14:45 〜 15:00
[SCG53-05] 緊急地震速報のための拡張IPF法（IPFx法）とSNSを利用した結果の見える化
キーワード：緊急地震速報、拡張IPF法、強震動、震源推定
We improve the integrated particle filter (IPF) method, which is a new source determination algorithm recently incorporated into the Japanese earthquake early warning system. The problem of the current IPF method is that it uses specific trigger information as an input and is therefore applicable to only limited seismic networks. This study proposes the extended IPF (IPFx) method to deal with continuous waveforms and merge all Japanese real-time seismic networks into a single framework.
The new source determination algorithm processes seismic waveforms in two stages. The first stage (single-station processing) extracts trigger and amplitude information from continuous waveforms. The second stage (network processing) accumulates information from multiple stations and estimates the location and magnitude of ongoing earthquakes based on Bayesian inference. In a 10-month continuous online test, this method performed better than the JMA EEW method in detecting earthquakes with a maximum seismic intensity >=3 as per the JMA manual catalog. By merging multiple networks into a single method, both the P-wave detection and the event detection speeds were significantly improved compared to those of the current JMA EEW method.
We also developed an automatic processing system to visualize the results of the IPFx method in real-time. We use “Net::Twitter”, a perl interface to the Twitter API. The results of the IPFx method (time histories of the estimated source parameters, time histories of the trigger information, and waveforms aligned as a function of distance from the estimated location) are uploaded online a few minutes after the event detection. Currently, the tweets are protected by password and open for only limited users. The advantage of using twitter is that we can access the results anytime and anywhere. We can monitor the performance of the online IPFx system without accessing servers. The results will be uploaded automatically without human interactions. The tweets also work as a super-rapid earthquake report. The IPFx method can detect earthquakes as small as M=1 for inland, so we are able to monitor the seismicity of small earthquakes with only a few minutes latency.
The new source determination algorithm processes seismic waveforms in two stages. The first stage (single-station processing) extracts trigger and amplitude information from continuous waveforms. The second stage (network processing) accumulates information from multiple stations and estimates the location and magnitude of ongoing earthquakes based on Bayesian inference. In a 10-month continuous online test, this method performed better than the JMA EEW method in detecting earthquakes with a maximum seismic intensity >=3 as per the JMA manual catalog. By merging multiple networks into a single method, both the P-wave detection and the event detection speeds were significantly improved compared to those of the current JMA EEW method.
We also developed an automatic processing system to visualize the results of the IPFx method in real-time. We use “Net::Twitter”, a perl interface to the Twitter API. The results of the IPFx method (time histories of the estimated source parameters, time histories of the trigger information, and waveforms aligned as a function of distance from the estimated location) are uploaded online a few minutes after the event detection. Currently, the tweets are protected by password and open for only limited users. The advantage of using twitter is that we can access the results anytime and anywhere. We can monitor the performance of the online IPFx system without accessing servers. The results will be uploaded automatically without human interactions. The tweets also work as a super-rapid earthquake report. The IPFx method can detect earthquakes as small as M=1 for inland, so we are able to monitor the seismicity of small earthquakes with only a few minutes latency.