The b-value in the 2011 Tohoku-oki high-slip zone stayed above its 1998-2011 level through 2024 after depth-mix, control and flag checks (JMA, M 3+); a faulting-type change is not excluded

Question: did the Gutenberg-Richter b-value in the high-slip zone of the 2011 Tohoku-oki earthquake return to its pre-mainshock level, as a rapid stress recovery claim (Tormann et al. 2015) says? Or did it stay high, and can changes in the event population or in the catalogue explain this? Data: the JMA hypocentre catalogue for 1998-2024, Slab2 interface depths and GCMT faulting types. Method: Aki-Utsu b at fixed Mc 3.0 and 3.5, b-positive with floors 2.0-3.0, reweighting to the pre-2011 mix of upper-plate, interface and slab events, and comparison with two control regions (off Hokkaido and Izu-Bonin). Uncertainty comes from bootstrap intervals. Result: in the high-slip box (37.5-39.0 N, 142.5-144.2 E), b at Mc 3.0 was 0.68 (n = 817) in 1998 to March 2011. It was 0.79 in April 2011 to 2013, 0.91 (n = 423) in 2014-2024 and 0.97 (n = 134) in 2019-2024. No post-mainshock window returned to the pre-mainshock level. The 2014-2024 change was +0.24 (95% interval 0.15 to 0.33). It was +0.28 with the pre-2011 depth mix, +0.42 at Mc 3.5 and +0.24 with b-positive at a 3.0 floor. Minus the control changes, it was +0.18 (0.09 to 0.28) against Hokkaido and +0.23 (0.13 to 0.32) against Izu-Bonin. The share of slab events increased from 0.29-0.43 to 0.82 in 2012-2013, but b increased inside the interface and slab classes separately. From 2016, most small events have automatic hypocentre flags. Because of this, b-positive at M 2.0 depends on the flag selection and we do not use it as evidence. Results at M 3.0 or more do not depend on it. The Hokkaido control also increased by 0.05-0.10. Limitations: GCMT shows that thrust events (20 of 24 before 2011) were replaced by normal-faulting events (10 of 17 in 2011-2013). We cannot control faulting type for small events, so this change can explain part or all of the increase. Offshore depths are uncertain, the late samples are small, we did not decluster the catalogue and we did not repeat the exact spatial sampling of Tormann et al.

b-value · Tohoku-oki earthquake · stress recovery · Gutenberg-Richter · b-positive · JMA catalog · Slab2 · Japan Trench · catalog artefacts · aftershocks · megathrust