Nowadays, technological advances in satellite imagery measurements as well as the development of dense geodetic and seismologic networks allow for a detailed analysis of surface deformation associated with active fault seismic cycle. However, the study of earthquake dynamics faces several limiting factors related to the difficulty to access the deep source of earthquake and to integrate the characteristic time scales of deformation processes that extend from seconds to thousands of years. To overcome part of these limitations and better constrain the role and couplings between kinematic and mechanical parameters, we have developed a new experimental approach allowing for the simulation of strike-slip fault earthquakes and analyze in detail hundreds of successive seismic cycle. Model rheology is made of multilayered visco-elasto-plastic analog materials to account for the mechanical behavior of the upper and lower crust and to allow simulating brittle/ductile coupling, postseismic deformation phase and far-field stress transfers. The kinematic evolution of the model surface is monitored using an optical system, based on subpixel spectral correlation of high-resolution digital images. First, results show that the model succeed in reproducing the deformation mechanisms and surface kinematics associated to the main phases of the seismic cycle indicating that model scaling is satisfactory. These results are comforted by using numerical algorithms to study the strain and stress distribution at the surface and at depth, along the fault plane. Our analog modeling approach appears, then, as an efficient complementary approach to investigate earthquake dynamics.
Abstract Earthquakes result from fast slip that occurs along a fault surface. Interestingly, numerous dense geodetic observations over the last two decades indicate that such dynamic slip may start by a gradual unlocking of the fault surface and related progressive slip acceleration. This first slow stage is of great interest, because it could define an early indicator of a devastating earthquake. However, not all slow slip turns into fast slip, and sometimes it may simply stop. In this study, we use a numerical model based on the discrete element method to simulate crustal strike‐slip faults of 50 km length that generate a wide variety of slip‐modes, from stable‐slip, to slow earthquakes, to fast earthquakes, all of which show similar characteristics to natural cases. The main goal of this work is to understand the conditions that allow slow events to turn into earthquakes, in contrast to those that cause slow earthquakes to stop. Our results suggest that fault surface geometry and related dilation/contraction patterns along strike play a key role. Slow earthquakes that initiate in large dilated regions bounded by neutral or low contracted domains, might turn into earthquakes. Slow events occurring in regions dominated by closely spaced, alternating, small magnitude dilational and contractional zones tend not to accelerate and may simply stop as isolated slow earthquakes.
Abstract We have developed a scaled analog model of a subduction zone simulating seismic cycle deformation phases. Its rheology is based on multilayered visco‐elasto‐plastic materials to account for the mechanical behavior of a continental lithospheric plate overriding a subducting oceanic plate. The seismogenic zone displays unstable slip behavior, extending at depth into a weak interface with stable slip properties. The model succeeds in reproducing interseismic phases interrupted by coseismic ruptures and followed by after‐slip. The experimental data catalog shows a broad variability of slip events from aseismic slow slips to fast dynamic lab quakes. Results also show the occurrence of both isolated and precursory slow‐slip events arising before the mainshocks. Given the absence of fluids in the model, the broad variability in slip event velocity can be attributed to fault roughness complexity. The model rheology induces also a key visco‐elastic coupling between the elastic overriding plate and the mantle wedge allowing, for the first time, to reproduce experimentally a realistic postseismic visco‐elastic relaxation phase. Preliminary results reveal that the tectonic loading rate modulates this visco‐elastic coupling. A low loading rate weakens it, which increase the amount of storable interseismic elastic deformation, and favors the occurrence of large megathrust events. A high loading rate strengthens it, which minimize the accumulation of interseismic elastic deformation, the slip‐event sizes, and promote aseismic creep. This new scaled‐analog subduction model is a complementary tool to investigate earthquake mechanics and improve the interpretation of geodetic and seismological records.
Le cycle sismique s’etend de la centaine a quelques milliers d’annees mais les mesures geodesiques et sismologiques s’etendent sur moins d’un siecle. Cette courte echelle de temps d’observation rend difficile la mise en evidence du role des parametres sismotectoniques clefs qui controlent la dynamique des failles actives. Pour palier a ce probleme d’echelle temporelle, j’ai developpe un nouveau modele experimental qui reproduit des microseismes le long d’une faille decrochante sur plusieurs centaines de cycles sismiques. Il est constitue de deux plaques de polyurethane lateralement en contact, reposant sur une couche basale de silicone, simulant le comportement mecanique d’une croute superieure elastoplastique couplee avec une croute inferieure ductile, respectivement. Pour chaque experience, environ 4000 mesures du champ de vitesses horizontales sont enregistrees. L’analyse des deplacements de surface au cours des phases intersismiques, cosismiques et postsismiques et leur comparaison aux failles sismogeniques montrent que le modele reproduit correctement les deformations proches de la faille et en champ lointain. J’ai aussi effectue des inversions du champ de vitesses en surface pour evaluer la distribution spatiale du glissement en profondeur le long du plan de faille. Pour comparer les experiences, j’ai developpe plusieurs algorithmes permettant d’etudier l’evolution spatio-temporelle des principaux parametres physiques et les processus de deformation de surface qui caracterisent le cycle sismique.
Mes premiers resultats suggerent que la vitesse de chargement tectonique imposee en champ lointain joue un role sur le cycle sismique en influencant la magnitude des seismes, leur temps de recurrence, ainsi que la capacite de la faille a generer des seismes caracteristiques. Une vitesse de chargement lente favorise l’occurrence de forts evenements caracteristiques et une vitesse rapide de nombreux microseismes de magnitude faible a moderee plus distribues le long de la faille. Ma premiere hypothese est que ce comportement est controle par le couplage fragile/ductile a la base des plaques de polyurethane. Pour une vitesse rapide, les forces visqueuses dans la couche basale augmentent de meme que ce couplage. Ce processus contraint la base de la faille a glisser a une vitesse proche de sa vitesse long-terme et induit un champ de contrainte plus heterogene le long de son plan qui favorise les microseismes de magnitude faible a moderee. Pour une vitesse lente, le silicone se comporte comme un fluide newtonien et les forces visqueuses diminuent considerablement, permettant a la faille de rester bloquee sur une plus longue periode et d’accumuler plus de deformation elastique. Les contraintes sont ensuite relaxees par de plus larges evenements sismiques.
Enfin, j’ai etudie le role joue par les variations de contrainte normale le long de la faille sur le glissement cosismique et le comportement long terme du systeme. Les resultats montrent que la distribution spatiale du glissement cosismique est fortement controlee par les variations de resistance de la faille et de l’accumulation des contraintes cisaillantes qui en resultent. Les evenements majeurs se produisent preferentiellement dans les zones d’asperite de contrainte cisaillante et leur distribution spatiale du glissement suit une tendance similaire a celle de la variation de contrainte normale le long de la faille. L’analyse revele aussi que l’heterogeneite de l’etat de contrainte initial influence la regularite du cycle sismique et le comportement long terme du modele. Les resultats de cette etude parametrique conforte ainsi l’hypothese selon laquelle la distribution du glissement cosismique le long des ruptures peut fournir des informations pertinentes sur l’etat de contrainte initial et pourrait ameliorer notre comprehension de l’alea sismique. Notre approche experimentale apparait donc, comme une methode complementaire et efficace pour etudier la dynamique des tremblements de terre.
<p>Fault damage zones present a renewal of interest to better understand stress perturbations around faults, earthquake&#8217;s ground motions and fluid flow in the upper crust. Although numerous studies provide significant amounts of data from a broad variety of rocks, the processes controlling fault damage development are not clearly understood and scaling properties in carbonate rocks remain poorly studied. D-T (displacement - DZ thickness) data compilations show strong scattering and are acquired using different methods and at different places along the faults (including tip, wall, link, or inner and outer damages), therefore rendering difficult a proper definition of the scaling relationship.</p><p>&#160;</p><p>First, we analyse fault/fracture systems at the outcrop and map scale and define displacement - thickness (D-T) scaling of fault damage zones using scanlines, in carbonate rocks in France and Spain. We determine fault displacement and damage zone thickness perpendicular to fault planes and far from fault tips for 12 selected faults in four study sites. The data show a logarithmic decrease of fracture frequency from the fault cores. This decrease is characterized by local frequency peaks corresponding to variably-linked secondary fault segments and abandoned tips within the fault damage zone. D-T data comprised between 0 and 100 m of net fault displacement show a nearly linear scaling with very little scattering. Including two additional data for D > 100 m, the best fit corresponds better to a power law. The linear scaling is explained by well-known processes of fault growth such as stress perturbations around faults and fault segment linkage. The non-linear trend shown by the largest faults suggests that at this scale the faults become restricted at their lower tips by the base of the brittle crust.</p><p>&#160;</p><p>Secondly, we analyse fault damage growth using analogue modelling of normal faults in a sand box. The model is composed by a 5 cm thick layer of dry sand deposited above a 2 cm thick ductile &#8220;kinetic sand&#8221; (sand and silicone) layer. The experiment is analysed in cross-section using image correlation allowing to calculate the velocity field and strain tensor over the fault zones including their damage pattern. Fault damage thickness obtained using the strain field appears to grow linearly with respect to shear displacement when the fault is contained into the dry sand layer. When the fault lower tip reaches the kinetic sand, fault damage begins to growth non-linearly with shear displacement, revealing that the brittle layer thickness is the main parameter governing the non-linear scaling.</p><p>&#160;</p><p>&#160;</p>
Abstract Fault damage zones strongly influence fluid flow and seismogenic behavior of faults and are thought to scale linearly with fault displacement until reaching a threshold thickness. Using analog modeling with different frictional layer thicknesses, we investigate damage zone dynamic evolution during normal fault growth. We show that experimental damage zone growth with displacement is not linear but progressively tends toward a threshold thickness, being larger in the thicker models. This threshold thickness increases significantly at fault segment relay zones. As the thickness threshold is approached, the failure mode progressively transitions from dilational shear to isochoric shear. This process affects the whole layer thickness and develops as a consequence of fault segment linkage as inferred in nature when the fault matures. These findings suggest that fault damage zone widths are limited both by different scales of mechanical unit thickness and the evolution of failure modes, ultimately controlled in nature by lithology and deformation conditions.
