Influence of instruments on the H/V spectral ratios of ambient vibrations

International audienceFor an optimal analysis of the H/V curve, it appears necessary to check the instrument signal to noise ratio in the studied frequency band, to ensure that the signal from the ground noise is well above the internal noise. We assess the reliability and accuracy of various digitizers, sensors and/or digitizer-sensor couples. Although this study is of general interest for any kind of seismological study, we emphasize the influence of equipment on H/V analysis results. To display the impact of the instrumental part on the H/V behavior, some series of tests have been carried out following a step-by-step procedure: first, the digitizers have been tested in the lab (sensitivity, internal noise...), then the three components sensors, still in the lab, and finally the usual user digitizers-sensors couple in lab and outdoors. In general, the digitizer characteristics, verified during this test, correspond well to the manufacturer specifications, however, depending on the digitizer, the quality of the digitized waveform can be very good to very poor, with variation from a channel to another channel (gain, time difference etc.). It appears very clearly that digitizers need a warming up time before the recording to avoid problems in the low-frequency range. Regarding the sensors, we recommend strongly to avoid the use of “classical” accelerometers (i.e., usual force balance technology). The majority of tested seismometers (broadband and short period, even 4.5 Hz) can be used without problems from 0.4 to 25 Hz. In all cases, the instrumentation should be checked first to verify that it works well for the defined study aim, but also to define its limit of use (frequency, sensitivity...)