Associations of emotional divergence in risk communication with forecast error type during typhoon Khanun

This article examines how rainfall forecast accuracy influenced public emotions during the landfall of Typhoon Khanun in South Korea. It combines precipitation forecast and observation data with more than 43,000 real-time posts from NAVER Report Talk, as well as search activity data from Google Trends, NAVER DataLab and Wikipedia. The study uses natural language processing to assess how different types of forecast errors shaped online risk communication and public sentiment.

The key takeaway is that forecast accuracy matters not only as a technical issue, but also as a factor shaping public trust, perception and emotional response during extreme weather events. The study finds that underestimated rainfall forecasts were associated with stronger fear, sadness and confusion, while overestimated forecasts could contribute to fatigue and feelings of over-preparedness. It highlights the need to communicate forecast uncertainty more clearly to support public trust, reduce emotional distress and improve risk communication during future disasters.