Sammendrag
Propagation of igneous intrusions, like dikes and sill, is the dominant transport mechanism of magma through the Earth’s lithosphere. Dikes and sills propagate by fracturing or pushing crustal rocks which deform according to a diversity of rheological behaviors, varying from elastic to plastic. Previous studies have shown that mechanical layering of rocks can affect the intrusion morphology, for example by initiating sills, as well as introducing different deformation mechanisms in layers of different strength. Previous layered laboratory models of dike propagation have focused on the behavior of intrusions in a host of the elastic, end-member rheology. Here we expand on these experiments by introducing layering to elastoplastic laboratory experiments. The experiments use granular materials which fail according to a combined Griffith-Mohr-Coulomb failure criterion, making them able to fail in both tension and shear. Two series of quasi-2D experiments were performed to study the effect of cohesive layering on igneous intrusions. Golden syrup (a magma analogue) was injected into granular materials (a crustal analogue) with layers of varying cohesions and depths at a stable rate. The first experimental series (Series A) studied the effect of weak layers in a strong host, while the second experimental series (Series B) looked at strong layers in a weak host. Digital image correlation was used to quantify displacements and shear strain in the experiments. I show that adding layers of different strength have a clear effect on propagating, viscous intrusions. Weak, non-cohesive layers in a strong host have a larger effect than strong, cohesive layers in a weak host. The weak layers are shown to temporarily halt the intrusion propagation, work as detachment zones and control the depth at which the overburden fails by fractures connecting the intrusion to the surface. Furthermore, we show that different emplacement mechanisms can successively occur in layered hosts. Applied to nature, the experiments help to explain field observations of preferential emplacement of sills in weaker rocks. The experiments also highlight the need for dynamical models of dike propagation where both dike generated stresses and interaction with failure of the host rock are taken into account.