Discussion on "Are there earthquake precursors that can lead to useful predictions?"

Abstract

In the past tens of years, many earthquake precursors were detected in the earthquake prediction exploration. And some of the precursors have passed the evaluation criterion of proposed earthquake precursors. Also the group characteristics for these proposed precursors were summarized. However, the earthquake prediction based on these precursors together with empirical statistical methods is not very successful. We consider that the reasons are as follows: (1) precursors are hard to select for each specific earthquake because of the large scale of the seismogenic zone; (2) the precursor observations on the existed earthquake cases are too short compared to the long recurrence time; (3) the seismogenic environment is complex and closely related to the dynamics and tectonic backgrounds, which caused the large differences of the precursory anomalies for each earthquake; (4) the precursory anomalies are observed on the surface of the earth. And they give only indirect reflection of the stress or strain changes in the deep area. The relationship between the observed anomalies and the precursors of earthquakes still needs be investigated in detail. Therefore, it is clear that the earthquake does not exactly follow the statistical behaviors summarized so far. We need direct measurements of underground stress/strain state for capturing earthquake precursors, instead of the measurements at individual sites. Earthquake prediction also need rely on physical models based on spatial-/temporal- dependent physical field, instead of empirical and statistical approaches.