[ESAC Student Seminar] Suyash Bire

“Teaching Oceanography to Machines”

Abstract: NOAA’s Climate Forecast System provides predictions of the state of the climate over three timescales — weekly, monthly, and seasonal. Predictability over monthly and seasonal timescales is enabled by coupling an atmospheric model to ocean, sea-ice, and land models. All of these models tend to be very computationally expensive. Recently, a number of data-driven atmospheric emulators have been proposed that can significantly decrease computational expense as compared to traditional numerical models at a minimal loss of accuracy. However, an effective ocean emulator is still missing. In this talk I will present a data-driven ocean model that makes predictions of up to two months. I will also peek into the internals of this model and explain how it “learns”.

ESAC Student Seminar Series

A forum for students and postdocs to share recent research, hone presentation skills, and build community among peers, sponsored by the EAPS Student Advisory Committee. Open to current EAPS graduate and undergraduate students and postdocs. Typically hosted on Thursdays during the semester, including pizza lunch.

Contact: esac.officers@gmail.com