The devastation caused by a large magnitude earthquake is well known. We remember the 1906 San Francisco damage and fatalities from old pictures we have seen. In Tangshen China, a quarter million people died during a Magnitude 7.6 earthquake. In contrast though, a year earlier, fatalities were avoided when another city in China was evacuated one day before a severe quake happened. How did that happen? Did they predict it? Well yes and no. The people noted changes over a period of months in ground positions and water levels. Extensive modifications in animal behavior were observed. In addition, tremors preceded the quake for several months. This information was well known and acted upon. The Magnitude 7.3 quake hit an evacuated city. This was more matter of observation and luck than any skill. The Chinese took a conservative and safe position by getting out of Dodge.
To predict an earthquake event would be a major public health coup. In order to do so, one needs to predict exactly where it will happen and when. The method should also be reproducible. It must also predict magnitude of the earthquake. At this time official organizations involved in earthquake predictions do not recognize any current method as being accurate. It cannot be done, now and by the nature of the beast, never will be done.
What we have are seismic maps produced for different areas that predict an earthquake likelihood of a certain magnitude over a large span of years. This gives the probability of an event in an area over time. Tremors and foreshocks are not necessarily predictive of a major event.
At the outset, we eliminate psychics and so on as not scientifically testable. They are unable to meet the criteria cited above. If we consider other possible methods and we come to studying electromagnetic fields and other related methodologies. Here scientists try to determine if electro-magnetic radiation is important. Several satellites are being used to study this. Some feel this is important for tipping off animals to an impending quake. However, the data for small animal behavior is lacking. The Japanese felt that catfish could predict changes but this has not panned out in reality. Small animal sensitivity is more anecdotal reporting with no scientific backing yet found. Others include luminescence changes in the atmosphere which was noted in the Kobe quake. Changes is visible radiation were noted and thought to be due to an outpouring of silicon crystals from the earth. Radon and other chemical changes in the soil or water have been investigated. Tidal changes were thought to be related. However, none of these methods can be used to meet the above criteria for successful predictions. Presently, there is no method for prediction of earthquakes and it is unlikely for us to find one.