Citation

Katzilieris K, Vlahogianni EI, Wang H. Safety Sci. 2022; 153: e105799.

Copyright

(Copyright © 2022, Elsevier Publishing)

DOI

10.1016/j.ssci.2022.105799

PMID

unavailable

Abstract

The objective of this paper is to analyze the household response behavior of road users during community evacuations in response to the 2018 Attica wildfires in Greece. To achieve this objective, empirical data were obtained from a questionnaire survey completed by residents from the affected areas by the wildfires of the East Attica in 2018. The design of the survey questionnaire was guided by the Protective Action Decision Model - PADM. Logistic regression models and machine learning techniques such as random forests were developed to identify the critical factors influencing the decision to evacuate or not, as well as the mode choice during evacuation.

FINDINGS reveal that the perception of risk, age, years in residence, the number of adults up to 65 years of age in residence, the attempt to obtain information before evacuation, gender, the existence of a prior warning, the number of minors in residence, the level of education and income were the most critical factors to the decision to evacuate during the wildfire. Regarding the mode choice, the most critical factors were the existence of an available vehicle, age, the attempt to obtain information before the evacuation and the perception of risk.

Language: en

Keywords

Decision-making; Evacuation; Logistic regression; Random forests; Wildfire