Understanding Weather Forecast Communication, Interpretation, and Use through Analysis of Twitter Data

Over the last few decades, advances in science and technology have dramatically improved the skill of weather forecasts and warnings. However, scientists still face major challenges in effectively conveying weather risk information to both professional and public audiences. At the same time, advances in information and communication technology are transforming how people access, share, and use weather-related information. This presentation will discuss work to understand and improve communication, interpretation, and use of risk information as hazardous weather approaches and arrives, in the context of the dynamic modern information environment. The research presented will focus on analysis of textual and image data collected from Twitter narratives during recent hurricanes, including Hurricane Sandy in 2012 and Hurricanes Harvey and Irma in 2017. The Twitter data offer a quasi-real-time record of what hurricane-related information was conveyed on Twitter, who generated and propagated that information online, and how that information was interpreted and used by Twitter users as the hurricane threats evolved. Results will examine the diffusion of different hurricane risk information on Twitter and the evolving perspectives of at-risk Twitter users as new risk information was communicated. By building understanding of how hazardous weather information is communicated and used in its dynamic real-world context, this work aims to help meteorologists, public officials, and others improve communication of and responses to weather-related risks.