On the existence of earthquake precursors

Earthquake prediction based on precursors can aim to provide fully quantified, time-varying, synoptic forecasts, which do not depart from physical and geological principles, and are amenable to formal testing. These features are in contrast to the traditional occultist or soothsayer style of prediction. The recently-advanced, pre-emptive hypothesis that earthquakes are intrinsically unpredictable, and precursors non-existent, is also amenable to testing: it is refuted by the well-known relations between mainshocks and aftershocks. These relations show that a set of aftershocks is to a high degree predictable from the mainshock, so that, as a matter of principle, the mainshock is a precursor to its aftershocks. This result is compatible with the power-law property of seismicity, on which the unpredictability hypothesis is based. Empirical research on most precursors is difficult because of the scarcity of data, and is still largely at the anecdotal stage. Additional difficulties at the experimental stage are exemplified by the failure of the Tokai and Parkfield experiments to advance the study of precursors as planned. A comparative abundance of data is available on seismicity anomalies, and research on this type of precursor is progressing towards the operational stage.