Multifrequential periodogram analysis of earthquake occurrence: An alternative approach to the Schuster spectrum, with two examples in central California

Abstract

Periodic earthquake occurrences may reflect links with semidiurnal to multiyear tides, seasonal hydrological loads, and ∼14 month pole tide forcing. The Schuster spectrum is a recent extension of Schuster's traditional test for periodicity analysis in seismology. We present an alternative approach: the multifrequential periodogram analysis (MFPA), performed on time series of monthly earthquake numbers. We explore if seismicity in two central California regions, the Central San Andreas Fault near Parkfield (CSAF-PKD) and the Sierra Nevada-Eastern California Shear Zone (SN-ECSZ), exhibits periodic behavior at periods of 2 months to several years. Original and declustered catalogs spanning up to 26 years were analyzed with both methods. For CSAF-PKD, the MFPA resolves ∼1 year periodicities, with additional statistically significant periods of ∼6 and ∼4 months; for SN-ECSZ, it finds a strong ∼14 month periodic component. Unlike the Schuster spectrum, the MFPA has an exact modified statistic at non-Fourier frequencies. Informed by the MFPA period estimates, trigonometric models with periods of 12, 6, and 4 months (Model 1) and 14.24 and 12 months (Model 2) were fitted to time series of earthquake numbers. For CSAF-PKD, Model 1 shows a peak annual earthquake occurrence during August-November and a secondary peak in April. Similar peaks, or troughs, are found in annual and semiannual components of pole tide and tide-induced stress model time series and fault normal-stress reduction from seasonal hydrological unloading. For SN-ECSZ, the dominant ∼14 month periodicity prevents regular annual peaking, and Model 2 provides a better fit (Δ R¯adjusted2: 2.4%). This new MFPA application resolves several periodicities in earthquake catalogs that reveal external periodic forcing.