BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Machine Learning for Stochastic Parametrisation - Hannah Christens
 en | Department of Physics\, University of Oxford
DTSTART:20201110T110000Z
DTEND:20201110T123000Z
UID:TALK152509@talks.cam.ac.uk
CONTACT:Tudor Suciu
DESCRIPTION:Atmospheric models used for weather and climate prediction are
  traditionally formulated in a deterministic manner. In other words\, give
 n a particular state of the resolved scale variables\, the most likely for
 cing from the sub-grid scale motion is estimated and used to predict the e
 volution of the large-scale flow. However\, the lack of scale-separation i
 n the atmosphere means that this approach is a large source of error in fo
 recasts. Over the last decade an alternative paradigm has developed: the u
 se of stochastic techniques to characterise uncertainty in small-scale pro
 cesses. These techniques are now widely used across weather\, seasonal for
 ecasting\, and climate timescales. While there has been significant progre
 ss in emulating parametrisation schemes using machine learning\, the focus
  has been entirely on deterministic parametrisations. In this presentation
  I will discuss data driven approaches for stochastic parametrisation. I w
 ill describe experiments which develop a stochastic parametrisation using 
 the generative adversarial network (GAN) machine learning framework for a 
 simple atmospheric model. I will conclude by discussing the potential for 
 this approach in complex weather and climate prediction models.\n\n"Zoom l
 ink":https://us02web.zoom.us/j/89975312802?pwd=dkpvc0M3RGhtY1JOUm5Hc1dCNm9
 IZz09 \n\nMeeting ID: 899 7531 2802 \n\nPasscode: 945323
LOCATION:https://us02web.zoom.us/j/89975312802?pwd=dkpvc0M3RGhtY1JOUm5Hc1d
 CNm9IZz09
END:VEVENT
END:VCALENDAR