Using machine learning and satellite radar to detect "silent" earthquakes in the Zagros Mountains, Iran

Earthquakes are the single biggest natural hazard facing Iran. 9 out of the 10 deadliest natural disasters in Iran in the 20th Century were earthquakes, including the deadliest individual event on record in the country; a single earthquake in 1990 that killed an estimated 50,000 people. These damaging earthquakes also cause major financial losses of hundreds of millions of dollars, and the combined human and financial losses present a major development challenge to the country.

A significant scientific obstacle that prevents accurate estimates of earthquake hazard, in Iran but also globally, is the unknown importance and frequency of “silent” earthquakes. In regular earthquakes, tectonic faults in the Earth’s crust release stress in sudden episodes of slip and cause violent shaking. But in silent or slow earthquakes, these same faults instead slip slowly over hours-to-years, releasing stress without radiating seismic waves. These silent earthquakes can have a major influence on the occurrence of regular earthquakes, and can raise or lower local earthquake hazard. But they are hard to measure, and so it is not known how common they are. Therefore, characterising when and where these silent earthquakes occur is critical for accurately estimating seismic hazard.

See more project case studies from Durham University