A statistical approach for comprehensively understanding crustal activity in terms of seismicity and strain rate (II)

For comprehensively understanding crustal activity, we have investigated the spatiotemporal relationships between different geophysical measures with a fine grid format by comparing between them in some representative inland regions of Japan. In the comparing process, in which we used the following 4 geophysical measures: dilatation rate, maximum shear strain rate, seismic energy, and the number of earthquakes, definition of a new statistical index clearly characterizing an aspect of crustal activity for each Japanese inland area is a key procedure to capture the feature of crustal activity. We defined a statistical index which expresses the strain rate averaged over the grids in a target area with a seismic measure being more than 80% of its maximum for each time period. Utilizing thus defined statistical index and noticing some regions surrounding the source areas of representative M6-class earthquakes including the 2004 Mid-Niigata Prefecture Earthquake, we previously reported a significant statistical tendency as follows: relatively larger earthquake energy was apparently released by smaller-magnitude earthquakes prior to the mainshocks in area(s) of smaller strain rates rather than in those of larger ones. We here show that the obtained statistical tendency, or the temporal change in the statistical index for the 2004 event, turned out to be caused by the decrease in strain rates in area(s) of relatively high seismicity, which had not almost been dislocated before the 2004 quake. This illustrates that comparison between different geophysical measures includes much more important information than the use of only one geophysical measure does in comprehensively understanding of crustal activity. Although the result is an important step for comprehension of crustal activity, we still have no information on whether or not the statistical tendency is a rare case in the Japanese islands. Therefore, we next need to examine and statistically classify the spatiotemporal relationships between geophysical measures for all other regions. For the next purpose, we further developed a searching tool with which we can automatically detect all the regions revealing a specific spatiotemporal relationship between geophysical measures, or the corresponding temporal change in the statistical index. Using the developed tool, we automatically searched for other regions in which the way of spatial contribution to a temporal change of the statistical index is very similar to that for the case of the 2004 Mid-Niigata Prefecture Earthquake. This would definitely be another fundamental step for understanding crustal activity from a statistical viewpoint. We also introduce the preliminary result obtained by the automatic search.