Linking ground motion measurements and macroseismic observations in France: a case study based on accelerometric and macroseismic databases

Comparison between accelerometric and macroseismic observations is made for three M w = 4.5 earthquakes, which occurred in north-eastern France and south-western Germany in 2003 and 2004. Scalar and spectral instrumental parameters are processed from the accelerometric data recorded by nine accelerometric stations located between 29 and 180 km from the epicentres. Macroseismic data are based on French Internet reports. In addition to the single questionnaire intensity, analysis of the internal correlation between the encoded answers highlights four predominant fields of questions bearing different physical meanings: (1) “vibratory motions of small objects”, (2) “displacement and fall of objects”, (3) “acoustic noise” and (4) “personal feelings”. Best correlations between macroseismic and instrumental observations are obtained when the macroseismic parameters are averaged over 10-km-radius circles around each station. Macroseismic intensities predicted by published peak ground velocity (PGV)–intensity relationships agree with our observed intensities, contrary to those based on peak ground acceleration (PGA). Correlation between the macroseismic and instrumental data for intensities between II and V (EMS-98) is better for PGV than for PGA. Correlation with the response spectra exhibits clear frequency dependence for all macroseismic parameters. Horizontal and vertical components are significantly correlated with the macroseismic parameters between 1 and 10 Hz, a range corresponding to both natural frequencies of most buildings and high energy content in the seismic ground motion. Between 10 and 25 Hz, a clear lack of correlation between macroseismic and instrumental observations exists. It could be due to a combination of the decrease in the energy signal above 10 Hz, a high level of anthropogenic noise and an increase in variability in soil conditions. Above 25 Hz, the correlation coefficients between the acceleration response spectra and the macroseismic parameters are close to the PGA correlation level.