University of Cambridge > Talks.cam > CEDSG-AI4ER > Developing data-driven models to emulate GCM’s

Developing data-driven models to emulate GCM’s

Add to your list(s) Download to your calendar using vCal

  • UserRachel Furner, University of Cambridge/ British Antarctic Survey
  • ClockTuesday 21 January 2020, 12:00-13:00
  • HouseBullard Lab, Seminar Room.

If you have a question about this talk, please contact Jonathan Rosser.

Chair: Scott Hosking Abstract: Climate models represent the best tools we have to predict, understand and potentially mitigate climate change, however these process-based models are incredibly complex and require huge amounts of high-performance computing resources. Machine learning offers opportunities to greatly improve the computational efficiency of these models. Here we discuss preliminary work looking to develop analogous data-driven models.

The talk will focus on two separate pieces of work:

We begin by presenting a neural network capable of replicating the behaviour of Lorenz model (an idealised dynamical system which includes chaos). This is used as a test-bed to assess the importance of the time-stepping method used in network based models of dynamical processes, and the loss functions used to train the model.

Secondly we present very preliminary results from work developing a linear regressor to emulate an idealised general circulation model of the ocean. The regressor is trained using the outputs from an idealised sector configuration of general circulation model (MITgcm). Our aim is to develop an algorithm which is able to predict the future state of the model to a similar level of accuracy. Some results from investigations into the sensitivity of the regressor to various inputs (e.g. temperature on different spatial and temporal scales, and meta-variables such as location information) will be presented.

Work with supervisors: Peter Haynes (University of Cambridge), Dan Jones (British Antarctic Survey), Dave Munday (British Antarctic Survey), Brooks Paige (UCL), Emily Shuckburgh (University of Cambridge)

This talk is part of the CEDSG-AI4ER series.

Tell a friend about this talk:

This talk is included in these lists:

Note that ex-directory lists are not shown.

 

© 2006-2021 Talks.cam, University of Cambridge. Contact Us | Help and Documentation | Privacy and Publicity