Advanced sensitivity calibration of the Los Angeles strong motion array

Results are presented of recent sensitivity calibration of 76 accelerographs (SMA-1) of the Los Angeles Strong Motion Array. These have pendulum-like transducers and optical recording system. One characteristic of their design is off-axis sensitivity, which is magnified by transducer misalignment. A new calibration procedure was applied, which considers off-axis sensitivity and measures the angles of misalignment (φ and ψ), as well as the incident angle of the light beam onto the film (θ₀). These are required (1) for accurate estimation of sensitivity, and (2) for proper instrument correction of recorded accelerograms which considers also cross-axis sensitivity and misalignment. These effects are important near large acceleration peaks (approaching and exceeding 1g), e.g. like the ones recorded near the source of the 1994 Northridge earthquake (M_L=6·4). This earthquake was recorded by 65 stations of the Los Angeles Strong Motion Array, at epicentral distances from 2 to 85 km. Histograms showing distribution of the misalignment angles, light beam incidence angle θ₀ (for unloaded position) and the transducer sensitivities are presented. These indicate that the misalignment angles are typically 1–1·5°, but may also be 3–4°. Angle θ₀ (usually neglected), is mostly between ±8°, but may reach ±12°. Assuming θ₀=0 leads to systematically smaller values of the measured sensitivity (e.g. by ∼3% for θ₀=8° and ∼4% for θ₀=12°). Comparison of the newly measured sensitivities with those measured prior to installation (in 1979/1980), s^old, shows that, in general, the new values are systematically smaller. The difference is typically within 5 per cent, but in some cases is as large as 10 per cent. Other principal sources of the observed differences and their mechanisms are discussed. Those include long-term changes in the transducers (e.g. change of stiffness, reflected in changes of the natural frequency) and differences in the calibration procedure (e.g. errors associated with manual reading film records with tilt test data, and with transducer and instrument housing misalignment). The presented results may be considered typical of similar strong motion arrays worldwide. © 1998 John Wiley & Sons, Ltd.