Agent-Based Modelling for Flood Evacuation as Inclusive Disaster Risk Reduction: Pilot Participatory Action Research with 11- and 12-year-old children from a Japanese School

Climate change has increased the prevalence of natural hazard threats like flooding across the world, which has resulted in a heightened risk in already flood-prone communities. As a result, more focus needs to be placed on climate change adaptation such as ensuring effective response (e.g. flood evacuation) to safeguard that livelihoods are protected in worst-case scenarios. This study was a participatory action research (PAR), engaging Japanese school children (aged 11 and 12) in Wakayama prefecture, Japan, in a series of interactive workshops focused on discussing flood evacuation and facilitating exposure to Agent-Based Modelling (ABM), which has potential to facilitate disaster preparedness learning in this context. As current flood evacuation is predominantly informed by topographic and demographic data, there is an exclusion of key impacting variables like social data (e.g. evacuation start times, etc.), and this research sought to include these. Through homework exercises issued to school children, social datasets were collected and included within a computational model of flood evacuation, creating an enhanced ABM-approach. Results illustrated that when comparing the enhanced model to an initial model that did not include social datasets, the addition offers more detailed and accurate insights into flood evacuation behaviour. Also, feedback from the school children that followed the workshops further established that engagement through the use of ABM raised awareness and interest towards their flood evacuation, which is essential to successful DRR. These findings suggest that consideration of variables beyond topography and demographics needs to be taken into account within future ABM in this context, and taking a participatory approach in ABM can have benefits to engage and educate samples affected by disasters. The study will need to be expanded to include the same approach within schools beyond Japan, and include other stakeholders where flooding is an increasing issue, and enlarge social variables used to ensure greater robustness in the modelling.

Article number 105180.