Abstract:
Spatially explicit hydrodynamic flood models can play an important role in natural hazard risk reduction. A key element of these models that make them suitable for risk reduction is the ability to provide inundation information of a hazard event. Such information can be critical for landuse planning, for mapping evacuation routes, and for locating suitable emergency.
The most important part of flood risk identification and management is the flood-prone area (extent) delineation. Flood-prone areas are those areas subject to inundation as a result of flooding with certain frequency. The determination of flood prone area requires considerable collation of historical data, accurate digital elevation data, discharge data and number of cross-
sections located throughout the watershed. A region-specific probabilistic model is developed combining information on (1) the probability of flooding, (2) the spatial extent of floods for different severity levels (or return periods), and (3) the consequences of these floods (e.g. damage assessment). Probabilistic flood events, is done by modeling events featuring the same statistical characteristics as the catalogue of historical flood events, but projected to span a period of several thousand years. It enables insurers to identify the flood risk of individual locations in detail and to substantially improve underwriting for flood risk.
A flood risk model was developed for the Romania for damage assessment due to floods. A regional specific probabilistic hazard model has been developed at commune level resolution. The model was GIS based, making use of a SRTM —90m horizontal resolution DTM and a network of the major rivers in the country, commune level residential exposure as of year 2006. The building stock is estimated in terms of material, age and height using data from the National Statistical Institute. A review of historical flooding was undertaken to define the worst and most widespread flood events. A number of synthetic events have been generated each providing peak flows at gauging stations throughout the country. The flood flows are propagated using one-dimensional hydraulic simulation of flood wave and then converted into flood levels using elevation data sets from DEM. The flood depths are then combined with maps of the communes to define an average flood depth per commune. Multi-variate Generalized extreme value distribution has been used in generating synthetic flows at the gauge station preserving the inter site correlation among the gauges. The flows were routed using of one-dimensional analysis using Hec-RAS. The extent of and depth of flooding were determined in GIS environment using ESRI's ArcView. The flood depths were combined with maps of the communes to define an average flood depth per commune. The model was calibrated using maps of the observed flood extents from 2005 and 2006. This study will help it formulatingthe catastrophe insurance schemes as a part of World Bank Hazard Risk Mitigation and Emergency Preparedness Project.
