. ge o - ph ] 3 0 O ct 2 00 2 Foreshocks and Earthquake Predictability

The observation of foreshocks preceding large earthquakes and the suggestion that foreshocks have specific properties that may be used to distinguish them from other earthquakes have raised the hope that large earthquakes may be predictable. Among proposed anomalous properties are the larger proportion than normal of large versus small foreshocks, the power law acceleration of seismicity rate as a function of time to the main-shock and their spatial migration toward the mainshock, when averaging over many sequences. Using Southern California seismicity, we show that these properties and others arise naturally from the simple model that any earthquake may trigger other earthquakes. This model puts all earthquakes on the same footing. We find that foreshocks precursory properties are independent of the mainshock size. This implies that earthquakes (large or small) are predictable to the same degree as seismicity rate is predictable from past seis-micity by taking into account cascades of triggering. 1 Foreshocks, mainshocks and aftershocks: hypothesis and predictions It has been recognized for a long time that large earthquakes are sometimes preceded by an acceleration of the seismic activity, known as foreshocks [1, 2]. In addition to the increase of the seismicity rate a few hours to months before large earthquakes, other properties of foreshocks have been reported, which suggest their usefulness (when present) as precursory patterns for earthquake prediction. Because foreshocks are rare and it is believed that a good proportion of them are forerunners of large events, specific physical mechanisms have been proposed for them with the hope of helping earthquake prediction [3]-[6]. In addition, anomalous precursory seismic activity extending years to decades before large earthquakes and at distances up to ten times the mainshock rupture size are often thought to require different physical mechanisms [7]-[12] than for foreshocks closer to the mainshock epicenters. The division between foreshocks, mainshocks, and aftershocks has a long and distinguished history in seismology. Within a pre-specified space-time domain, foreshocks are usually defined as earthquakes (above the background rate) preceding a larger earthquake (mainshock), which is itself followed by an increase in seismicity of smaller earthquakes (aftershocks). However, recent empirical and theoretical scrutiny suggests that this division might be arbitrary and physically artificial [13]-[17]. Since the underlying physical processes are not fully understood, the qualifying time and space windows used to select aftershocks, mainshocks and aftershocks are more based on common sense than on hard science. If the space-time window is extended and …