CGD Seminar- Hot models and assessed warming: challenges for the community to effectively use CMIP6

Gavin Schmidt and Zeke Hausfather

Seminar
Apr. 25, 2023

11:00 am – 12:00 pm MDT

MESA LAB MAIN SEMINAR ROOM AND ON THE CGD YOUTUBE CHANNEL

Webcast

Main content

Historical climate models have generally performed well in projecting temperatures in the years after they were published. Models have long had a range of equilibrium climate sensitivity (ECS) of between 2C and 4.7C, which has closely matched the assessed range of climate sensitivity (1.5C-4.5C or 2C-4.5C) in past IPCC reports. However, in CMIP6 a subset of models – around 30% of the total – have an ECS higher than any models in CMIP5 (e.g. exceeding 4.7C), resulting in higher future warming projections in the CMIP6 multimodel mean (MMM) than in comparable scenarios in CMIP5 and a poor match to historical temperature records (though they may perform skillfully against climatology or other variables). At the same time, the community has used multiple lines of evidence to help constrain climate sensitivity, and the AR6 found a likely ECS range of 2.5C to 4C with a very likely range of 2C to 5C. This mismatch between CMIP6 models and other lines of evidence led the AR6 to create "assessed warming" projections that differ significantly from the CMIP6 MMM. Here we discuss the development of these "assessed warming" projections, as well as some of their shortcomings that have prevented widespread adoption by the community, and suggest paths forward in a post-model-democracy world.