Motivation and Overview of Hydrological Ensemble Post-processing

In this introduction to this chapter on hydrologic post-processing, we discuss the different but complementary directives that the “art” of post-processing must satisfy: the particular directive defined by specific applications and user needs; versus the general directive of making any ensemble member indistinguishable from the observations. Also discussed are the features of hydrologic post-processing that are similar and separate from meteorological post-processing, providing a tie-in to early chapters in this handbook. We also provide an overview of the different aspects the practitioner should keep in mind when developing and implementing algorithms to adequately “correct and calibrate” ensemble forecasts: when forecast uncertainties should be characterized separately versus maintaining a “lumped” approach; additional aspects of hydrological ensembles that need to be maintained to satisfy additional user requirements, such as temporal covariability in the ensemble time series, an overview of the different post-processing approaches being used in practice and in the literature, and concluding with a brief overview of more specific requirements and challenges implicit in the “art” of post-processing.