Observational seismology has a large array of algorithms at its disposal to locate an earthquake from waveform recordings at a number of seismograph stations. The task in any case is to locate an event in three spatial coordinates (hypocentre), in general longitude, latitude, depth, and in time (origin time). Since different seismic phases travel with different propagation velocities through the earth, the identification of individual phases and their delays, which depend on the distance from the earthquake source location, already constrain origin time and source to station distance. In particular the location problem in two spatial coordinates (epicentre) and the determination of origin time is well posed. Determination of depth is in general not a well posed problem, due to the geometry of the problem where seismic stations are confined to the earth’s surface.
The problem setting for Earthquake Early Warning (EEW) is different since the time needed to observe and analyze the full wave-form over the total duration of an earthquake cannot be afforded. It thus comes at no surprise that algorithms which determine the location of a source from just the first arrival times of a wave (in the case of earthquakes, the P-wave) have initially been developed in acoustical engineering (Friedlander, 1987; Schau and Robinson, 1987; Huang and Benesty, 2000; Pirinen, Pertila, and Visa, 2003; Pirinen, 2006; Gillette and Silverman, 2008) rather than in seismology. Complete seismograms provide a lot more information that just first arrival times of a single phase.
A similar remark applies to algorithms proposed for estimating the final magnitude of an earthquake from just a few seconds of P-wave recordings. This approach remains controversial among seismologists and more recent research suggests that incorporating real-time displacement data from Global Navigation Satellite Systems (GNSS) may provide more robust estimates and updates while a large earthquake unfolds.
The problem setting for Earthquake Early Warning (EEW) is different since the time needed to observe and analyze the full wave-form over the total duration of an earthquake cannot be afforded. It thus comes at no surprise that algorithms which determine the location of a source from just the first arrival times of a wave (in the case of earthquakes, the P-wave) have initially been developed in acoustical engineering (Friedlander, 1987; Schau and Robinson, 1987; Huang and Benesty, 2000; Pirinen, Pertila, and Visa, 2003; Pirinen, 2006; Gillette and Silverman, 2008) rather than in seismology. Complete seismograms provide a lot more information that just first arrival times of a single phase.
A similar remark applies to algorithms proposed for estimating the final magnitude of an earthquake from just a few seconds of P-wave recordings. This approach remains controversial among seismologists and more recent research suggests that incorporating real-time displacement data from Global Navigation Satellite Systems (GNSS) may provide more robust estimates and updates while a large earthquake unfolds.