From Wikipedia, the free encyclopedia.

An earthquake prediction is a prediction that an earthquake of a specific magnitude will occur at a specific location and time. Seismologists are not currently able to predict earthquakes with such accuracy, instead they focus on calculating the seismic hazards of a region and the probabilities that a given earthquake will occur.
Earthquakes were once thought to be random geologic events without cycles or patterns. In some ways, earthquakes behave similar to the stock market, or types of events in nature that have properties described by chaos theory.
Like stocks, the pattern of earthquakes is quite capable of being correlated with anything, sometimes based on only one observation. Most empirical sciences, particularly the medical sciences, prefer to have hundreds, thousands, or even millions of data points before drawing conclusions regarding such things as the link between cigarettes and cancer, or fatty-foods and heart disease.
With regard to predictive seismology, humans have tried to associate an impending earthquake with such things as animal behavior, electromagnetic fields, weather conditions, unusual clouds, radon gas in ground water, water level in wells and so on - thereby suggesting that the dataset of observed seismicity is dependent on a large number of external variables.
Controversy arises as a result, since conclusions usually should not be made from a small data set unless a well understood physical phenomenon is present, particuarly when the data set is noisy or there are questions regarding how it is gathered.
Based on the historic record of the various published efforts to predict a quake, it might be easy to conclude that earthquake prediction is usually an imprecise, but nonetheless scientifically and socially useful art.