A Bayesian network analysis on the interactions of flood resilience measures in the Vietnamese Mekong Delta

Abstract The Vietnamese Mekong Delta (VMD) is a critical agricultural region but is highly vulnerable to the adverse impacts of climate change, particularly flooding. Nevertheless, flood resilience in the VMD has improved through engineering and adaptive practices initiated by local governments and farmers. However, the understanding of the internal interactions among flood resilience measures remains limited. Cho Moi district, An Giang province, VMD, has experienced flooding but shown resilience since the implementation of the South Vam Nao project, making it a representative case for flood resilience investigation. This study employs qualitative methods and Bayesian network (BN) analysis to examine the interrelationships among flood resilience measures in Cho Moi district to address the research gap. The results indicate that the project has facilitated coordination between the local governments and citizens, enhancing farmers’ capacity to reduce flood impacts. A BN inference test, assuming that all measures perform at 100% ‘High’ efficiency, illustrates that the current system configuration performs strongly overall. ‘High-dike construction’ is also highlighted as the central measure facilitating the transformation of agricultural practices. Yet, without the presence of high dikes, farmers remain the most critical factor in sustaining resilience, regardless of the adequacy of government-led measures. The study recommends redesigning the height of the high-dike system to mitigate landslide risk and address other environmental problems, such as nutrient-rich alluvium loss and soil compaction, while ensuring adequate flood protection. The study introduces an innovative framework for flood resilience assessment, helping governments understand their water management systems and identify necessary improvements to enhance flood resilience.