New τc regression relationship derived from all P wave time windows for rapid magnitude estimation

Two issues related to the average period t(c) early-warning parameter are the magnitude saturation effect on large earthquakes and considerable scatter for small earthquakes. To reduce the effect of these two issues on earthquake early-warning systems, we introduce a new tc regression relationship derived from all P wave time windows (PTWs) in high-pass filtered (T = 0.075 Hz) strong-motion data for three damaging moderate-to-large earthquakes. Our results show that this relationship provides a better and more stable magnitude prediction than those derived from 3 s PTW without a saturation effect on large earthquakes with M < 7.5. It is expected that fewer false alerts (those outside the magnitude uncertainty tolerance) would be issued. Additionally, a reduction of the initial PTWs to 1-2 s and evolutionary calculation with an expanding window allow more lead time for small-to-moderate events.