Elena I. Novikova

The Modified Mercalli intensity and the geometry of the sedimentary basin as scaling parameters of the frequency dependent duration of strong ground motion

Abstract Several new empirical equations of the frequency dependent duration of strong earthquake ground motion are presented. The duration is considered as being composed of two parts: (1) the duration of stong motion as it is observed at recording stations located on basement rocks, and (2) the prolongation of this duration for stations located on sediments. The first part, called the ‘basic duration’, is modelled in terms of the Modified Mercalli intensity and (in some cases) the hypocentral distance. The depth of the sediments under the station, the distance from the station to the rocks surrounding it, and the angular measure of the size of those rocks (as seen from the station) are chosen as the parameters for modelling the prolongation of the duration. The new empirical equations are compared (a) with each other, (b) with our previous models which used similar ‘prolongation’ terms, but the ‘basic duration’ was expressed in terms of the magnitude of the earthquake and the source-to-station distance, and (c) with models with ‘intensity-type’ ‘basic duration’, but with a simplified ‘prolongation’ term (the geological conditions at the stations are modeled by lumping all the sites into three groups: basement rock, sediments and intermediate geology). This collection of models is found to have good internal consistency. The choice of the proper model depends on the availability of the earthquake and site parameters. The residuals of the empirical regression equations are found to have similar distribution functions for all the models. An explicit functional form for such distributions is proposed, and the frequency dependent coefficients are found for all the models of duration. This allows one to predict (for each set of earthquake and site parameters) the probability of exceedance of any given level of duration of strong ground motion at a given frequency.

Digital Instrument Response Correction for the Force‐Balance Accelerometer

Correction for the instrument response of acceleration records obtained by force balance accelerometers (FBA) is necessary (a) to eliminate phase distortion at high frequencies, and (b) to broaden the useful frequency band up to, and beyond, the corner frequency of the system. The proposed algorithm contains operations in the time domain only and can be applied to any digitized record obtained from a FBA with known characteristics. The adequacy of the procedure depends on the accuracy of the information on the transducers constants: damping ratio and natural frequency. An appropriate testing procedure is also presented.

Modified Mercalli intensity scaling of the frequency dependent duration of strong ground motion

Abstract New empirical models of the duration of strong ground motion in terms of the Modified Mercalli intensity at the recording station are presented. Two groups of regression equations are considered: one explicitly includes the dependence of the duration on the distance to the source, and the other excludes this dependence. The Modified Mercalli intensity serves as a parameter in both types of models. The models of the first type are more descriptive, but are also more region dependent, because the regional dispersion and attenuation laws are ‘built into’ the frequency dependent regression coefficients. For a given site intensity, the duration grows when the distance from the source to the recording site increases. For a given distance from the source, the dependence of the duration on the site intensity is more complex. At low frequencies, the duration of strong motion decreases when the intensity increases, while at high frequency it grows with increasing intensity. A smooth transition from one type of dependence to another occurs at intermediate frequencies. When compared to basement rock sites, the duration of strong motion at sedimentary sites is prolonged by about 5 s at frequencies near 1 Hz. The prolongation of the duration on the soft soils can be as much as 7 s. The influence of the type of soils on the duration is stronger at higher frequencies ( f = 0.3−25 Hz), while the effect of the presence of sedimentary deposits can be observed at lower frequencies ( f = 0.15−2 Hz). The residuals of the empirical regression equations were also studied, and their distribution function is proposed.

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 (θ 0 ). 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 θ 0 (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 θ 0 (usually neglected), is mostly between ± 8°, but may reach ±12°. Assuming θ 0 = 0 leads to systematically smaller values of the measured sensitivity (e.g. by ∼3% for θ 0 = 8° and ∼4% for θ 0 = 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.