 |
| Titel |
A multi-criteria parameterisation strategy for the hydrological modelling of storm events in an agricultural catchment |
| VerfasserIn |
D. W. Hallema, R. Moussa |
| Konferenz |
EGU General Assembly 2009
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 11 (2009) |
| Datensatznummer |
250019561
|
|
|
|
|
|
| Zusammenfassung |
| The parameterisation of physically based hydrological models is a key factor in evaluating
their performance and improving their adequacy. In comparison to natural catchments, the
parameterisation of agricultural catchment models requires a special approach since
agricultural practice, such as tillage, introduces a temporal variability in the hydraulic
conductivity of the soil. This study aims to develop and evaluate a parameterisation strategy
for the hydrological modelling of storm events in a Mediterranean agricultural
catchment. The adopted methodology accounts for a spatio-temporally variable
land use, combined with the presence of an artificial drainage network and the
occurrence of short duration-high intensity storm events. A parameterisation of 43
events covering a 12-year period was carried out applying MHYDAS (Distributed
HYdrological Modelling for AgroSystems) to the Roujan catchment (0.91 km2) in southern
France.
Parameterisation of MHYDAS is particularly difficult considering the large number of
spatio-temporally variable parameters involved. To incorporate this variability, the catchment
was divided into hydrological units that are considered hydrologically homogeneous, mostly
parcels separated by field boundaries and ditches. The parameterisation was performed in
four parts. Firstly, the appropriate flood routing equation was selected for each channel reach
in the drainage network: kinematic wave where possible and diffuse wave for the other
reaches. Secondly, the boundary conditions (geometry, connectivity and roughness of
hydrological units and ditches) were assumed equal for all events, while soil hydrodynamic
properties and initial soil moisture content were taken variable in time. A third step
was to calibrate individual events by manually tuning the average overland- and
channel flow celerity, saturated hydraulic conductivity, and two coefficients for the
channel infiltration and exfiltration fluxes. The predictive power of each simulation
was assessed using multi-objective functions related to volume, peak flow, and the
Nash-Sutcliffe coefficient. The last step was a global calibration and validation on 33 events.
The results of the calibration of individual events have demonstrated that events
can be properly parameterised using the current strategy, assuming that channel
exfiltration does not contribute to runoff. It was concluded that the average overland-
and channel flow celerity, and the channel infiltration coefficient have remained
relatively constant. This study has demonstrated that extensive data collection in
combination with a multi-criteria parameterisation strategy is an important tool
for calibrating and evaluating hydrological models for agricultural environments. |
|
|
|
|
|
|