MPE Wednesday 

Schedule MPE WEDNESDAY, 9TH OCTOBER 

 12:30 – 14:00 AT ICSM 402B – Lunch

14:00 – 15:00 AT ICSM 402 – Prof Nedjeljka Žagar on Multiscale, multivariate aspects of atmospheric predictability

15:00 – 15:30 AT ICSM 402B – Refreshments

15:30 – 16:15 AT ICSM 402 – Aythami talk 

16:15 – 17:00 AT ICSM 402 – Tasmin Talk

17:00 AT ICSM 402B Drinks Reception

————————————————————————

Prof Nedjeljka Žagar (Meteorologisches Institut, Universität Hamburg) on Multiscale, multivariate aspects of atmospheric predictability

Abstract:

Imperfections in the numerical weather prediction models and uncertainties in the initial state for prediction together with chaotic nature of atmospheric motions impose limitation on the accuracy of day-to-day weather forecasts. Many studies of the errors in predictions have focused on the extra-tropical, quasi-geostrophic dynamics and often considered the error-free large-scale initial state. In contrast, the global numerical weather prediction (NWP) models and ensemble prediction systems are characterized by significant initial uncertainties at large scales, especially in the tropics. I will present the scale-dependent properties of analysis and forecast uncertainties using the data from a state-of-the-art ensemble prediction system and from an observing-system simulation experiment using a perfect model. Then I will review a long tradition of representing the growth of errors in NWP by simple parametric models and introduce a new such model. Applying the new model to the multivariate, multiscale forecast error data, I will discuss evolving limits of practical predictability in the past few years and their sensitivity to the reduction of analysis uncertainties at many scales.

More information on Prof Zagar’s group can be found here.  

Prof Zagar’s personal page can be found here. 

————————————————————————-

Tasmin Symons on The theory and challenges of Gaussianity on the sphere.

Abstract:

This talk aims to give a gentle introduction to the theory and practice of Gaussian random fields (GRFs) parametrised by the sphere. Such objects are ubiquitous in geostatistics, and care must be taken as it cannot be taken for granted that these fields behave in the same way as their better-studied Euclidean counterparts. We discuss some classical and recent results characterising GRFs in space and space-time, where space is the sphere, and the challenges of translating this theory into practice.

 ———————————————————————

Aythami de Leon Bethencourt

Title: TBA

Abstract: TBA