In this talk, Dr. Judith Berner will discuss how forecast skill in two NSF NCAR models (CESM2 and MPAS-A) depends on land and ocean states, not just atmospheric initial conditions. She will also assess if AI systems can capture this state-dependent predictability. Findings indicate land initialization is the main predictability source after the third week of a forecast.
The event is hybrid and will be held at the NCAR-Foothills Laboratory at 3450 Mitchell Lane, room FL2-1022 (the Large Auditorium). Refreshments will be served at 1:45 p.m. MT, followed by an informal discussion from 3:00 p.m. to 3:30 p.m. MT. Remote participants may watch the
live webcast and submit questions via Slido.