Earthquake forecasting

Earthquake forecasting is a specialized branch of geophysics. It focuses on seismology, which is the study of earthquakes. Forecasting involves the probabilistic assessment of seismic hazards. This means scientists estimate the likelihood of damaging earthquakes. They look at frequency and magnitude in specific areas. These estimates usually cover long periods like years or decades.

It is important to distinguish forecasting from earthquake prediction. Prediction attempts to pinpoint an exact time and location. It also tries to name the exact magnitude. Currently, reliable earthquake prediction is not achievable. Forecasting is different because it estimates probabilities. It provides a range of possibilities rather than a single certain event. This distinction helps researchers focus on what is actually possible.

One major theory is the elastic rebound theory. Even very stiff rock is not perfectly rigid. When huge tectonic plates move, they apply a large force. The Earth's crust begins to bend or deform under this pressure. This deformation is known as strain. Eventually, the strain becomes too great for the rock to hold. The rock breaks, usually at an existing fault. This slippage is the earthquake itself. The rock then rebounds to a less deformed state. This process releases energy as seismic waves.

Scientists also use the characteristic earthquake model. This model looks at specific segments of a fault. These segments have fixed lengths and specific rock types. Because these properties are fixed, earthquakes on one segment often have similar characteristics. They tend to have a similar maximum magnitude. They also require a specific amount of accumulated strain to occur. Because plate motion is continuous, strain builds up steadily. This means these earthquakes may recur at somewhat regular intervals.

Another approach is the seismic gap model. Tectonic plates slip past each other at contact points. Every section must eventually slip during the long-term cycle. However, different sections are at different stages of strain. The seismic gap model suggests the next big quake occurs in gaps. These are areas where recent seismic activity has not relieved the strain. This model was used for many Pacific Rim forecasts. However, recent studies suggest it may not forecast large earthquakes well. Some research shows a long quiet period does not always increase potential.

History shows that these scientific models can fail. In the 1970s, scientists were very optimistic about prediction. By the 1990s, many began to question if prediction was possible. A famous example involves the Parkfield earthquakes. Researchers used data from 1857, 1881, 1901, 1922, 1934, and 1966. They forecasted an earthquake around 1988 or 1993. However, the earthquake did not occur until 2004. This failure raised doubts about the characteristic earthquake model.

Modern forecasting is highly practical for public safety. In California, the UCERF3 is the latest official forecast. It was prepared by the Working Group on California Earthquake Probabilities. This group includes the USGS and the California Geological Survey. The forecast examines 2,606 fault subsections. It helps estimate the chance of an M 6.7 earthquake. This magnitude is similar to the 1994 Northridge earthquake. These estimates help set building codes and insurance rates. They also guide disaster planning and public policy.

Scientists also study electromagnetic changes as potential signals. There is an active debate regarding these signals. A ten-year study in California found modest magnetic signals. These occurred 24 to 72 hours before certain earthquakes. Another study used Swarm satellite data to find patterns. They found a correlation between magnetic fields and TEC anomalies. These anomalies appeared one to seven days before earthquakes of M 4.0 or higher. These signals are part of the complex study of earthquake precursors.