Parts | Days | Selection | Search | Updates | Downloads | Help
DY: Fachverband Dynamik und Statistische Physik
DY 29: Posters II
DY 29.44: Poster
Thursday, March 29, 2012, 17:00–19:00, Poster A
Towards first-passage-time prediction in temperature time series — •Anja von Wulffen and Holger Kantz — Max Planck Institute for the Physics of Complex Systems, Dresden, Germany
Current operational weather forecasts are based on detailed models run for a lead time of about ten days. Efforts to issue seasonal forecasts further into the future using similar models are still very experimental and deal with a high uncertainty. On these longer time scales, very specific prediction tasks such as the first passage time until a threshold crossing, e.g. the time until first frost, are of significant interest.
In order to see whether for such specific questions predictions based solely on a statistical analysis of the time series might not only require less resources than the full models but also yield better results, we study these first passage time distributions for actual temperature data and evaluate their potential predictability.
We also investigate the possibility of improvements through incorporating a second time series supplying information about the relevant slower-evolving atmospheric patterns such as the North Atlantic Oscillation.