Calibration and QC tests of seismic monitoring instrumentation

Accuracy of seismic monitoring relies on the precise calibration and quality control (QC) of recording instruments to ensure reliable recovery of ground motion. Seismic sensors do not measure ground motion directly they record signals that are mixed with the instrument response, which must be characterized and removed to obtain physically meaningful displacement, velocity, or acceleration records. This thesis presents a systematic evaluation of seismic instrument performance through controlled calibration and QC experiments, with particular emphasis on response deconvolution and frequency-domain analysis. A series of laboratory experiments were conducted using a single-axis shaking table to excite four identical Sentinel GEO seismic sensors, each combining a MEMS triaxial accelerometer and an embedded triaxial geophone. Controlled sinusoidal and broadband frequency-sweep excitations were applied to assess sensor sensitivity, linearity, frequency response, and temporal stability. Complementary datasets acquired using a dedicated QC device (IsamGeo) which is analyzed to sensor behavior under varying excitation amplitudes. Data processing and analysis were performed using the Obspy framework, including instrument response removal, detrending, filtering, and computation of frequency spectra, spectrograms, power spectral density (PSD) and root mean square (RMS) metrics. The results demonstrate that the sensors exhibit stable, linear responses within their designed bandwidths, with dominant spectral peaks accurately reproducing the imposed excitation frequencies. Minor sensor-to-sensor variations were observed, reflecting inherent differences in individual transfer functions rather than instrumental malfunction. The study confirms that controlled shake-table testing combined with frequency-domain and time–frequency QC techniques provides a robust steps and tips for seismic instrument calibration. The findings highlight the importance of empirically characterizing instrument responses rather than nominal manufacturer specifications, improving the seismic monitoring data for both natural hazard assessment and structural monitoring applications.