M5 event excess persisted on days 7-30 within 200 km of 95 deep M6.5 to below 8 earthquakes in 2001-2024

We tested whether excess nearby events after deep earthquakes persist beyond day 7. We reused a fixed USGS ComCat snapshot and 95 selected mainshocks deeper than 300 km in 2001-2024. The mainshocks had preferred magnitude 6.5 to below 8. We counted M5 or larger events with a depth-aware 200 km rule. On days 7-30 there were 32 after counts and 15 before counts. Their difference was 0.00778 events per mainshock per day, with a conditional 95% three-year block interval of 0.00375 to 0.01256. The days 1-7 excess rate was 0.03684. The early-minus-late rate interval was positive in the main test but included zero after removal of future doublets or overlapping windows. The late excess stayed positive in those checks, with M5.5 neighbors and with Earth chord distance. Its interval included zero at 100 km and 300 km. This is a time-window benchmark in previously used data. It does not identify caused aftershocks, a decay exponent or physical recovery. Sparse counts, mixed magnitude types, selection, location error, regional dependence and catalog revision limit inference. Code, frozen data, pair counts, uncertainty draws, a figure and an independent scalar audit are included.

deep earthquakes · temporal excess · aftershock counts · USGS ComCat · block bootstrap · event persistence