Spatial and Temporal Prediction of Flash Floods in Ungauged Area Based on GIS

Abstract

A GIS-based method for predicting the spatial and temporal characteristics of flash floods in ungauged area is proposed in this paper. The critical rainfall method is used to carry on temporal prediction for flash floods. First in basins with raingauges, the critical rainfall is estimated by analyzing rainfall records; in ungauged basins with investigated disaster records, however, it is supposed that the occurrence frequency of natural disasters and that of rainfall are the same, and the critical rainfall is verified by surveying the occurrence frequency of disasters; then the Inverse Distance Weighted interpolation method is used to map the critical rainfall and derive certain critical values of rainfall in ungauged basins. In order to carry on spatial prediction, the modified rational formula method is adopted to calculate the design flash flood of ungauged basins first. Genetic Algorithm is used to search for the parameters (n 1,n 2,u,m) in gauged basins. Geostatistic analysis module in ArcGIS is used to map storm parameters n 1\n 2. Regression analysis method is used to get the empirical relationship between u and θ and that between m and θ. Finally in ungauged basins, by using the map to confirm n 1\n 2, by using the empirical formula to get m and u, the design flash flood can be calculated with the rational formula. With contour map and simple hydraulic method, flash flood disaster risk map can be drawn out. It has been applied to Urumchi region successfully.