Floods as the Dominant Disaster Type in Explaining Disaster Impacts: A Dominance Analysis of Affected Population in Indonesia

Abstract

Understanding the types of disasters that contribute most to social impacts is crucial for formulating effective disaster management policies. However, most previous studies assess disaster impacts using a limited regression approach to explain the relative importance of various correlated disaster types. This study aims to identify the dominance hierarchy of various disaster types in explaining variations in the number of affected populations in Indonesia. Using the 2021 Village Potential Survey (Podes) data covering 84,096 villages/sub-districts and six main disaster types: floods, landslides, droughts, winds, forest and land fires, and coastal erosion, this study applies a dominance analysis approach based on the incremental contribution to the coefficient of determination (R²). The analysis reveals a highly unequal disaster impact structure, with floods emerging as the most dominant disaster and consistently outperforming other disaster types across all model specifications. Droughts and winds follow, while landslides, forest and land fires, and coastal erosion contribute relatively little to variations in the number of affected populations. These findings emphasise the importance of a dominance-based approach in determining disaster mitigation priorities and provide a strong empirical basis for formulating more effective and evidence-based disaster management policies.