Wood-Anderson torsion seismograms are synthesized by convolution with cosited broadband digital seismograph recordings and compared with the photographically recorded Wood-Anderson seismograms at Berkeley and Mt. Hamilton. The objective is to use routinely the synthesized seismograms to determine local magnitude. Comparison of local magnitude values calculated from both synthesized and photographic seismograms shows that the variation in local magnitude between the two types of seismograms is less than the variation among the photographic seismograms. Consequently, broadband digital recordings can provide reliable local magnitude estimates without introducing significant bias.
It is shown that the static magnification of standard (Ts = 0.8 sec) Wood-Anderson torsion seismographs, determined from measurement of the free period and tilt sensitivity, is 2080 ± 60, and not 2800 as often reported. This discrepancy is due to the erroneous assumption that the taught-wire suspension in a Wood-Anderson seismograph does not distort significantly, thus leading to an incorrect estimate of the radius of gyration and consequently to a theoretical geometrical magnification of 2800.