A note on the use of strong-motion data from small magnitude earthquakes for empirical ground motion estimation

A common practice in the derivation of equations for the prediction of strong ground motions is to supplement the still sparse data from large damaging earthquakes with data from smaller earthquakes, which is much more abundant. In this paper, it is shown that even though strong-motion records from small magnitude earthquakes are often affected by recording problems, such as low A/D convertor resolution, they often can yield accurate response spectral accelerations (SAs). The paper then highlights some possible problems with the practice of combining strong-motion data from small and large magnitude earthquakes. These possible problems include: the difficulty of obtaining estimates of the independent parameters (magnitude, distance and faulting mechanism); more rapid decay of ground motions from small earthquakes compared to ground motions from large earthquakes; higher dependence of ground motions on magnitude for small earthquakes compared to large earthquakes; and larger variability in ground motions from small earthquakes compared to large earthquakes.