Development of generalized loss functions for rapid estimation of flood damages: a case study in Kelani River basin, Sri Lanka

Abstract

Assessment of infrastructural vulnerability to natural hazards, and subsequent economic loss, can make important contributions to future disaster risk minimization. The recent endeavor is to ascertain and evaluate risk globally, which can provide a framework to identify unique regional vulnerabilities, the mobilization of international investments, and cross-country risk comparison. This would require a concerted effort for the detailed classification of building exposures and vulnerability models. This study presents the design and efficacy of flood-vulnerability models for structural building types. The study uses an empirical approach, with data gathered from survey questionnaire, for direct estimation of flood damages in the Kelani River basin in Sri Lanka. Survey questionnaires were administered in the flood-prone areas of the basin, and depth-damage functions were established for four (4) structural building types that were identified based on the relationship between inundation depths and flood damage ratio. Event-based flood hazards were simulated using the Flo-2D model. Building exposures and densities were derived from remote sensing data, using integrated thematic land cover feature indices and supervised image classification. A modified mathematical loss model was employed to simulate flood damages to each building category for a disastrous flood event in the Kelani River basin. Simulated damages and post-flood survey showed reasonable comparativeness. The models can be employed for loss estimation of future damages and risk-reduction planning for flood disaster in Sri Lanka.