 |
| Titel |
Evaluating the performance in the Swedish operational hydrological forecasting systems |
| VerfasserIn |
Ilias Pechlivanidis, Thomas Bosshard, Henrik Spångmyr, Göran Lindström, Jonas Olsson, Berit Arheimer |
| Konferenz |
EGU General Assembly 2014
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 16 (2014) |
| Datensatznummer |
250092876
|
| Publikation (Nr.) |
 EGU/EGU2014-7239.pdf |
|
|
|
|
|
| Zusammenfassung |
| The production of hydrological forecasts generally involves the selection of model(s)
and setup, calibration and initialization, verification and updating, generation and
evaluation of forecasts. Although, field data are commonly used to calibrate and initiate
hydrological models, technological advancements have allowed the use of additional
information, i.e. remote sensing data and meteorological ensemble forecasts, to improve
hydrological forecasts. However, the precision of hydrological forecasts is often
subject to uncertainty related to various components of the production chain and data
used.
The Swedish Meteorological and Hydrological Institute (SMHI) operationally produces
hydrological medium-range forecasts in Sweden using two modeling systems based on the
HBV and S-HYPE hydrological models. The hydrological forecasts use both deterministic
and ensemble (in total 51 ensemble members which are further reduced to 5 statistical
members; 2, 25, 50, 75, 98% percentiles) meteorological forecasts from ECMWF
to add information on the uncertainty of the predicted values. In this study, we
evaluate the performance of the two operational hydrological forecasting systems
and identify typical uncertainties in the forecasting production chain and ways
to reduce them. In particular, we investigate the effect of autoregressive updating
of the forecasted discharge, and of using the median of the ensemble instead of
deterministic forecasts. Medium-range (10 days) hydrological forecasts across 71
selected indicator stations are used. The Kling-Gupta Efficiency and its decomposed
terms are used to analyse the performance in different characteristics of the flow
signal.
Results show that the HBV and S-HYPE models with AR updating are both capable of
producing adequate forecasts for a short lead time (1 to 2 days), and the performance
steadily decreases in lead time. The autoregressive updating method can improve
the performance of the two systems by 30 to 40% in terms of the KGE. This is
mainly because the method has a significant impact on the improvement of discharge
volume. S-HYPE seems to perform slightly better than HBV in the longer lead
time, probably because the S-HYPE system is capable of updating the lake water
level, which has an impact on the longer lead times. Moreover, the deterministic
and ensemble HBV systems with AR updating perform fairly similar for all lead
times.
Keywords: Hydrological forecasting, S-HYPE, HBV, Operational production,
Kling-Gupta Efficiency, Uncertainty. |
|
|
|
|
|
|