FLOOD INUNDATION MAPPING - A CASE STUDY

Abstract

A flood inundation map project needs geographic data from traditional aerial mapping for terrain modeling and map presentation and cross-section data for hydraulic calculations. The hydrologic analysis is needed for determination of discharges or water levels that correspond to the final flood inundation maps. Finally, hydraulic analysis facilitates computation of water surface profiles for flood discharges of different return periods. The water surface profiles are used for depiction of flood inundation maps. The objective of this dissertation work is to present a case study of Godavari River in the 143 km upstream reach of Polavaram dam site for the preparation of flood inundation map. The popular one-dimensional hydraulic model named HEC-RAS model is used for hydraulic analysis. The input data requires sufficient number of cross-sections wide enough and spaced closely to represent the floodplain and these are derived from SRTM 90M digital elevation model data through Remote Sensing and GIS. Steady flow data constitute 50,100 and 500 yrs. return period floods from frequency analysis. Inundated areas are determined using Geographical Information System (GIS). HEC-RAS is an integrated system of software designed for interactive use in a multi-tasking environment. HEC-RAS has the ability to import three dimensional (3D) river schematic and cross section data created in a GIS. While the HEC-RAS software only utilizes two-dimensional data during the computation; the three dimensional information is used in the program for display purpose only. After completion of hydraulic analysis, the computed water surface profiles can be exported back to the GIS system for development and display of a flood inundation mapping. A triangular irregular network (TIN) is generated from the available elevation data. The terrain TIN is then transformed into a regular grid with grid size 15-20 meters. In the same way, a TIN is generated between the cross sections. The cross sections have water elevations that represent 50, 100 and 500 years return periods. Different TINs are then transformed into water surface grid with same definition as the terrain grid. The ►1 inundated areas are identified by subtracting the land surface grid from the water surface grid, resulting in positive values in inundated areas. The final product is smoothed polygons representing inundated areas with a specified return period. Flood inundation mapping is an important component of the non-structural measures. The results of the study suggest that cross-sections generated using Remote Sensing and GIS data, coupled with little manipulations, are in close agreement with the cross-sections observed at limited number of sites in the considered study reach of Godavari basin. The flood values derived from frequency analysis, using Gumbel distribution, for 50, 100 and 500 years return periods comes out to be 97604, 98948 and 96259 cumec, respectively. The inundated areas computed using HEC-RAS for these return period floods are of the order of 10.8, 11.65 13.69 sq. km., respectively. A sensitivity analysis of FRL reveals the existence of a power relationship between FRL and the corresponding inundated area.