Diagnosing Rossby wave train errors in numerical weather forecasts (RW-FCST)

点击翻译按钮获取中文摘要

Severe weather over Europe is known to be occasionally associated with long-lived wave trains. This suggests potential predictability on the time-scale of a week and even longer — a potential which is far from being exploited today. As a prerequisite for improved forecasts this project aims to diagnose deficiencies in current weather forecast systems regarding the generation, propagation, and decay of such long-lived wave trains. During phase 1 we demonstrated that also in a climatologicalsense severe weather over Europe is associated with distinct long-lived wave trains. We developed a novel diagnostic to quantify predictive skill regarding wave trains. The project will be continued during phase 2 by significantly advancing and extending the analysis. First, we shall apply our new diagnostic to systematically quantify forecast errors in long data sets. In addition, we shall explore and develop the utility of so-called “wave activity flux” diagnostics, which are distinct from and partly complementary to the more conventional diagnostics including our own one. We shall apply our diagnostics to ensembles of forecasts in order to learn how well these are able to quantify forecast uncertainty regarding long-lived wave trains. Overall, this project pursues similar goals as RW-GUIDE, but uses vastly different and largely complementary methods.