Onshore Japanese crustal mainshocks had foreshocks about twice as often within 10 km of Holocene volcanoes in 1998-2024, and aftershock counts did not decrease with heat flow (JMA, M3+)

We tested if surface heat flow controls earthquake clusters in the onshore crust of Japan, as reported for southern California. We used the JMA catalog for 1998-2024 (depth at most 25 km; 83,850 events at M 2.0 or more; replication with 214,374 events at M 1.5 or more), nearest-neighbor cluster trees, GHFDB 2024 heat flow and the GVP Holocene volcano list. All intervals come from a 1-degree spatial block bootstrap. Of 345 M 3+ mainshocks within 10 km of a Holocene volcano, 24.9% had a foreshock. Beyond 80 km, 11.5% of 1,312 mainshocks had one (ratio 2.17, 95% interval 1.49-3.01; M 1.5 catalog ratio 2.02). The volcano-distance slope stayed negative in both periods, after removal of large sequences, with threshold changes and with heat flow, depth and productivity controls. In the full sample, heat flow added no clear information to volcano distance. More than 40 km from a volcano, high and low heat-flow areas had the same foreshock share. Aftershock productivity did not decrease with heat flow (+0.072 log10 units per decade, -0.038 to +0.205); this does not agree with the southern California inverse relation. Productivity decreased with mainshock depth (-0.12 log10 units per 10 km). Tree shape (average leaf depth) had no clear heat-flow link. Limitations: the heat-flow field is sparse; where it is dense, heat flow and volcano distance are collinear (r = -0.78) and their effects cannot be separated. Onshore selection excludes offshore parent events. The results are observational.

earthquake clustering · foreshocks · aftershock productivity · heat flow · volcanoes · Japan · JMA catalog · nearest-neighbor · swarms · crustal seismicity