Speaker: Iuliia Polkova (Univ. Hambourg)

Presentation covers initialization of climate predictions in general.
I will try not to bore audience with the analyses of skills or
mechanisms for predictability, though the presentation relies on such
in already published papers. Instead…  it would be interesting to
have an interactive conversation on initialization, even though it
might not be everyone’s topic, and to talk about challenges of this
field of research and maybe identify interesting directions also with
input from the audience. As a teaser for discussion, I have collected
some ideas behind different initialization and assimilation methods
for decadal predictions and my speculations why they fulfilled or not
fulfilled the expectations of their developers.  I also mention
whether uninitialized predictions can replace initialized predictions
and speculate about what we can try to learn from simpler climate
models in terms of initialization and data assimilation. I suppose the
presentation would be of interest for prediction system developers but
hopefully not only. If someone wants to know how observations enter
prediction systems and implications of this for the quality of
predictions, they should drop by. Hopefully, we can have interesting
discussions.

Please note this seminar has been postponed to: Wed Apr 20, 14:00, Salle 417

Ce séminaire a été reporté au : Mercredi 20 avril, 14:00, Salle 417