Experimental testing of processes controlling the structuring of major continental rifts

Assessing the seismic hazard in continental strike-slip zones (faults that are usually vertical and along which two blocks slide horizontally in relation to each other), such as the San Andreas Fault in the USA or the Levant Fault in the Near East, involves integrating past and present seismic activity, tectonic loading rates and, critically, fault geometry into a single model. The combination of these different observations enables us to compile synthetic catalogs listing a sufficiently large number of earthquakes to calculate the probability of a specific rupture scenario occurring.

Publication date: 17/09/2020

Press, Research

Related themes : Earth System Science, Natural Hazards

In this context, apart from the potential uncertainties associated with the measurements, the assessments of seismic activity and the rate of tectonic loading are difficult to discuss. Conversely, the geometry of strike-slip faults, and more specifically their structuring along the fault, is far from being the subject of consensus. So, depending on the approach chosen, the results vary drastically: the models consider either continuous faults or faults that are more or less segmented, with rules governing the propagation of a seismic rupture from one segment to another.

A team of researchers from IPGP, CY Cergy Paris Université and Sorbone Université have attempted, upstream of hazard modelling, to highlight a systematic segmentation of faults, to characterise the relevant spatial scale to be considered and to identify the physical parameter(s) controlling this segmentation. To do this, the scientists carried out analogue basal shear experiments in sand boxes. One of the properties of sand is that it localises deformation and behaves like an analogue of a brittle material, even though it does not store elastic energy. The researchers therefore took photographs of the surface of all the experiments at regular time intervals. Using the MicMac image correlation code, they were able to measure surface deformations precisely, and in particular the appearance of segmentation.

A/ Typical surface rupture map for a continental earthquake (1992 Landers earthquake in California, Mw=7.2). The rupture is not rectilinear, but breaks up into several distinct segments. B/ Growth diagram of a strike-slip fault initiated by a series of shear-oblique Riedel cracks at the base of the brittle zone of the continental crust (~15 to 20 km). These cracks have a helical geometry. In a 2nd phase, deformation coalesces along a shear zone that intersects these Riedel cracks. C/ Map view of a sandbox experiment with the trace of Riedel slits intersected by shear. However, these cracks continue to control the lateral geometry of the shear.

Shear experiment. The surface deformation of the material, in this case the rotation of the sand grains, is measured by correlating successive optical images acquired during the experiment. The 3 snapshots show the successive stages of deformation localization: 1) Riedel splits, oblique to the main shear direction and disconnected from each other. 2) As the cumulative displacement (indicated top left) increases, the shear localizes and a fault zone appears, connecting the Riedel slots and accommodating an increasing proportion of the deformation at the expense of these slots, which are gradually abandoned. 3) The shear zone is completely localized in the center of the experiment. The geometry of this zone is not perfectly regular, however, as geometric asperities that laterally structure the fault zone remain visible where the Riedel cracks were initially located. This experiment shows how the initial emplacement of the shear zone leaves a long-term imprint that structures the geometry of the fault zone by defining a regular segmentation along the shear. The diagram summarizes this observation in 3D.

The main parameters studied that were likely to have an impact on the different experiments were the internal friction of the material (the type of sand and the way it is distributed in the box), the basal friction (the nature of the contact between the sand and the bottom of the box) and the thickness of the sand. This last parameter proved decisive in controlling the size of the segments, with a constant ratio, close to 1, between the thickness of the sand and the length of the segments. The influence of the other parameters is only secondary and only changes this result marginally.

By comparing these data with previously documented measurements of the segmentation of surface ruptures in large unstripping earthquakes, the scientific team was able to show that the ratio between segment length and thickness of the seismogenic crust is the same as that observed experimentally between segment length and sand thickness, suggesting that the physical mechanisms involved are the same.

This study therefore reveals the existence of a segmentation of major continental rifts, the spatial dimensions of which are controlled by the thickness of the seismogenic crust, corresponding to the thickness of the sand in the multiple analogue experiments carried out. The thickness of the seismogenic crust varies relatively little from one continental context to another, which explains why the segmentation of faults as observed during major earthquakes always seems to be the same.

This work provides a physical framework for the notion of fault segmentation, which until now has been mainly descriptive, and thus paves the way for better integration of fault geometry in hazard models by pointing out the spatial scales that need to be considered more specifically.

Ref : M. Lefevre, P. Souloumiac, N. Cubas, Y. Klinger; Experimental evidence for crustal control over seismic fault segmentation. (2020) Geology ; 48 (8): 844-848. doi: 10.1130/G47115.1

See all news

Experimental testing of processes controlling the structuring of major continental rifts

Contacts