Surface and basal crevasses are deep fractures in ice that are formed throughout the whole Antarctic ice sheet. They are a direct precursor of formation of rifts in ice shelves and a consequent calving or breaking of icebergs. Modelling of this process is crucial for better estimation of the ice mass balance in Antarctica. Whereas there have been a number of parameterisations that can describe calving in a simple way they do not include all the complexity of calving. A more comprehensive calving model needs to include not only parameterisation of calving at ice front but on the history of ice that lead to a calving event such as formation, propagation and advection of crevasses. Towards improving calving parameterisation in ice sheet models I implement a statistics-based method for formation of surface and basal fractures and their propagation. This method allows finding a better fit to observations and provides a stronger basis for a calving model. Combing a set of created sub-models with advection and spatial propagation of crevasses I develop a preliminary calving model. It is the first calving model that includes the full history of fractures, from their initiation to the final break off the ice front.