Vein spacing in extending, layered rock: The effect of synkinematic cementation

Cemented fractures (veins) commonly show mm-scale spacing, even in layerbound arrangements in beds of cm- to dm-scale thickness. The relief of tension around such layerbound fractures should preclude nearby fracture propagation and result in cm- to dm-scale fracture spacing. We hypothesize that cement precipitated during vein opening could re-establish tension across veins and lessen the effects of relieved tension, thus decreasing fracture spacing. We test this hypothesis using a computer-based numerical model. The model consists of a 2D triangular lattice of nodes connected by elastic springs. The lattice is stretched by holding the left-boundary stationary and moving the right-boundary to the right at constant velocity. The lattice consists of three layers; springs within the middle fracturing layer fail upon stretching past a given critical length. Springs within the upper and lower matrix layers are indestructible. We tune the model parameters to produce the familiar regular spacing of barren fractures (joints). Then we perturb this system by adding cement within fractures as the fractures propagate and widen. Cementation is simulated by extending failed springs across the space between the nodes on which they are rooted and re-attaching springs once they reach across. Re-attached springs are assigned a new neutral length equal to their current length on re-attachment. The primary effect of cementation is to make fractures narrower and more closely spaced. Thus the model highlights the resistance to fracture widening by cement as a potential reason why veins can be more closely spaced than joints. Individual fractures open and seal multiple times when the stiffness of cemented springs is lower than that of the host-rock springs, suggesting that natural crack-seal vein opening is associated with persistent mechanical weakness at extant fractures. Modeled veins are more irregularly spaced than modeled joints, but the vein patterns remain more regularly spaced than would be expected for a random arrangement. Therefore the model does not explain systematic clustering of natural veins