Rapid Assessment of Seismic Intensity Based on Sina Weibo —A Case Study of the Changning Earthquake in Sichuan Province, China
2021
Abstract Disaster information acquisition and assessment in China primarily depends on disaster-related governmental departments at all levels. As new challenges faced by disaster emergency management arise, a gap still exists in the timeliness and completeness of rapid disaster assessment. The development and popularity of social media has opened up new channels. Social media data, containing massive amounts of disaster information, has the advantages of timeliness, efficiency, and multiple spatiotemporal scales and can supplement existing methods. Based on the dominant social media platform in China, Sina Weibo, this paper proposes a method of extracting intensity information from Weibo texts, including social perceptions such as shaking, emergency reaction, mood and visible damage, which can support the rapid assessment of seismic intensity. By taking the Changning earthquake in Sichuan Province, China, as a case study, we verify the feasibility of hazard information extraction from Weibo. The results show that the intensity information derived from Weibo in the short-term can respond to a sudden earthquake promptly, then effectively identify the main earthquake affected area, locate the possible orientation close to the meizoseismal area and depict the situation in 10 minutes. We also propose a grid-based correction method, which synthesizes the hotness-intensity matrix and a seismic attenuation-based model. Compared with taking the seismic attenuation model directly, the proposed correction method improves the accuracy of correct recognition rate of seismic intensity assessment to 82% on average and reduces the false recognition rate significantly. The proposed framework of rapid assessment and correction reveals that the combination of social perception and hazard intensity plays a valuable and promising role as a supplement of traditional disaster assessment approach.
Keywords:
- Correction
- Source
- Cite
- Save
- Machine Reading By IdeaReader
29
References
0
Citations
NaN
KQI