Overall Conclusion and Discussion - Response of Flood Events to Land Use and Climate Change - page 221

mainly in the upstream where a small part of the channel was missing. The
results obtained from (e) Topo to Raster, (f) Kriging, Gaussian and (i) Kriging,
Spherical are considered to be in better agreement with the observed drainage
network. By focusing on details and the interpretation of results, it is concluded
that the drainage networks generated by the Kriging, Spherical allowed a much
more accurate delineation of the stream network. Focusing more on details, as
would be expected, the Spherical was the best overall modeled drainage net-
work, and it was comparatively more accurate than the river from the DEMs
generated with other interpolators, even in the floodplain which is a flat area.
Objective (b):
Within hydrodynamic modeling, the output is affected to a large extent by
model input parameters such as cross-section, Manning's value, etc. Therefore,
before starting the hydrodynamic simulation, it was necessary to prepare all input
data as follows:
• First, the Ubaye River was divided into three parts: (1) the upstream or chan-
nelized part of the river in the city area; (2) the Middle part of the river in the
natural environment when the river morphology starts to widen; and (3) the
downstream part of the river in the natural environment when the river starts to
narrow again. Cross-sections should be added at locations where changes occur
in slope, width, and roughness. Therefore, to select the cross-section's location
with respect to all the above mentioned parameters, 10 cross-sections were
applied along the river and the data was manually imported to the SOBEK (1D-
2D) model. The interval between cross-sections was selected at 150 m, with a
total river section of 10 km in length in the HEC-RAS (1D) model. All the
cross-sections were created based on a 5 m resolution DEM using the HEC-
GeoRAS model. In regard to the modeling representing any prefect results, it
should be noted that these cross-sections also needed to be manually modified
before being used in flood simulations (SOBEK and HEC-RAS).
• While using Manning's equation, the selection of a suitable value of n is the most
important single parameter for the proper estimation of velocity in an open
channel. To be sure about the correct selected roughness value for the channel, the
model was run for different values from 0.02 to 0.07. The result of hydraulic depth
was compared with observed data at the Abattoir Bridge. The testing results show
that the friction value of 0.05 in the channel represents the best agreement between
modeled and observed water height measured close to the Abattoir station.
• Due to a lack of available recorded data over the case study area, calibration
analysis was limited to the observed height of water levels found in the literature
and field reports of the 2008 flood. The adjustment was performed using different
friction values in the channel. Considering the lack of data regarding the river
morphology, the model was considered to have been a successful calibration.