Search

In Canada, flood analysis and water resource management, in general, are tasks conducted at the provincial level; therefore, unified national-scale approaches to water-related problems are uncommon. In this study, a national-scale flood risk assessment approach is proposed and developed. The study focuses on using global and national datasets available at reasonably fine resolutions to create flood risk maps. First, a flood hazard map of Canada is developed using topography-based parameters derived from digital elevation models namely Elevation Above Nearest Drainage (EAND) and Distance From Nearest Drainage (DFND). This flood hazard mapping method is tested on a smaller area around the city of Calgary, Alberta, against a flood inundation map produced by the City using hydraulic modeling. Second, a flood exposure map of Canada is developed using a land-use map and the satellite-based nightlight luminosity data as two exposure parameters. Third, an economic flood risk map is produced, and subsequently overlaid with population density information to produce a socioeconomic flood risk map for Canada. All three maps of hazard, exposure, and risk are classified into five classes, ranging from very low to severe. A simple way to include flood protection measures in hazard estimation is also demonstrated using the example of the city of Winnipeg, Manitoba. This could be done for the entire country if information on flood protection across Canada were available. The evaluation of the flood hazard map shows that the topography-based method adopted in this study is both practical and reliable for large-scale analysis. Sensitivity analysis regarding the resolution of the digital elevation model is needed to identify the resolution that is fine enough for reliable hazard mapping, but coarse enough for computational tractability. The nightlight data are found to be useful for exposure and risk mapping in Canada; however, uncertainty analysis should be conducted to investigate the effect of the overglow phenomenon on flood risk mapping. [ABSTRACT FROM AUTHOR]

Worldwide, floods are the most frequently observed natural catastrophes that lead to considerable damages to both life and property. Financial losses due to the damages caused by natural catastrophes are in most places largely uninsured, in both developing and developed countries. Determining losses due to different flood types and/or even other perils can help to evaluate the cost of mitigation as well as quantify the structure and the cost of insurance and reinsurance protection. In general, the concept of catastrophe models consisting of hazard, exposure, and vulnerability components is of prime importance to estimate projected losses. In the present study, a probabilistic approach to flood loss estimation is presented from an insurance perspective and tested on a river basin prone to flooding in South India. Hazard is estimated by determining the flood extent and inundation depths for different return periods along a river network using a 2D hydrodynamic model – TUFLOW (BMT WBM). The exposure reflects the financial liability within the area of interest such as industrial parks, warehouse, commercial districts, public infrastructure etc. The vulnerability is determined by developing a relationship between property value and flood levels. To arrive at probabilistic flood hazard, a large number of flood events are generated using advanced statistical processing based on copulas. A vine copula approach using block maxima of streamflow values is used to generate synthetic events and subsequently; the losses corresponding to each of the generated events are estimated. The key outcome of such an approach is the potential flood loss profile expressed as an exceedance probability (EP) curve. Uncertainties associated with determining flood magnitudes and flooding extents for different return periods and their effect on the estimated loss is also evaluated. [ABSTRACT FROM AUTHOR]