Spectral corrections for propagation and site effects and their implications for stress drop estimation: application to the 2016 Mw 6.0 Amatrice earthquake

Stress drop is a fundamental parameter controlling the high-frequency content of ground motion, yet its estimation remains highly uncertain and often inconsistent across methodologies. Empirical spectral studies of Central Italy M w 6.0 mainshock yield stress drops of ∼3–4 MPa, while ground motion simulations require ∼10–18 MPa to reproduce observed amplitudes, particularly near source. In this study, we investigate the sensitivity of stress drop estimates for the 2016 Mw 6.0 Amatrice earthquake to the attenuation and site parameters adopted for correcting the observed S-wave spectra. Two sets of attenuation and site parameters were used for the spectral correction, each derived using a different methodology: KS method (κ + Site) and CN method (Coda Normalization). Stress drop is then estimated using the spectral fitting method (SP) applied to the KS- and CN-corrected spectra. The corrections based on the KS parameter set produce stress-drop estimates of approximately 3–4 MPa, whereas the CN-based corrections yield estimates closer to the 10–18 MPa range. The corrected spectra are generally consistent at distant stations but show different distance dependencies and substantial differences in near-source spectral amplitudes at frequencies up to 4 Hz. Using the CN-derived parameters up to 4 Hz, the average stress-drop estimate is approximately 8 MPa, while values of up to approximately 70 MPa are obtained at some near-source stations, including stations located in the forward-rupture direction. These results demonstrate that stress-drop estimates are sensitive to the adopted attenuation and site parameters and to the datasets and station distributions used to derive them. Accounting for these effects is essential for improving the physical consistency of ground motion models and seismic hazard assessments.