X-ray CT scanners produced cross-sectional images through models. Because of the time involved in data acquisition and the required cooling of the X-ray source, these X-ray CT scanners could only compute a 3-D image on the basis of sequential cross-sectional slices at the end of an experiment. Technical developments in recent years have now resulted in more powerful X-ray CT techniques. We performed analogue model experiments within the investigation field of a spiral CT scanner. This new-generation scanner revolves around the model and allows a 4-D analysis of the deforming model by generating time-lapse 3-D volumetric images. Time needed for 3-D data acquisition depends on analogue material, X-ray dose intensity, slice spacing and size of investigated area. Analogue modeling is of particular interest when working in complex structural regimes (where lateral changes in 3-D geometry are common) because 3-D imaging of models can provide constraints for geological interpretations and seismic analysis of complex zones. Moreover, knowledge of structural evolution through time is important to determine the potential for migration and trapping of hydrocarbons.

Our experiments simulated structures in transfer zones produced by shortening of multilayer models using brittle and ductile analogue materials. The effects of lateral changes in analogue material on the resulting structures were investigated. Computer visualisation techniques were used to create animations on the basis of CT images and to analyze the structural evolution of our models in 3-D. Visualisation shows that lateral changes in analogue material result in the development of transfer zones. In such zones shallow dipping lateral or oblique ramps connect further advanced thrust fronts with less far-travelled thrust fronts.

Colletta, B, Letouzey, J, Pinedo, R, Ballard, JF, Bale, P, Geology, 19, 1063-1067, (1991).

Mandl, G, Mechanics of Tectonic Faulting: Models and Basic Concepts, (1988).

0814

L01 : 4A/04 : F4

Erosive Mass Transfer at Convergent Margins: Geological Evidence, Analog Models, Mass Transfer Modes

Nina Kukowski (nkukowski@geomar.de)¹ &

Juergen Adam (adam@geowiss.uni-hamburg.de)²

¹ GEOMAR, Wischhofstr. 1-3, D-24148 Kiel, Germany

² Department of Geology and Palaeontology, University of Hamburg, Bundesstr. 55, D-20146 Hamburg, Germany

Non-accreting margins comprise more than half of the length of active convergent margins. However, mass transfer modes at erosive margins are not well understood. Seismic profiles across numerous margins give evidence for short term frontal tectonic erosion affecting accretionary complexes at the toe of the overriding continental fore arc crust which may be caused by subducting asperities as seamounts or aseismic ridges and the rough surface of the subducting oceanic crust. Geological data infer long term frontal and/or basal erosion processes at non-accreting continental margins with significant trench retreat rates. Because this long term tectonic erosion occurs at the base of the forearc at greater depths, seismic imaging of related features is difficult. Thus, for a better understanding of erosive mass transfer at convergent margins it is necessary to combine the geological and geophysical data with mechanical concepts and analog models.

Scaled sandbox models with excess output (with a subduction window) provide insight in possible mass transfer modes at erosive margins. In the low basal friction case, basal erosion occurs in the footwall of a wedge-scaled out-of-sequence-thrust initiated on top of the subduction window located at the inner base of the deformable backstop. Trenchward overthrusting of the non erosive hangingwall block causes a pronounced fore-arc high with a leading depression. Tectonically eroded material will be removed from the arcward part of the accretionary wedge and from the adjacant backstop segment, both situated in the footwall block of the out-of-sequence-thrust, as shown by a significant decrease of the wedge geometry. This modeled deformation pattern images the formation of a fore-arc basin trapped between inner and outer fore-arc highs. Contemporaneously, a frontal accretionary wedge may remain to exist or even grow, if there is some material supply indicating the simultaneity of frontal accretion and basal erosion.

In a high basal friction mode, nowhere significant growth is observed in the entire wedge. In sandbox models with heterogenous input and where a new basal decollement develops within the incoming sediment pile, basal underplating and frontal erosion occur simultaneously. Basal underplating also causes significant uplift and surficial extension of the continental margin. The combination of high basal friction and frontal erosion leads to an increasing taper and remarkable slope break of the frontal wedge. This characteristic topographic feature also is confirmed by bathymetric data of inner trench slopes at various erosiveconvergent margins. The results of the experiments imply that retreating margins may grow vertically to obtain a stable high frictional wedge geometry and therefore offer new concepts for the subduction of both, frontal and mid-arc material consisting of previously accreted sediments or frame work rock of the fore-arc crust.

1130

L01 : 4A/05 : F4

Erosive Mass Transfer at Convergent Margins: Constraints and Application of Coulomb Wedge Analysis to Fore-Arc Systems

Jürgen Adam (adam@geowiss.uni-hamburg.de)¹ &

Nina Kukowski (nkukowsk@geomar.de)²

¹ Department of Geology, University of Hamburg, Bundesstrasse 55, D-20146 Hamburg, Germany

² GEOMAR Research Centre for Marine Geosciences, Wischhofstrasse 1-3, D-24148 Kiel, Germany

Actually, frictional wedge models concerning Coulomb plastic material provide the most actual physical concept for upper crustal deformation in convergent settings and allow to explore the dynamics, kinematics, mass transfer patterns and the controlling mechanism of exogenetic processes. In practise the capability of frictional wedge modelling is confirmed by numerous dynamic analyses of accretionary complexes and foreland thrust belts. For the study of the dynamics concerning long term erosive mass transfer at non-accretive convergent margins we apply scaled analog sandbox models and an enlarged frictional wedge concept to the entire brittle fore-arc crust.

In terms of Coulomb wedge dynamics, frontal and basal erosion occur only under particular conditions. Frontal erosion will be triggered by the modification of the wedge shape due to variation of the wedge base by subducting topographic asperities of the oceanic crust shifting wedge dynamics into overcritical and subcritical states (e.g. Japan trench). Long term orogenic processes like mass transfer by basal erosion and underplating at non-accretive convergent margins is not restricted to the toe of the overriding plate but modify the entire fore-arc crust (e.g. Northern Chile). Here enlarged frictional wedge analysis for the brittle fore-arc crust describes the active state of stress, mass transfer patterns, rheological conditions and dynamics controlling tectonic erosion. Progressive subduction processes and associated mass transfer shape the wedge geometry of the fore-arc system adjusting a dynamic equilibrium between internal deformation, basal erosion and surficial mass transfer.

Therefore, at erosional convergent margins without significant recent variations of geodynamic or exogenetic factors, the fore-arc wedge geometry reflects the active dynamic processes and allows to analyse the rheological properties and dynamic processes of tectonic erosion. Ongoing basal erosion only occurs in high frictional Coulomb wedges at the verge of existence limit. In this case wedge internal detachments develop parallel to the locked subduction interface and material of the wedge base will be transported away by the subducting plate within a high strain deformation or melange zone. The process is comparable with the subduction channel model for sediment subduction. Changing rheological conditions with increasing depth (e.g. strain hardening, dewatering processes) control the unlocking of the subduction fault and re-entering of eroded material into the fore-arc wedge by underplating.

The different modes of mass transfer by tectonic erosion, subduction and underplating are controlled by variations of the dynamic state within the fore-arc wedge system and are reflected by the actual wedge geometry and active deformation processes. The resulting concepts for erosive mass transfer are supported by analog models and observations at active non-accretive margins.

Hans Ramberg Symposium ­ Analogue Modelling of Tectonics

Hans Ramberg Symposium ­ Analogue Modelling of Tectonics

Hans Ramberg Symposium ­ Analogue Modelling of Tectonics

Hans Ramberg Symposium ­ Analogue Modelling of Tectonics

Hans Ramberg Symposium ­ Analogue Modelling of Tectonics

Hans Ramberg Symposium ­ Analogue Modelling of Tectonics

Hans Ramberg Symposium ­ Analogue Modelling of Tectonics

Hans Ramberg Symposium ­ Analogue Modelling of Tectonics

Hans Ramberg Symposium ­ Analogue Modelling of Tectonics