Publication: Optimization of flood inundation simulation using multi-resolution digital elevation model
Date
2020-05-01
Authors
Kadir, Muhammad Azraie Abdul
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Several studies have been made to improve the flood inundation simulation
derived from coarser Digital Elevation Model (DEM) such as using modification
spatial distribution of DEM and modification of the river cross section. All models
focused on the DEM. Up until now, optimization studies of DEM resolution with
flood inundation distribution relationship is scarce. This is one of the key
components of flood phenomenon which requires high accuracy for further action in
the flood risk management. Additionally, inflow value could also be used to optimize
flood inundation simulation without disturbing the coarse DEM. In this study, the
relationship between DEM resolution and flood inundation prediction was
determined in an attempt to propose a new optimization method of flood inundation
prediction. This was established by incorporating the inflow value into the
relationship between DEM resolution and flood inundation. A hydraulic model of
Sungai Bertam, Cameron Highland was developed at a scale of 1:25 to conduct
experiments at different inflow value. This result was used to validate the 2D
numerical simulation of HEC-RAS model. Then, the high resolution DEM was
resampled into several coarser DEM to re-simulate the flood inundation. Both DEM
and flood inundation results (Water Surface Elevation (WSE), flood extent, and flood
area) were used to determine the effect of inflow parameter on the optimization of
flood inundation simulation. Additionally, the optimization was also validated
experimentally using the hydraulic model. Hydraulic model experiment indicated
that the geometry of the channel (shape and slope) plays an important role in flood
inundation and inflow of 64 l/s provided a significant flood inundation at the
downstream of the river. A good agreement between hydraulic and HEC-RAS model
was found except at low inflow (25 m3/s and 50 m3/s), velocity (maximum Mean
Relative Absolute Error (MRAE) = 0.408) and flow travel time (R2 = 0.595). This is
due to the limitation of diffusive wave algorithm, 2-dimensional model and effect of
grid configuration. Relationship between DEM resolution and WSE gave the best R2
with minimum R2 = 0.9845 at the inflow of 150 m3/s. Increasing DEM resolution
also increased the Mean Absolute Error (MAE) and MRAE index. Optimization
using the relationship between DEM resolution and WSE together with several
inflow values gives significant improvement up to 72 % of MRAE index for flood
extent and 131.53 % of MRAE index for flood area. Validation using the hydraulic
model experiment at an inflow of 200 m3/s indicated that the optimization improves
the MRAE index of WSE by 23%. The study concluded that flood inundation
simulation can be improved by modification of the inflow parameter, as well as the
correlation between DEM resolution and flood inundation prediction.