Development of A Real-Time Shaking Map System Using Low Cost Acceleration Sensors and Its Application for Earthquake Early Warning

Real-time signals from Palert (low-cost MEMS acceleration sensor) network are used to develop a real-time shaking map system based on the Earthworm environment in Taiwan. This system could provide real-time intensity map and estimate magnitude which is determined by shaking covered areas without earthquake location process. In order to derive an empirical strong motion attenuation relationship between the shaking covered areas and their corresponding earthquake magnitudes, we collected the strong motion records from 42 large crustal earthquakes recorded by the Taiwan Strong Motion Instrumentation Program (TSMIP) stations. The shaking covered areas (A) of certain peak ground acceleration (PGA) could be related to M_A as follow: M_A=0.002*(PGA)*log(A)+0.279*log(A)+4.236 Records of three earthquakes are used to test our system performance, they are the Hualien earthquake (M_L 5.6, M_W 5.5) occurred on 7th March 2013, the Nantou earthquake (M_L 6.1, M_W 5.9) occurred on 27th March 2013, and the Nantou earthquake (M_L 6.3, M_W 6.2) occurred on 2nd June 2013. Results show that the first report could be provided at about 10 seconds after the earthquake occurrence and the magnitudes are reported as 5.7, 5.7 and 5.8 for Hualien and two Nantou events, respectively. Finally, the stable report could be obtained at about 20, 19 and 17 seconds after the earthquake occurred and the magnitudes are reported as 5.5, 5.9 and 6.0 for Hualien and two Nantou events, respectively. Base on the result from this study, the real-time shaking map system could provide rapidly real-time shaking map and estimate earthquake magnitude within 1 minute even tens of seconds. It will play an important role in seismic hazard mitigation.