Crossing the behavioral threshold: social amplification and climate risk management in urban flood evacuation
4 affiliations
- 1Faculty of Economics, Shiga University, Japan; National Research Institute for Earth Science and Disaster Resilience, Japan; The Center for Public Service, Portland State University, USA; Corresponding author at: Faculty of Economics, Shiga University, 1-1-1 Bamba, Hikone, Shiga 522-8522, Japan.
- 2Japan Meteorological Agency, Japan
- 3Disaster Prevention Research Center, Aichi Institute of Technology, Japan
- 4Crisis and Environment Management Policy Institute, Japan
Abstract
Early warnings save lives only when they move people to act. Yet during recent urban flood disasters in Japan, evacuation compliance remained extremely low, revealing a substantial intention–action gap at the heart of climate risk management. This study addresses that gap by proposing an Evacuation Threshold Framework (ETF), which extends the Protective Action Decision Model by conceptualizing the conversion from intention to action as a non-linear, threshold-gated process. The framework is tested using a factorial survey experiment conducted in Hiroshima, Japan ( n = 1324), in which each respondent evaluated identical flood scenarios under both Standard Valuation and Inferred Valuation framings. The latter is used as the primary inferential basis because it yields more conservative estimates of evacuation propensity and helps attenuate self-presentation bias. The results yield three main findings. First, the behavioral activation threshold is calibrated at approximately 0.7 on a 0–1 scale, well above the conventional 0.5 cutoff, indicating substantial psychological inertia before intention translates into action. Second, evacuation intention is driven less by logistical frictions than by social and environmental cues: visible peer evacuation exerts a strong contagion effect, while improved shelter privacy substantially raises predicted evacuation rates. Third, even under extremely high perceived risk, predicted evacuation rates remain near zero when pull and contagion cues are absent, indicating that information alone is insufficient to close the intention–action gap. These findings suggest that effective climate risk management must move beyond information transmission toward the behavioral design of protective environments — lowering psychological thresholds and amplifying visible action. Although grounded in urban flood risk in Japan, the proposed framework offers a generalizable behavioral perspective for climate risk management across hazards and settings.
Related
- Integrating social and spatial dimensions in tsunami evacuation modeling: a case study of Panggarangan coast, Indonesia
- Testing the theory of planned behavior in predicting evacuation intentions during storms
- Mapping hazard-access mismatches in wildfire evacuation systems: a multi-facility accessibility framework
- Inter-hemispheric contrast in the sensitivity of polar amplification to sea-ice coverage conditions
- Contour-crossing calculations quantify decadal changes in the pCO2 impacts of inorganic chemical weathering in the North American Arctic
BibTeX
@article{tanaka2026crossing,
title = {{Crossing the behavioral threshold: social amplification and climate risk management in urban flood evacuation}},
author = {Katsuya Tanaka and Kazuhide Akaishi and Takashi Yokota and Fujio Kusano},
journal = {Climate Risk Management},
year = {2026},
month = oct,
doi = {10.1016/j.crm.2026.100901},
url = {https://doi.org/10.1016/j.crm.2026.100901},
note = {CC BY},
}
RIS
TY - JOUR
TI - Crossing the behavioral threshold: social amplification and climate risk management in urban flood evacuation
AU - Katsuya Tanaka
AU - Kazuhide Akaishi
AU - Takashi Yokota
AU - Fujio Kusano
PY - 2026
DA - 2026/10/30
JO - Climate Risk Management
DO - 10.1016/j.crm.2026.100901
UR - https://doi.org/10.1016/j.crm.2026.100901
AB - Early warnings save lives only when they move people to act. Yet during recent urban flood disasters in Japan, evacuation compliance remained extremely low, revealing a substantial intention–action gap at the heart of climate risk management. This study addresses that gap by proposing an Evacuation Threshold Framework (ETF), which extends the Protective Action Decision Model by conceptualizing the conversion from intention to action as a non-linear, threshold-gated process. The framework is tested using a factorial survey experiment conducted in Hiroshima, Japan ( n = 1324), in which each respondent evaluated identical flood scenarios under both Standard Valuation and Inferred Valuation framings. The latter is used as the primary inferential basis because it yields more conservative estimates of evacuation propensity and helps attenuate self-presentation bias. The results yield three main findings. First, the behavioral activation threshold is calibrated at approximately 0.7 on a 0–1 scale, well above the conventional 0.5 cutoff, indicating substantial psychological inertia before intention translates into action. Second, evacuation intention is driven less by logistical frictions than by social and environmental cues: visible peer evacuation exerts a strong contagion effect, while improved shelter privacy substantially raises predicted evacuation rates. Third, even under extremely high perceived risk, predicted evacuation rates remain near zero when pull and contagion cues are absent, indicating that information alone is insufficient to close the intention–action gap. These findings suggest that effective climate risk management must move beyond information transmission toward the behavioral design of protective environments — lowering psychological thresholds and amplifying visible action. Although grounded in urban flood risk in Japan, the proposed framework offers a generalizable behavioral perspective for climate risk management across hazards and settings.
SN - 2212-0963
ER -