Evaluation of flood forecasting-response systems II

Abstract

system model and computational methodology have been developed which evaluate the worth of flood forecast - response systems in reducing the economic damage caused by floods. The efficiencies of the forecast system, the response system, and the overall system may be individually obtained and compared. In this report the case study of Milton, Pennsylvania, was extended and further case studies were performed including a large residential section of Victoria, Texas, and all the residences in Columbus, Mississippi. These locations show better forecast and response efficiencies than obtained for Milton, Pennsylvania. The difference is attributed to longer forecast lead times at Columbus and Victoria. Sensitivity analyses were run at all three locations. These show the effects of many system factors, such as the time required to produce, disseminate and respond to a forecast, on the efficiency of the system. The forecast efficiency improves significantly as these times are reduced. Further analysis of the response system based on human factors involved has led to the development of a simulation model of the process by which the floodplain dweller determines the appropriate response to a flood warning. Investigation of ways to extend the methodology to evaluate regions lacking the detailed data used for the case studies has indicated more problems than answers. Extrapolation based on overall system efficiency related to published regional and national flood damage estimates was used to provide an approximate value of the flood forecast - response system for two regions and for the nation.A listing of simplicities and approximations which make computations tractable but which may affect accuracy is given. Finally, an evaluation of the work accomplished for this project and suggestions for the constructive use of the flood forecast -response system model and computational procedures is given.