Investigating Corner-Frequency Uncertainties: Insights from Six Earthquakes (Mw 3.2–3.8) in Ridgecrest, California, and the Korean Peninsula

ABSTRACT Accurate corner-frequency (fc) measurements are crucial for precise stress-drop (Δσ) estimates, essential for understanding earthquake mechanics and predicting ground motions, directly impacting seismic hazard assessments. This study investigated fc variations for six earthquakes (Mw 3.2–3.8) across different tectonic regions: two from the 2019 Ridgecrest sequence in California (R1 and R2) and four from the Korean Peninsula�the 2022 Goesan mainshock (GM) and its foreshock (GF), the 2020 Haenam mainshock (HM) during a swarm-like sequence, and an aftershock (PA) of the 2017 Pohang earthquake. We employed the empirical Green’s function approach, using cross-correlation coefficients (CCs) to select suitable Green’s functions. We explored variations in frequency band, window length, and CC thresholds for cross-correlation, followed by deconvolution and spectral fitting. Our results indicate substantial differences in spectral shapes and fc estimates among the six target events, reflecting diverse source characteristics: R1, R2, and GM (all Mw 3.8) exhibit distinct fc distributions; PA (Mw 3.6) shows the most complex source time functions with one of the lowest fc estimates. Variations in CC criterion and the upper bound of the frequency band used in the CC calculation significantly affected the mode and shape of the fc estimate distribution for each target event, sometimes splitting the distribution into higher and lower fc groups. We adjusted the variables to emphasize the low-frequency components, aiming to recover the overall structure of the source time function. This led to persistent differences in Δσ estimates across the events, despite uncertainties in the individual fc estimates. However, achieving this consistency was challenging without the adjustment, as errors in the fc estimates propagated on a cubic scale. Our findings provide a clear example that systematic fc estimation can obscure stress-drop differences among events by increased uncertainties in Δσ estimates for individual events, especially with different tectonic settings and source complexities.