Group seminar on 04. January, 14:15 CET
Atmospheric Predictability: Why Butterflies Are Not of Practical Importance
Chen Wang
I will present an article entitled “Atmospheric Predictability: Why Butterflies Are Not of Practical Importance.” by Durran and Gingrich (2014).
In a classical paper by Lorenz (1969), it is shown that the errors from the smallest scales can propagate upscale and destroy the atmospheric predictability. However, Durran and Gingrich (2014) show that tiny relative errors in the large scales may dominate the predictability because they have larger background kinetic energy. First I will briefly introduce the smoothed saturation Lorenz-Rotunno-Snyder (ssLRS) model used in this study and its response to a few types of hypothetical initial errors. Then I will show the initial error distribution computed from real weather forecasts. The error growth from ssLRS model with similar initial error distribution approximately resembles the error growth of the mesoscale weather forecasts. Furthering this, experiments with initial errors in selected scales show that errors in large scales dominate the error growth. The slopes of the spectra also control the upscale and downscale propagation speed. Finally, I will briefly discuss the limits of the study and relate it to my own research.