 |
| Titel |
Stepwise calibration procedure for regional coupled hydrological-hydrogeological models |
| VerfasserIn |
Baptiste Labarthe, Lena Abasq, Chantal de Fouquet, Nicolas Flipo |
| Konferenz |
EGU General Assembly 2014
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 16 (2014) |
| Datensatznummer |
250096178
|
| Publikation (Nr.) |
 EGU/EGU2014-11672.pdf |
|
|
|
|
|
| Zusammenfassung |
| Stream-aquifer interaction is a complex process depending on regional and local processes.
Indeed, the groundwater component of hydrosystem and large scale heterogeneities control
the regional flows towards the alluvial plains and the rivers. In second instance, the local
distribution of the stream bed permeabilities controls the dynamics of stream-aquifer water
fluxes within the alluvial plain, and therefore the near-river piezometric head distribution. In
order to better understand the water circulation and pollutant transport in watersheds,
the integration of these multi-dimensional processes in modelling platform has
to be performed. Thus, the nested interfaces concept in continental hydrosystem
modelling (where regional fluxes, simulated by large scale models, are imposed at local
stream-aquifer interfaces) has been presented in Flipo et al (2014). This concept has been
implemented in EauDyssée modelling platform for a large alluvial plain model
(900km2) part of a 11000km2 multi-layer aquifer system, located in the Seine basin
(France).
The hydrosystem modelling platform is composed of four spatially distributed modules
(Surface, Sub-surface, River and Groundwater), corresponding to four components of the
terrestrial water cycle. Considering the large number of parameters to be inferred
simultaneously, the calibration process of coupled models is highly computationally
demanding and therefore hardly applicable to a real case study of 10000km2. In order to
improve the efficiency of the calibration process, a stepwise calibration procedure is
proposed.
The stepwise methodology involves determining optimal parameters of all components of
the coupled model, to provide a near optimum prior information for the global calibration. It
starts with the surface component parameters calibration. The surface parameters are
optimised based on the comparison between simulated and observed discharges (or
filtered discharges) at various locations. Once the surface parameters have been
determined, the groundwater component is calibrated. The calibration procedure is
performed under steady state hypothesis (to minimize the procedure time length)
using recharge rates given by the surface component calibration and imposed fluxes
boundary conditions given by the regional model. The calibration is performed using
pilot point where the prior variogram is calculated from observed transmissivities
values.
This procedure uses PEST (http//:www.pesthomepage.org/Home.php) as the inverse
modelling tool and EauDyssée as the direct model. During the stepwise calibration process,
each modules, even if they are actually dependant from each other, are run and calibrated
independently, therefore contributions between each module have to be determined. For the
surface module, groundwater and runoff contributions have been determined by
hydrograph separation. Among the automated base-flow separation methods, the
one-parameter Chapman filter (Chapman et al 1999) has been chosen. This filter
is a decomposition of the actual base-flow between the previous base-flow and
the discharge gradient weighted by functions of the recession coefficient. For the
groundwater module, the recharge has been determined from surface and sub-surface
module.
References :
Flipo, N., A. Mourhi, B. Labarthe, and S. Biancamaria (2014). Continental hydrosystem
modelling : the concept of nested stream-aquifer interfaces. Hydrol. Earth Syst. Sci. Discuss.
11, 451-500.
Chapman,TG. (1999). A comparison of algorithms for stream flow recession and base-flow
separation. hydrological Processes 13, 701-714. |
|
|
|
|
|
|