search

chrisroadmap / fair-calibrate

Calibration package for fair (>=2.1)
MIT License
2 stars 0 forks source link

TCRE needs fitting (calibration) #74

Closed chrisroadmap closed 1 year ago

chrisroadmap commented 1 year ago

This one is up for debate, because I think it is the weakest and most speculative of the AR6 climate assessments (other than projected SSP warming - which I only sense check).

With everything else calibrated and distribution-fitted (ECS, TCR, observed warming, observed OHC, aerosol forcing), the TCRE has not been. In AR6 we used the Chapter 5 assessment and calculated TCRE from TCR and airborne fraction. This is annoying for calibration as we need to do separate 1% runs on top of the 1.5 million member emissions-driven ensemble.

It's come to light when I use the FaIR (non-TCRE calibrated) calibration in other work. FaIR 1.6-AR6 had a median ZEC50 of -0.05 K in the 750 and 1000 Tg bell experiments (compare MAGICC -0.09 K, CMIP6 ESMs -0.078 K). FaIR 2.1-AR6 has a median ZEC50 of zero - so we're bang on chapter 5, but further from chapter 4 😆

chrisroadmap commented 1 year ago

TCRE as assessed by AR6 is a poor constraint, so while it has been included as a diagnostic, we don't constrain to it (present day CO2 does this). Addressed in #84.