 |
| Titel |
Identification of the HYPE hydrological model over the Indian subcontinent |
| VerfasserIn |
Ilias Pechlivanidis, David Gustafsson, Berit Arheimer |
| Konferenz |
EGU General Assembly 2014
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 16 (2014) |
| Datensatznummer |
250092794
|
| Publikation (Nr.) |
 EGU/EGU2014-7155.pdf |
|
|
|
|
|
| Zusammenfassung |
| Large-scale hydrological modelling has the potential to encompass many river basins, cross
regional and international boundaries and represent a number of different geophysical and
climatic zones. However the performance of this type of model is subject to several sources of
uncertainty/error which may be caused by, among others, the imperfectness of driving inputs,
i.e. regional and global databases. This uncertainty further leads to wrong model
parameterisation and incomplete process understanding. Data assimilation aims to utilize
both hydrological process knowledge (as embodied in a hydrologic model) and information
that can be gained from observations; hence information from model predictions and
observations is synergistically used to improve performance. This study presents a
methodology, drawn on experience from modelling with the HYPE model in the Indian
subcontinent (covering a modelled area of 4.9 million km2), to enhance identification of
highly parameterised large-scale hydrological models. The model was set up using
available large-scale datasets on topography, land use, soil, precipitation, temperature,
lakes, reservoirs, crop types, irrigation, evaporation, snow and discharge. A stepwise
automatic calibration is carried out to avoid, to a certain extent, errors incurring in some
model processes and being compensated by introducing errors in other parts of
the model. In addition, information from remote sensing data is assimilated in the
model to drive identification of parameters that control the spatial distribution of
potential evapotranspiration. Results show that despite the strong hydro-climatic
gradient over the domain, the model can adequately describe the hydrological process
in the Indian subcontinent. Overall, the median Kling-Gupta Efficiency (KGE)
increased from 0.08 to 0.64 during the calibration process using 43 stations of monthly
discharge series over the period 1971 to 1979. Finally, decomposition of the KGE
(i.e. into terms describing agreement in correlation, bias and variability between
observed and modelled streamflow series) allowed a thorough understanding of model
inadequacies.
Keywords
Large-scale hydrological modelling, HYPE, model identification, Kling-Gupta Efficiency,
remote sensing, India |
|
|
|
|
|
|