 |
| Titel |
Regional stochastic estimation of the groundwater catchment for distributed hydrological modelling |
| VerfasserIn |
Th. Wöhling, L. Samaniego, B. Selle, R. Kumar, M. Zink |
| Konferenz |
EGU General Assembly 2012
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 14 (2012) |
| Datensatznummer |
250065303
|
|
|
|
|
|
| Zusammenfassung |
| Rainfall-runoff modeling typically assumes that the groundwater catchment boundary
coincide with the topographic one. While this is often a reasonable assumption for large and
and mesoscale catchments (> 103 km2), this assumption may lead to large errors of
streamflow in small scale catchments (-¤ 102 km2), in particular in certain geological
settings. The Ammer catchment (135 km2) in the upper Neckar river basin (Germany) is a
prime example where groundwater and topographic catchment boundaries are significantly
distinct from each other. The catchment is characterized by a complex sequence of fractured,
karstic Triassic rock formations. These strata gently dip into ESE direction governing
groundwater flow. Analysis of tracer experiments conducted in the 1970s indicates that the
boundary overlap could be less than 80 percent. Further, a modelling study of the upper
Neckar river basin using the distributed hydrological model mHM showed Nash-Sutcliff
efficiencies (NSE) < 0.4 for simulated runoff in the Ammer sub-basin whereas higher
efficiencies (NSE ~ 0.7) were obtained for most of the other 21 sub basins in the region.
In this study we present a methodology to simultaneously estimate the regional
groundwater catchment boundaries of the Ammer and its surrounding basins. In a
first step we derive the best possible fit between mHM simulated and observed
runoff for the individual sub-basins in the Ammer region and determine the trade-off
between the fits of the individual basins using the muliobjective optimization method
AMALGAM. We further present a strategy to estimate the regional groundwater
catchment boundaries with the aim to improve runoff predictions in the Ammer
catchment while not deteriorating runoff predictions in the surrounding basins. Our
strategy involves a modification of the mHM model to account for ground water
import/export from neighboring catchments while maintaining full mass balance of
the surrounding basins. Groundwater catchment boundaries are then obtained by
innovative stochastic optimization techniques based on Simulated Annealing that are
constrained by expert knowledge about the hydrological system, e.g. a minimum
overlap of groundwater and topographical catchment boundaries. The methods
developed herein are useful for both plausibility and hypothesis testing as well as
hydrological modelling of small scale catchments where conventional models fail due
to the mismatch between groundwater and topographical catchment boundaries. |
|
|
|
|
|
|