 |
| Titel |
Comparison of 2D numerical models for river flood hazard assessment: simulation of the Secchia River flood in January, 2014 |
| VerfasserIn |
Iuliia Shustikova, Alessio Domeneghetti, Jeffrey Neal, Paul Bates, Attilio Castellarin |
| Konferenz |
EGU General Assembly 2017
|
| Medientyp |
Artikel
|
| Sprache |
en
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 19 (2017) |
| Datensatznummer |
250138295
|
| Publikation (Nr.) |
 EGU/EGU2017-1271.pdf |
|
|
|
|
|
| Zusammenfassung |
| Hydrodynamic modeling of inundation events still brings a large array of uncertainties. This
effect is especially evident in the models run for geographically large areas. Recent studies
suggest using fully two-dimensional (2D) models with high resolution in order to avoid
uncertainties and limitations coming from the incorrect interpretation of flood dynamics and
an unrealistic reproduction of the terrain topography. This, however, affects the computational
efficiency increasing the running time and hardware demands. Concerning this point, our
study evaluates and compares numerical models of different complexity by testing them on a
flood event that occurred in the basin of the Secchia River, Northern Italy, on 19th January,
2014. The event was characterized by a levee breach and consequent flooding of over 75 km2
of the plain behind the dike within 48 hours causing population displacement, one death
and economic losses in excess of 400 million Euro. We test the well-established
TELEMAC 2D, and LISFLOOD-FP codes, together with the recently launched
HEC-RAS 5.0.3 (2D model), all models are implemented using different grid size (2-200
m) based on the 1 m digital elevation model resolution. TELEMAC is a fully 2D
hydrodynamic model which is based on the finite-element or finite-volume approach.
Whereas HEC-RAS 5.0.3 and LISFLOOD-FP are both coupled 1D-2D models. All
models are calibrated against observed inundation extent and maximum water depths,
which are retrieved from remotely sensed data and field survey reports. Our study
quantitatively compares the three modeling strategies highlighting differences in terms
of the ease of implementation, accuracy of representation of hydraulic processes
within floodplains and computational efficiency. Additionally, we look into the
different grid resolutions in terms of the results accuracy and computation time.
Our study is a preliminary assessment that focuses on smaller areas in order to
identify potential modeling schemes that would be efficient for simulating flooding
scenarios for large and very large floodplains. This research aims at contributing
to the reduction of uncertainties and limitations in hazard and risk assessment. |
|
|
|
|
|
|