Skip to content
Svenska
To start page
  • Report
  • Hydropower

Verification of run-off data from three Swedish catchments comparing ECMWF and WTM

| 2015:204 | BJØRGE JANSEN
In this project an analysis have been performed, where the prognoses of runoff from two models, EC (European Centre) and WTM (Weather Terrain Model), are compared for three Swedish catchments.

The runoff prognoses are run through a HBV-model. WTM has shown some improvement in temperature and precipitation forecasts compared to EC for observation stations, and the basis for the analysis are to see if those improvements can be recreated in runoff-predictions.

The analysis shows that both EC and WTM in general have too low runoff for two catchments, Kultsjön and Ottsjön, and this bias is slightly larger in WTM. For Burvattnet it is the other way around, EC and WTM has slightly too much runoff. It is shown that the deviations is largely due to the calibration (“Pcorr”) in the HBV-model. Calibration of the HBV-model is adapted to another model, which seems to be the reason for much of the deviation between prognoses and observed runoff. Without this calibration, it seems like runoff in both EC and WTM would have been closer to the observed.

Comparison between the two models is, as a consequence of this, not relevant. Future similar test will need to address this issue. There is also potential in improving the analysis by using a more sophisticated HBV-model and in choosing catchments where the real topography differ more from the topography in the weather-model.