Articles | Volume 20, issue 9
Hydrol. Earth Syst. Sci., 20, 3561–3579, 2016
https://doi.org/10.5194/hess-20-3561-2016
Hydrol. Earth Syst. Sci., 20, 3561–3579, 2016
https://doi.org/10.5194/hess-20-3561-2016

Research article 06 Sep 2016

Research article | 06 Sep 2016

Error reduction and representation in stages (ERRIS) in hydrological modelling for ensemble streamflow forecasting

Ming Li1, Q. J. Wang2, James C. Bennett2, and David E. Robertson2 Ming Li et al.
  • 1CSIRO Data61, Floreat, WA, Australia
  • 2CSIRO Land and Water, Clayton, VIC, Australia

Abstract. This study develops a new error modelling method for ensemble short-term and real-time streamflow forecasting, called error reduction and representation in stages (ERRIS). The novelty of ERRIS is that it does not rely on a single complex error model but runs a sequence of simple error models through four stages. At each stage, an error model attempts to incrementally improve over the previous stage. Stage 1 establishes parameters of a hydrological model and parameters of a transformation function for data normalization, Stage 2 applies a bias correction, Stage 3 applies autoregressive (AR) updating, and Stage 4 applies a Gaussian mixture distribution to represent model residuals. In a case study, we apply ERRIS for one-step-ahead forecasting at a range of catchments. The forecasts at the end of Stage 4 are shown to be much more accurate than at Stage 1 and to be highly reliable in representing forecast uncertainty. Specifically, the forecasts become more accurate by applying the AR updating at Stage 3, and more reliable in uncertainty spread by using a mixture of two Gaussian distributions to represent the residuals at Stage 4. ERRIS can be applied to any existing calibrated hydrological models, including those calibrated to deterministic (e.g. least-squares) objectives.

Please read the corrigendum first before accessing the article.
Download

The requested paper has a corresponding corrigendum published. Please read the corrigendum first before downloading the article.