Abstract
AbstractDifferent post-processing techniques are frequently employed to improve the outcome of ensemble forecasting models. The main reason is to compensate for biases caused by errors in model structure or initial conditions, and as a correction for under- or overdispersed ensembles. Here we use the Ensemble Model Output Statistics method to post-process the ensemble output from a continental scale hydrological model, LISFLOOD, as used in the European Flood Awareness System (EFAS). We develop a method for local calibration and interpolation of the post-processing parameters and compare it with a more traditional global calibration approach for 678 stations in Europe based on long term observations of runoff and meteorological variables. For the global calibration we also test a reduced model with only a variance inflation factor. Whereas the post-processing improved the results for the first 1-2 days lead time, the improvement was less for increasing lead times of the verification period. This was the case both for the local and global calibration methods. As the post-processing is based on assumptions about the distribution of forecast errors, we also present an analysis of the ensemble output that provides some indications of what to expect from the post-processing.
Highlights
Ensemble forecasting has a long tradition in meteorological forecasting but has only been widely applied in hydrological forecasting since a few years
We will first present the results from local calibration validation of the ensemble model output statistics (EMOS) parameters with different transformations
The continuous ranked probability score (CRPS) generally increases when the spatial penalty is included in the calibration, but for some catchments it even declines, indicating that the first iteration found a local optimum
Summary
Ensemble forecasting has a long tradition in meteorological forecasting but has only been widely applied in hydrological forecasting since a few years. Previous studies such as Cloke and Pappenberger (2009) and Wetterhall et al (2013) showed how the use of ensemble streamflow forecasts has increased during the last years. Whereas the variance can be seen as an indication of the reliability, it is usually assumed that the ensembles need postprocessing (Gneiting et al 2005) and that this is essential for hydrological variables (Hemri 2018). For meteorological and hydrological forecasting, the ensembles will often exhibit both bias and dispersion errors
Sign-in/Register to view highlights & summary 
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call