Thrust tectonics are dealt with on the basis of primarily experiments focusing on the dynamics of a developing thrust belt and on understanding and predicting normal-sequence and out-of-sequence thrusting. Field examples are presented in addition to the examples of sandbox-model experiments. The results have improved the insight into thrust-belt-forming mechanisms; the validity of the conclusions is supported by natural examples.The experimental program was aimed at examining the effect of changes in a selection of key parameters in thrust tectonics on the geometry and the successive phases in the development of thrust sheets. Sandbox experiments were used to analyse the effects of basal friction, detachment lithology, basement relief and syntectonic sedimentation. Multilayer experiments were performed to simulate the effects of ductile detachment lithologies.It was found that a thrust belt's cross-sectional geometry is formed in a dynamic process during which the wedge may develop from subcritical through critical to supercritical and back to critical again.The process is illustrated with sandbox experiments, analysed by time-lapse computed tomography scans and in-situ stress measurements. On the basis of the sandboxmodel results and the natural examples, we conclude that a critical examination of the boundary conditions of a fold-andthrust belt and of changes in these conditions during the deformation process enables predictions about the geometry and kinematics of the thrust belt.

Sandbox models of thrust tectonics with multiple detachment levels allow alternative interpretations of fault patterns and evolution, even though they are well controlled experiments. Palinspastic reconstruction resolves these ambiguities and leads to refined interpretations. It reveals that if there are thick viscous interlayers, all non-fault related contraction can be accommodated by folding, without tectonic compaction. It also confirms the complete decoupling of faulting above and below a thick weak viscous layer, and the initial straightness and regular spacing and dips of thrusts above a decollement of low basal friction. A weak elasto-plastic interlayer or decollement merely has a lubricating effect, without complete decoupling. Many natural thrust systems are underlain by a ductile substratum and have more than one decollement level. These examples almost always exhibit a complex thrust belt geometry and are apparently formed by different thrusting events that can be difficult to interpret and to time relative to one another. Weak decollement materials and interlayers have a strong effect on the structural style of thrust belts at the moment of detachment. However, other material properties, in particular the viscosity, play a dominant role during the structural development following detachment (the overthrust phase). The objective of this study was to improve structural interpretation of data from areas of thin-skinned thrust tectonics with weak ductile detachment horizons (i.e. excluding large scale basement involvement). The approach involved: • the use of sandbox experiments to provide controlled structural models with different decollement rheologies, and • the refinement and validation of selected cross-section interpretations with palinspastic reconstruction. The result is a set of better constraints on structural interpretation of cross-sections in thrust tectonics. In natural cases, the most common weak decollement lithologies are over-pressured shales and evaporites, in particular rock salt. In our experiments we have simulated rock salt by using silicone putty (‘SP’), a viscous material that can flow slowly as a Newtonian fluid. We modelled overpressured and/or very soft clay with an oil-water emulsion (‘OWE’), which is a ductile, weak elasto-plastic, non-viscous material in which localized shear zones can form, as in any other elasto-plastic solid.