 |
| Titel |
The investigation on the effects of model parametrization on daily water flux simulations at various spatial resolutions |
| VerfasserIn |
Rohini Kumar, Luis Samaniego |
| Konferenz |
EGU General Assembly 2010
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 12 (2010) |
| Datensatznummer |
250038424
|
|
|
|
|
|
| Zusammenfassung |
| Accurate and reliable predictions of streamflow and other spatio-temporal distribution of
water fluxes and state variables such as soil moisture, evapotranspiration, among others, are
needed for efficient management of water resources at a mesoscale. Distributed hydrologic
models that provides such predictive abilities requires, among other things, a robust
parametrization method that estimates the spatial fields of model parameters. The
performance of such models greatly depends upon the parameterization method employed
and a spatial scale chosen for modeling purposes.
The objective of this study is to explore the effects of distributed model parameterizations
on the prediction of streamflow and other water fluxes at various spatial resolutions.
For this purpose, two parameterization methods namely: Hydrological Response
Units (HRU) and Multiscale Parameter Regionalization (MPR), were employed in a
grid based mesoscale hydrologic model (mHM). The HRU method uses a static
categorical classification scheme (e.g. k-NN method) to group the grid cells into
homogenous units based on the basin physical characteristics (e.g. topography, soil,
vegetation). Unique sets of model parameters are assigned to each HRU through the
calibration process. The MPR method, on the other hand, establishes a dynamic
quasi-continuous functional relationships between model parameters with basin
characteristics. In this case, the global parameters that establishes such functional
relationships are estimated through calibration, instead of model parameters of each grid
cells.
mHM with both parameterization methods was applied in the upper catchment of Necak
river (area of approximately 4000 km2) to simulate the spatio-temporal dynamics of
hydrological processes at three spatial resolutions: 2, 4, 8Â km. Model parameters for both
methods at each scale were separately estimated using dynamically dimensioned search
algorithm. The results indicated that both parameterization methods performed more or less
same for the streamflow simulation at all modeling scale as long as calibration was
performed. In cases, when model parameters were transferred from calibration scale to other
scales, the HRU method showed the significant deterioration for the daily streamflow
simulations, as compared to those obtained through the MPR method. Moreover,
HRU, as opposed to MPR, showed a significant bias in the conservation of mass
balance of spatially distributed water fluxes and state variables. Additionally, the soil
moisture patterns obtained with MPR was more realistic than those obtained through
HRU method, when compared against proxies derived from daily MODIS images
(NASA). |
|
|
|
|
|
|