Assessment of Urban Flood Susceptibility and Inundation through Bivariate Statistics with Synthetic Aperture Radar: Insights for Spatial Planning in Pekanbaru City, Indonesia

Abstract

Flooding has become one of the most dangerous hydrometeorological disasters, affecting the sustainability of cities in the future. This study aims to assess flood susceptibility using a frequency ratio approach and evaluate the spatial planning in Pekanbaru City, Indonesia. Flood locations were derived from synthetic aperture radar data to prepare actual flood data. In this area, the identification of physical environmental parameters was conducted using various datasets such as slope, landform, curvature, topographic wetness index, distance from rivers, rainfall, soil texture, and soil depth. Furthermore, a weighted assessment of all thematic layers was calculated based on the frequency ratio of actual flood events in the observation area. The overall data related to the location of the events were divided, with 70% for model development and 30% for validation. The results showed that actual flood events affected 18 km², with an accuracy of 84.21%. Flood susceptibility was categorized into four levels such as very high (11.36%), high (58.04%), medium (24.78%), and low (5.81%). An accurate assessment of potential flood susceptibility, measured by the operational characteristic curve (AUC), showed a prediction rate of 76.30% and a success rate of 78.45%. However, considering the implications for spatial patterns, flood susceptibility affects cultivated areas covering 381.16 km², which are spread almost throughout the urban area. High to very high susceptibility areas indirectly cause disaster losses and impact community activities. The misalignment between the distribution of flood susceptibility and spatial planning in urban areas needs to be addressed to anticipate other potential hazards.