EGS XXVI General Assembly, Nice, France, March 2001

NP6. Quantifying predictability (co-sponsored by OA):

Report

This session comprised of 49 orally presented papers arranged around the following topics: General issues of predictability; Analysis and targeted observations; Initial value uncertainty; Ensembles; Verification of probabilistic forecasts; and Application of probabilistic forecasts. As the main topics indicate, a wide range of problems were addressed from theoretically oriented questions to the practical issues related, for example, to the use of ensemble forecasting in hydrological applications. The first presentation under each topic was given by an invited speaker, aimed at providing an introduction of a particular area within the main topic to help non-specialists familiarize themselves with different areas of research. The oral presentations were complemented by a poster session with five papers. The two-day program was attended by an active audience of 40-80 scientists from around the world. At the end of the program, an open discussion session on "Predicting forecast uncertainty: Real life applications" followed. Despite the late starting time of this session, the discussion was well attended by an active audience. After short introductions the attendees discussed the questions why and how to communicate weather related risk/uncertainty to the user community. In the course of the discussion it became clear that the scientific community agrees on the usefulness of a probabilistic approach to numerical prediction, yet there is still only a limited set of forecasts generated by ensemble systems. Studies on the potential economic value of ensemble-based products for specific users (severe weather, floods and other hydrological, energy, agriculture, shipping, insurance, etc. applications) that could promote a wider use of ensemble products should be carried out. The scientific comunity should send a clear signal that despite the fact that the current operational ensemble systems are not perfect, they can be used to generate forecast products that can have more value than those based on single deterministic forecasts. A selective set of presentations from the NP6 session are planned to be published in a special issue of Nonlinear Processes in Geophysics. 

Zoltan Toth