THERE IS NO SUCH THING AS A FREE FAULT

Fault systems may evolve to minimize the total work on the system. Numerical models permit direct evaluation of individual terms within the work budget including frictional heating, internal work, tectonic work along the model boundaries, work against gravity and seismic work. More enigmatic than these is the work required to propagate faults. While this term may be smaller than other terms in the work budget, observations of long-lived faults suggest that Wprop is not negligible. Numerical simulations of triaxial experiments produce good match to observation when the work of creating new fault surfaces is considered. Numerical simulations of sandbox experiments show that new forethrusts lower the total work but only when the total work gain exceeds the cost of the new fault. The drop in work with new fault formation can be used to estimate the cost of fault growth in the sandbox.