MAPPING AREAS EXPOSED TO EROSION AND WATER FORCES DURING
EXTREME FLOODS IN STEEP TERRAINMICHAL PAVLÍČEK, ODDBJØRN BRULAND
Department of Civil and Environmental Engineering, Norwegian University of Science and Technology, Trondheim, Norway.
Introduction Results
Discussion and conclusions
References
Methods
The extent and potential consequences of floods in large water courses are mostly well
mapped in Norway (NVE 2018). Here, the flood risk is related to inundation which has
been mapped using the hydraulic routing of the floods of different return periods. The
risks in small and steep catchments are not that well mapped. The faster response and
the forces due to high water velocities induce another risk dimension that is
significantly more challenging to handle.
The main goal of this study is to examine the critical points in the river channel during
flash floods in steep rivers. In this study, critical points are spots where water could find
its way out from the river channel, which could be caused by erosion and deposition of
sediments.
Fig. 1. Aerial view of Utvik flood event from the fjord (VG 2017), with original river
thalweg (blue line) and bridges.
TELEMAC – MASCARET was selected as a software package to carry out simulations
of flash floods in steep rivers. Telemac’s 2D module was used to run 2D hydrodynamic
simulations (i.e. rigid terrain) in the horizontal plane. The code solves Saint-Venant
(shallow water) equations in non-conservative form (Hervouet 2007). Regarding the
morphodynamic simulations (i.e. including sediment transport and river’s bed
evolution), Sisyphe module was used. the module solves river’s bed evolution with the
sediment mass conservation equation (Exner equation) (Tassi 2017).Fig. 5. Results of hydrodynamic (left) and final morphodynamic (right) simulations. Displayed variable is maximum water
depth; purple lines represent flow paths during flood event.
Fig. 6. Results of hydrodynamic simulation (left, maximum shear stress), final morphodynamic simulation (center, bed
evolution in the end of simulation) and bed evolution measured after real flood event (right, bed evolution (Kartverket
2018)); purple lines represent flow paths during flood event.
Fig. 7. Results of hydrodynamic simulation. Maximum water depth (left); maximum shear stress
(right); purple lines represent flow paths during flood event.
Hydrodynamic simulation
The simulation described next is first applied to Utvik and the same parameters and
set-up were used in Innvik afterwards.
Fig. 2. Flow hydrograph of Utvik‘s flood on July
24, 2017 (Bruland 2018).
By discretization, the domain was divided into
triangular mesh elements: size of 20 m in the
fjord, 2 m in the river channel and of 3 m in the
rest of the domain.
A flow hydrograph (Fig. 2) of Utvik’s flood event
was assigned as the inflow open boundary
condition. Constant water surface elevation 0 m
a.s.l. was attached to the outflow open boundary
condition.
Different values of Manning roughness coefficient (n) were assigned: 0.045 for the river
channel and the fjord bed, 0.025 for the roads, and 0.100 for the rest of the domain
(i.e. possible inundation area).
Finite element method was used for resolution of Saint-Venant equations. Constant
viscosity turbulent model was assumed.
0
50
100
150
200
0 2 4 6 8 10 12 14 16 18 20 22 0
Q [
m3
/s]
t [h]
Bruland, Oddbjørn. 2018. Extreme flood in small steep cathcment case Utvik. In XXX Nordic
Hydrological Conference. Bergen, Norway.
Hervouet, Jean-Michel. 2007. Hydrodynamics of Free Surface Flows: Modelling with the Finite
Element Method ( John Wiley & Sons: UK).
Kartverket. 2018. Høydedata, Accessed 13/12/2018. https://hoydedata.no/LaserInnsyn/.
NVE. 2018. Kartlegging, Accessed 13/12/18. https://www.nve.no/flaum-og-skred/kartlegging/.
VG. 2017. Per Inge Verlos (58) hus sto midt i flommen: – Helt ufattelig Accessed 13/12/2018.
https://www.vg.no/nyheter/innenriks/i/GVbeQ/per-inge-verlos-58-hus-sto-midt-i-flommen-helt-ufattelig
Tassi, Pablo. 2017. Sisyphe User Manual Version 7.2.
The results presented in this poster showed that the hydrodynamic simulation could be
used to determine the capacity of the river channel and finding the critical points due to
the shear stress. However, it does not take into account the creation of the new flow
paths due to erosion and deposition processes.
As for the morphodynamic simulation, as can be seen in Fig. 5, the inundation area in
both types of simulations is not much different. The results of the current
morphodynamic simulation present a non-erodible bed, thus, a negligible volume of
transported sediment in the river channel. Therefore, the set-up of the simulation
should be tested and improved on the simpler cases with steep slope, covering an
erodible bed in the river channel.
Erosion and sedimentation processes are important to determine the critical points
during flash floods, hence, the bed load transport in steep terrains needs to be
investigated further.
Further research will focus on numerical and physical modeling of the phenomena.
Numerical simulations of lab experiments and cases with field measurements will be
carried out to calibrate and validate the models. More numerical models will be used
(TELEMAC-MASCARET, REEF3D, HEC-RAS).
InnvikHydrodynamic
simulation was
carried out in
Innvik so far.
As it can be seen
in Fig. 7, the river
channel in Innvik
is straighter than
Utvik’s and there
is no clear critical
point as in Utvik.
UtvikDue to the instabilities, non-erodible bed in the river channel was assumed for the final
simulation set-up. Sediment diameter in the inundation area was set up 0.5 mm.
Meyer-Peter and Müller formula was used for bed load transport and no suspension
load was simulated. It was assumed an active layer thickness of sediment 2.5 m. The
results of the final morphodynamic simulation are presented in Fig. 4 (right) and Fig. 6
(center).
In the hydrodynamic simulation (Fig. 5, left), the main flow path is located in the river
channel and the other paths matches quite well with flow paths from the real flood.
As for the final morphodynamic simulation, results show that the water found new flow
paths on the left side of the bridge 02 (Fig. 5, right). But the inundation area is similar
as in the hydrodynamic simulation.
As it can be seen in Fig. 6, the results of shear stress from hydrodynamic simulation
are suitable to find the critical point besides bridge 02.The flooding of Utvik
(western Norway, Fig. 1)
in July 2017 is
reconstructed using
TELEMAC–MASCARET
numerical simulating
software. The
methodology is also
tested and compared in
the neighboring river in
Innvik. Coarse
sediment, boulders and
rocks are located in the
river bed and the slope
of the river channel in
Utvik and Innvik is steep
(ca. 3-17%).
In the morphodynamic
simulation, instabilities
in the bed evolution
were observed in the
reaches with the
steepest slope and
steep river banks. The
examples of the
instabilities in Utvik can
be seen in Fig. 3 and
4. The instabilities in
the river channel were
observed in all
simulations with
different set-ups.
Fig. 3. Longitudinal profile of the river channel in Utvik. Black line represents the original
river bed, red and green lines represent river bed obtained from morphodynamic
simulations with different sediment diameters.
Fig. 4. Cross-section of the river channel in Utvik in the reach with slope 17 %. Black
line represents the original river bed, red and green lines represent river bed obtained
from morphodynamic simulations with different sediment diameters.
0
20
40
60
80
200 250 300 350 400 450 500 550 600 650 700 750 800
Elev
atio
n [
m a
.s.l.
]
Distance [m] original river bed d50 = 0.01 m d50 = 0.001 - 0.1 m
17 %
3 %
12
17
22
27
32
40 50 60 70 80 90 100 110 120 130 140
Elev
atio
n [
m a
.s.l.
]
Distance [m] original river bed d50 = 0.01 m d50 = 0.001 - 0.1 m
Morphodynamic simulation
Several combinations of sediment parameters, bed load transport formulas, mesh size
and numerical set-up were tested to find out the best match with the flow paths
observed during the flood event (Fig. 1).