Tweeting Tsunami: Influence and Early Warning in British Columbia

Amanda Oldring, Alexander Brand, Antonina Milekhina


Early warning has been studied on Twitter, but Canadian research is exceptionally rare. British Columbia (BC) is vulnerable to tsunamis, and warning systems are not ubiquitous. Establishing pre-event networks can contribute to understanding early warning dissemination potential in BC. This study investigated a 1,932 follower network sampled from the Twitter account @NWS_NTWC, a key tsunami notification source for BC. It combined social network analysis, content analysis, and a user survey with 120 identified opinion leaders to investigate stakeholder connections and how influencers operate as a proxy for describing the network’s capacity to warn areas in proximity to tsunami risk. Social network analysis identified the network’s composition and influencers, termed opinion leaders. Profile data content analysis provided the users’ stakeholder group descriptions and locations. An opinion leader survey was conducted to perform validity checks against the network findings. The results describe how networks can optimize warnings on Twitter. Key outcomes include a longitudinal baseline, network-driven decision--making techniques, strategies for alerting at-risk areas, and a method for detecting influential users. This study concludes that remote rural communities are accessible by the CSN, that they contain opinion leaders, and that they have their own unique stakeholder group compositions. Although representation is low, it is still existent. Following this line of thought, Emergency management can improve warning capacity by fostering ties between the CSN’s most peripheral groups: citizens, academics, hobbyists, and the public sector. The government, in particular, could expand connections internally as well as between citizens, first responders, and the public sector. The network data suggest that first responders, especially rural ones, and the public sector need to work on ties to media and government to increase warning range. Overall, greater mutual awareness of stakeholder groups is needed. In this way, pre-event networks not only visualize the voices at the table but also help to strategically target improvement initiatives.



← Schedule