It has been asserted that earthquakes cannot be predicted (Geller et al., 1997). This may be true to some extent for the present, but it should not stop further research. The discussion of stress changes in this article gives an example to show that what people observe near the surface may not be the situation at depth. Nonetheless, such surficial observations reflect the movements of the crust associated with earthquakes.
Before the great M 7.8 Tangshan, China, earthquake in 1976, the Douhe and Zhaogezhuang stress-monitoring stations on the supposed fault zone observed significant tensile pulses of ground stress normal to the fault zone. The data are both reliable and reasonable. Analysis of the stress field adjacent to a fracture front shows that, when the fracture front passes by the measuring point, the stress normal to the fracture will first rise and then drop, while the stress at an angle of 30° to the fracture has much less significant variation in the observations. The pulses are attributed to the near-surface tensile fractures generated along the fault zone.