Causal Effects of Perceived False Alarm Ratio on Flood Protective Actions during Heavy Rain Warnings in Japan: A Case Study of Tropical Storm Yun-Yeung in 2023

Abstract Although the accuracy of weather forecasts and warnings has improved over time, the issue of false alarms is ongoing. Frequent false alarms can undermine public trust, potentially discouraging individuals from taking protective action in response to future forecasts and warnings. Understanding the causal effects of perceived false alarms on protective actions is thus critical. This study investigated the effect of the perceived false alarm ratio (FAR) on protective actions during heavy rain warnings in Japan, as well as the heterogeneity of the effect across individuals. The analysis utilized data ( n = 419) from a survey conducted among residents of Chiba Prefecture, where a heavy rain warning was issued for Tropical Storm Yun-yeung in 2023. In this study, a high perceived FAR prior to the tropical storm was treated as the treatment variable, while the adoption of protective actions during the warning period served as the outcome variable. Propensity score analysis and causal forest methods were employed to estimate causal effects and their heterogeneity. The results revealed that the perceived FAR had nonsignificant causal effects on any type of protective action, with no substantial heterogeneity across individuals. These results suggest that a higher perceived FAR does not necessarily inhibit people’s protective actions and that the effects do not necessarily vary across individuals. This study provides valuable insights into the relationship between false alarms and public responses and suggests possible directions of future research.