Development Of Distributed Grid-Based Hydrological Model And Floodplain Inundation Management System

A physical based, distributed hydrological model was developed to route overland flows during isolated HISD storms. The model has operated on a grid or cell basis and routed the excess rainfall over the grids, conforming to the DEM-derived drainage paths, to the basin outlet. The rainfall-runoff hydrological modelling was implemented in MATLAB 7.0. The system has integrated GIS, RS, DEM, data management capability and a dynamic basin model within a common Windows environment. The simulation algorithms of the rainfall-runoff model have operated on grid bases compatible with the MATLAB programming language, which has been used to write instructions to many grid-based operations. Due to the MATLAB architecture, the system has been proven successful for large-scale basin modeling, which requires high level resolution, record keeping and technical transfer. The model has estimated the runoff using the Soil Conservation Service-Curve Numbers (SCS-CN), determined by the land use/ land cover and the hydrological soil group found in each grid. The overland flow mechanics were described by the diffusion wave approximation of St Venant equations, which were numerically solved for depth of flow and runoff by the finite volume method (FVM). The grid cell physical properties such as topography, land use, soil, and Manning’s roughness’ coefficient were extracted from published maps for discretized cells of the Klang River basin(KRB) using a GIS. The land use/cover classes were derived from interpreted information of Landsat TM imagery using the combined object-oriented segmentation - fuzzy logic algorithm. The DEM of 90m resolution, used to calculate slopes that generated runoffs, was derived from radar data sets (C-band) of the Shuttle Radar Topography Mission (SRTM) using the interferometric approach. Four criteria were used for the assessment of the model performance - Model bias, Nash– Sutcliffe and model efficiencies for both low and high flows during both calibration and validation periods. The results showed the advantages of integrating RS, DEM and GIS with hydrologic simulation in generating runoff processes in the spatial domain, attaining as well fairly high precision simulation with the general hydrologic trends well captured by the model. This study has also involved the application of flood modeling, which has integrated the results of the grid-based overland flow routing model into MIKE11 onedimensional hydrodynamic model. The discharge hydrographs were extracted from the grid-based overland flow routing model in ASCII format and imported into MIKE11 hydrodynamic modeling system. The MIKE11 model was developed based on surveyed, stream cross-section data to perform hydrodynamic simulation of the flooding process. The MIKE11 modeling was applied to the Klang River system comprising 9 main tributaries. The analysis has considered the river system with and without Stormwater Management and Road Tunnel (SMART) project, which involve structural flood mitigations measures including retention ponds, bypass tunnel and flow diversions, where the river physical condition was modified accordingly. Hourly data for flow were created into compatible MIKE11 time series in a separate file as input to the parameter editors. Initial and boundary conditions were based on the inputs for MIKE11 operational analysis. It has been found that the modeled predictions of depth and discharge matched observed data. A good agreement between the simulated and observed data was achieved for rating curves with RMSE = 0.96, 0.94, 0.95, and 0.97 at respective calibration points. From the results revealed by the MIKE11 modeling simulation, there were evidences that SMART was useful for flood mitigation of Klang River Basin. For instance at Tun Perak Bridge, the normal level for the Klang River was 25m, the alert level was 28m and the danger level was 29.5m. The value from the simulation showed that the maximum water level without SMART was 32m. However this level with SMART was only 27.8m which did not exceed the alert and danger level at Tun Perak Bridge. This area is the most critical part of KL. Once the water level from the Klang River exceeds the flood wall, the whole KL will be badly flooded. Finally, the results of the runoff modeling were integrated in MIKE-GIS model for flood inundation mapping. A digital planimetric view and topographic mapping of the floodplain was developed using the three-dimensional floodplain visualization approach through the integration of a digital terrain model. This model was synthesized from MIKE11 stream cross-sectional coordinate into a digital surface model, generated from aerial stereo pair photos using Ortho Engine PCI image processing software. The resulting formulated surface model provided a good representation of the general landscape and contained additional details within the stream channel. Integration of 3D-GIS and spatial analytical techniques together with hydrologic and hydraulic modeling processes has enhanced the visualization and display techniques for visual presentation and generation of flood inundation maps for early warning and contingency planning.

Preview PDF FK_2008_27.pdf Download (600kB)

Actions (login required)

View Item