Consistency and Uncertainty of Multimethod Source-Parameter Estimates for the 2016–2017 Amatrice–Visso–Norcia Seismic Sequence, Central Italy

ABSTRACT Accurate characterization of earthquake source parameters is essential for understanding rupture processes and developing physics-based ground-motion models. Among these parameters, stress drop (Δσ) inferred from source spectra is particularly relevant, but remains highly uncertain because of trade-offs among path attenuation, site amplification, and methodological assumptions. Here, we evaluate the consistency and epistemic uncertainty of stress-drop estimates for the largest events of the 2016–2017 Amatrice–Visso–Norcia seismic sequence in Central Italy. We compare five independent approaches: spectral fitting, spectral ratios (SRs), nonparametric inversion of source, path, and site terms (generalized inversion technique [GIT]), coda spectral calibration (coda calibration tool), and coda envelope spectral ratios (coda envelope ratio tool). Scaling relations are derived for each technique using method-specific datasets (or published data when available). To enable direct comparison, we also analyze a subset of ∼100 events (Mw∼6.3 to 1.7) that are common to multiple methods, as well as a subset of 22 events (Mw∼6.3 to 3.3) analyzed by all five techniques. The subsets allow us to compare intermethod variability in seismic moment (M0), corner frequency (fc), and stress drop (Δσ). Although the absolute values differ between the techniques, M0–fc trends and the scaling relations for Δσ and apparent stress show consistent slopes and strong correlations. Event-by-event comparisons and Sammon’s maps further indicate similar spectral shapes and source parameters across methods. For example, GIT yields higher stress drops with an average bias of 0.185 in log10Δσ, whereas SR shows the largest negative deviation (−0.203). Across all methods, the intermethod standard deviation is 0.16. For the 22 events that are common to all methods, stress-drop intermethod variability remains moderate (standard deviation of log10Δσ∼0.2log units). Overall, stress-drop estimates are broadly consistent across methods. The proposed multimethod framework quantifies epistemic uncertainty and supports the development of more reliable physics-based seismic hazard models.