Geohazard types, households' behavioral decision-making, and livelihood strategies: Empirical evidence from Southwest China

Households' behavioral decision-making is critical for disaster prevention and mitigation, and also strengthening settlement resilience. To investigate the relationship between geohazard types and behavioral decision-making, this study focused on landslides and debris flow in Southwest China's mountainous regions. Using random sampling, 63 villages threatened by two different geohazard types in Chongqing and Sichuan were studied, then 1019 sample questionnaires in total were kept. We used statistical methods to analyze differences in households' individual characteristics and behavioral decision-making as a result of different geohazard types. Propensity score matching was used to determine whether different geohazards varied the behavioral decision-making of households. There were four main findings. (1) Following model fitting, the sample's mean absolute deviation reduced from 56 % before matching to 4.9 %, demonstrating that the kind of geohazard has an important effect in behavioral decisions, after controlling for other variables. (2) Peer effects in relocation behavior regardless of the type of geohazard threat to the household(ATT = 0.0098,P > 0.1). (3) Various geohazard characteristics resulted in different behavioral decisions, and different behavioral decisions reflected the characteristics of multiple geohazards. (4) Different geohazards had different attributes, and households had diverse perceptions and impressions of them, leading to differences in behavioral decisionmaking. From a micro perspective, this study focuses on households' behavioral decisionmaking in disaster prevention under the influence of various types of geohazards, confirming that people's behavioral decision-making is influenced by geohazard type, which contributes to the theoretical discussion of disaster prevention behavior in academia and provides a richer perspective for bottom-up disaster risk governance.