Although it is known that most global earthquakes will concentrate at the plate boundaries, there is no reliable method of accurately predicting the time, place and magnitude of an earthquake. Most current research is concerned with minimising the risk associated with earthquakes, by assessing the combination of seismic hazard and the vulnerability of a given area. Many seismic countries, however, have research programs based on identifying possible precursors to major earthquakes. This includes the study of dilatancy, how rocks crack and expand under the increased stress associated with the earthquake. Some major earthquakes, but not all, are heralded by the occurrence of foreshocks. which can be detected by dense local monitoring networks. Other instruments can measure changes in the levels of radon gas, electrical and magnetic properties, velocity changes of seismic waves and changes in topography. Long term monitoring and examination by these sensors is required as some or all of these factors may change due to the opening of cracks prior to the earthquake.
All attempts to predict earthquakes have, however, been generally considered as failures and it is unlikely that accurate prediction will occur in the near future. Efforts will, instead, be channelled into hazard mitigation. Earthquakes are difficult or impossible to predict because of their inherent random element and their near-chaotic behaviour.