Driver Decisions Based on Flood Warning Information

Floods may lead to devastating losses, and their risks should be communicated to people to avoid these potential losses. Risk communication is vital in assisting people to make educated decisions based on the potential upcoming event (Chen, 2020). A multi-granularity approach to risk communication allows some users to understand the risk in general while others receive detailed information about the risk at hand (Jorgensen et al., 2015). Flood risk communication may be conveyed through mobile navigation applications, yet there is a lack of literature supporting the design of these types of warnings. The risk of driving through a flood also depends on the vehicle, the flood depth, and possibly the driver’s gender and past flood experience (Pearson & Hamilton, 2014; Drobot et al., 2007).

This study aimed to examine how people perceive risk, trust, and which factors influence decision-making when given a flood warning through a mobile navigation application in a simulated driving scenario. A total of 93 participants were recruited to take part in this 60-minute study. Participants first filled out the Motion Sickness Susceptibility Questionnaire (MSSQ; Golding, 2006) to ensure they had a low likelihood of experiencing motion sickness in the driving simulator. Those eligible to continue conducted a practice drive in the STISIM Drive (Build 3.20.03) driving simulator, where they manually drove a vehicle through an urban setting with the Logitech G27 Racing Wheel and pedals. After the practice drive, they continued with the first of four flood scenario drives, which were ordered with a Latin square design. The participant was tasked to drive to a restaurant and the navigation system would provide directions on their dashboard. Near the end of the drive, the participant was presented with one of the four flood-information types (flood, no flood, flood of 6 inches, flood of 6 inches maximum) and had to decide whether to take the alternate route presented by the navigation system, or continue driving on the original route, and potentially through a flood. After, the scenario ended and participants were directed to answer survey questions testing their understanding, measuring their perceived risk (Simon et al., 2000), and their trust in the navigation system (Jian et al., 2000). After repeating this process until they had driven through all four flood scenarios, they were dismissed.

The results showed that participants were more risk-avoidant than risk-seeking and kept the same route less when a flood of any type was presented compared to when no flood was presented. However, participants did not avoid the flood more when given a general flood warning than a flood warning with the depth of the flood, showing the level of detail did not significantly influence the drivers’ decisions. It is possible that those who drove through the flood overestimated their vehicle’s ability to drive through the flood. Gender did not significantly predict the participants’ decisions, but perceived risk did. Results also showed that there was no difference in trust between the flood information conditions. Trust was consistently high regardless of the warning presented.

This study helps deepen the understanding of human decision-making in driving through floods, and how drivers perceive flood warnings. It also implies that providing any type of flood information helps the drivers’ decision-making process. One limitation of this study is that all scenarios followed a similar route and had the same destination. Future studies may include more dynamic scenarios with different destination locations.