Earthquake frequency distribution and the mechanics of faulting

The level of intraplate seismicity in Japan generally shows a positive correlation with the density of Quaternary faulting. In southwest Japan, where intraplate seismicity is concentrated on land, rates of seismic moment release ( ) are similar when calculated from either the 400‐year historical record of seismicity or geologically determined slip rates of Quaternary faults. A data set of 18 earthquakes with seismic moments (M 0 ) ranging from ∼0.01 to 3×10 27 dyn cm shows a relationship between rupture length l and M 0 (log M 0 = 23.5+1.94 · log l ). When seismic moment on each Quaternary fault is assumed to occur in discrete events every T = M 0 / years (where M 0 is estimated for a rupture extended over the entire fault length, and is proportional to the slip rate of each Quaternary fault), the moment frequency distribution of earthquakes (log N =A − B · log M 0 ) predicted from the geologic record is virtually identical to that seen with the 400‐year record of seismicity. In contrast, if it is assumed that earthquakes on each fault occur according to the Gutenberg‐Richter relation, we obtain poor agreement with the observed seismicity. Thus, while regional seismicity satisfies the relation log N =A − B · log M 0 (or equivalently, log N = a − b · log M, where M is magnitude), it appears that seismicity on individual faults does not. This further implies that the primary factor that leads to the magnitude frequency distribution in regional seismicity studies is the relative distribution of the slip rates and lengths of preexisting faults.