Anomalous decrease of relatively large shocks and increase of the p and b values preceding the April 16, 2016 M7.3 earthquake in Kumamoto, Japan

TL;DR

This study analyzed seismicity changes before the 2016 Kumamoto M7.3 earthquake, finding a decrease in large shocks and increases in p and b values, indicating stress reduction possibly due to slow slip growth.

Contribution

It provides new insights into seismic precursors by quantifying changes in seismicity rates, p, and b values prior to a major earthquake, linking them to stress alterations.

Findings

Significant decrease in M>=3 earthquake occurrence one day before the mainshock.

Increased p value (~1.5) indicating faster decay of seismicity.

Significant increase in the b value (~0.3-0.5) in the northern area.

Abstract

The 2016 Kumamoto earthquakes started with a magnitude (M) 6.5 quake on Apr 14 on the Hinagu fault zone (FZ), followed by active seismicity including an M6.4 quake. Eventually, an M7.3 quake occurred on Apr 16 on the Futagawa FZ. We investigated if any sign indicative of the M7.3 quake could be found in the space-time changes of seismicity after the M6.5 quake. We determined in advance the threshold magnitude, above which all earthquakes are completely recorded. We then showed that the occurrence rate of relatively large (M >= 3) earthquakes significantly decreased 1 day before the M7.3 quake. Significance of this decrease was evaluated by one standard deviation of sampled changes in the rate of occurrence. We next confirmed that seismicity with M >= 3 was well modeled by the OU law with p ~ 1.5 +/- 0.3, which indicates that the temporal decay of seismicity was significantly faster than a typical decay with p = 1. The larger p value was obtained when we used data of the longer time period in the analysis. This significance was confirmed by a bootstrapping approach. Our detailed analysis shows that the large p value was caused by the rapid decay of the seismicity in the northern area around the Futagawa FZ. Application of the slope (the b value) of the GR frequency-magnitude distribution to the spatiotemporal change in the seismicity revealed that the b value in the northern area increased significantly, the increase being db = 0.3-0.5. Significance was verified by a statistical test of db and a test using bootstrapping errors. Based on our findings, combined with the results obtained by a stress inversion analysis performed by the NIED, we suggested that stress near the Futagawa FZ had reduced just prior to the occurrence of the M7.3 quake. We proposed, with some other observations, that a reduction in stress might have been induced by growth of the slow slips on the Futagawa FZ.