In the United States, the areas that most frequently experience earthquakes are the coast of California, Oregon, and Washington, the southern coast and Aleutian Islands of Alaska, Hawaii, and the mountain west from the Rocky Mountains to the Pacific coast. The central and eastern United States rarely experience significant earthquakes.

Earthquake epicenters compiled on a map show that, on a global basis, most earthquakes occur around the rim of the Pacific Ocean, in the mountains of southern Asia from China to the Middle East, and in the Mediterranean Sea area. Earthquake epicenters also trace the mid-ocean ridges across the floors of the oceans.

Because nearly all earthquakes occur on faults, determining seismic risks on a finer scale largely consists of identifying, mapping, and studying active faults in each state or region. However, many active faults are hidden, either because any scarps they formed at the surface have been eroded or covered by sediments, soil, and vegetation, or because they are blind faults. A hidden fault is often not identified and located until one or more significant earthquakes has occurred on it and the seismic waves have been studied to determine its location and type of fault motion.

Information used to determine the seismicity of an area includes:

• frequency of earthquakes in the past, as deduced from:
  • historic records
  • geologic studies that examine evidence of the prehistoric earthquake record
• location of known active faults
• seismologic data collected on recent earthquakes that have occurred in the area
• tectonic setting of the area in terms of proximity to plate boundaries, and information about the plate boundary if one is nearby
• stress and strain being experienced by the crust in that area based on measurements from GPS equipment and from stress and strain measurements conducted in boreholes
• underground geologic layers and structures in that area based on cross-sections from geologic mapping, data from drilling, and remote imaging of deeper layers of the crust and mantle

Based on this information, the seismic risk of a particular area can be quantified statistically. For example, the odds of a major earthquake happening in the next century, or in the next 10 years, can be estimated for a specific seismic zone.

However, no scientific method has yet been developed that can predict precisely when the next earthquake in a specific region will happen, where it will happen, or what its magnitude will be. Scientists have looked into using such possible pre-earthquake indicators as ground tilting, changes in well water levels, changes in radon gas in groundwater near fault zones, changes in electrical conductivity in the earth around faults, changes or patterns in seismic activity that can be measured by seismometers even though it is not felt by humans, and strange animal behavior which, according to numerous, largely unconfirmed anecdotes, takes place before an earthquake. But, so far, none of these types of data have been found to lead to reliable earthquake predictions.

People have also looked into correlations between earthquakes and phases of the Moon, earthquakes and the time of day (such as dawn when the Sun is first shining on the ground), and so on. No connections have been found between earthquakes and these other types of phenomena.

As the study of seismicity stands now, we can identify which areas on earth will undergo major earthquakes in the coming decades and centuries, we can delineate which areas on earth are at risk for the most powerful types of earthquakes, and map the coastal areas that are most at risk of being inundated by a tsunami, but we cannot pinpoint in advance the date or location of the next major earthquake.