Uncertainty Estimation in Flood Inundation Mapping: An Application of Non‐parametric Bootstrapping

• Published in: River research and applications Vol. 33; no. 4; pp. 611 - 619
• Main Authors: Faghih, M., Mirzaei, M., Adamowski, J., Lee, J., El‐Shafie, A.
• Format: Journal Article
• Language: English
• Published: Bognor Regis Wiley Subscription Services, Inc 01.05.2017
• Subjects: business enterprises; Companies; Computer programs; Computer simulation; computer software; Delineation; Disaster management; disaster preparedness; Emergency preparedness; Flood damage; Flood hazards; Flood management; Flood mapping; Flooding; Floods; Flow rates; generalized extreme value distribution; Geographic information systems; Geographical information systems; Hydrodynamics; hydrologic models; Information systems; infrastructure; Insurance; Land use; Malaysia; Mapping; Modelling; Motivation; Nonparametric statistics; non‐parametric bootstrap sampling; planning; Precipitation; Remote sensing; Residential areas; River basins; Rivers; Satellite navigation systems; Software; stakeholders; Stream discharge; Stream flow; topography; Uncertainty; uncertainty analysis; Water levels; water management; Water resources; Water resources management; watersheds; Malaysia
• ISSN: 1535-1459; 1535-1467
• DOI: 10.1002/rra.3108

Disaster prevention planning is affected in a significant way by a lack of in‐depth understanding of the numerous uncertainties involved with flood delineation and related estimations. Currently, flood inundation extent is represented as a deterministic map without in‐depth consideration of the inherent uncertainties associated with variables such as precipitation, streamflow, topographic representation, modelling parameters and techniques, and geospatial operations. The motivation of this study is to estimate uncertainties in flood inundation mapping based on a non‐parametric bootstrapping method. The uncertainty is addressed through the application of non‐parametric bootstrap sampling to the hydrodynamic modelling software, HEC‐RAS, integrated with Geographic Information System (GIS). This approach was used to simulate different water levels and flow rates corresponding to different return periods from the available database. The study area was the Langat River Basin in Malaysia. The results revealed that the inundated land and infrastructure are subject to a flooding hazard of high‐frequency events and that the flood damage potential is increasing significantly for residential areas and valuable land‐use classes with higher return periods. The proposed methodology, as well as the study outcomes, of this paper could be beneficial to policymakers, water resources managers, insurance companies and other flood‐related stakeholders. Copyright © 2017 John Wiley & Sons, Ltd.