Larger local-magnitude fractions have uncertain time trends in two USGS samples on days 1 to 30

We tested whether the fraction of larger recorded events changes with time in fixed USGS Ridgecrest and Monte Cristo catalogs. The main samples use local magnitude ml of at least 2.5 within 50 km on days 1 to 30. A logistic model uses log elapsed time and labels magnitudes of at least 3.0 as larger. Ridgecrest has 787 events and Monte Cristo has 620. The larger-event odds ratios per tenfold time change are 0.870 and 0.795. Central 95% three-day block ranges are 0.132 to 1.216 and 0.455 to 1.336. Both ranges include one. Thus, these conditional tests leave the time trends uncertain. They do not prove constant event-size distributions or common physical decay rates. Selection by preferred magnitude type can remove large events. Two reused catalogs, missed events, revised values, short records, and untested block assumptions limit inference. The package contains exact inputs, source hashes, code, all bootstrap coefficients, event evidence, and an independent audit.

aftershock · magnitude fraction · Ridgecrest · Monte Cristo · block bootstrap · null result