T2 - Towards seamless prediction of extremes
Other researcher: Federico Grazzini (Postdoc)
There is a growing interest in the weather forecast community to complement direct output from numerical weather prediction (NWP) models with other diagnostics tools designed specifically for extremes events. In the previous transfer project T1 a strategy was designed to classify precipitation extremes in Northern Italy. Working in collaboration with the regional meteorological service ARPAE-SIMC, we developed diagnostics to follow the unfolding of the extreme event across scales from its earliest stage in the forecast, to its actual occurrence. Building on this success for a weather event in a particular region, this project aims at expanding and generalizing this dynamical methodology to other regions of Europe and into the sub-seasonal forecast range (10-30 days), and at creating prototype forecast products at ECMWF as well as at ARPAE-SIMC. The project provides the forecaster a "narrative" of the unfolding of an extreme event, which describes the predictability of the relevant dynamical processes. The project supports the forecasting and warning chain with important dynamical understanding which goes beyond relying on black box forecast products from imperfect models. To this end a truly seamless diagnostic framework is developed. The accompanying prototype products will demonstrate how dynamical thinking and process-oriented predictors can serve as a "warning bell" in critical weather events, at all forecast ranges. As a first action, the project will transfer the insight of T1 into a novel forecasting method for extreme precipitation events in N-Italy. Next, we will develop a unified, seamless framework which systematically documents the linkage of the large-scale flow to extremes and their dynamical drivers for other European regions. This will be complemented by a prototype forecast framework which integrates existing forecast products for extremes with forecast probabilities computed for a hierarchy of indices describing the dynamical precursors of an extreme. The diagnostic framework and prototype forecast products will be implemented in close collaboration with ECMWF and ARPAE-SIMC and for both precipitation and temperature extremes. Focusing on the predictable scales of motion and processes, our proposed seamless approach will provide a way to increase preparedness for forecasting extremes from the medium to extended forecast ranges (10-30 days), ranges where classical forecast products are inaccurate. The need for better guidance across lead times is becoming even more substantial in view of the increasing likelihood of extreme events in a warming climate and remaining deficiencies in numerical models for the foreseeable future. Finally, the project will contribute to the key goal of WMO’s HIWeather project to strengthen warning procedures across scales.