Thread by @hausfath

Theres been a slew of media stories about climate models running hot. While some new models have much higher climate sensitivity, most new models do not. Multiple lines of evidence suggest climate sensitivity is still likely ~3C (+/- 1.5C) thebreakthrough.org/issues/energy/…

A thread: 1/11

Climate modeling groups around the world currently running the latest generation of climate models that will be featured in the upcoming 2021 IPCC 6th Assessment Report (AR6). Around 30 of the expected 100 different models have reported climate sensitivity so far. 2/11

Nine of the 30 CMIP6 models currently available have a climate sensitivity higher than any CMIP5 models. These have pulled up the average, despite the fact that around 40% (13) of the models have a climate sensitivity lower than the 3.2C average of CMIP5. #notallmodels 3/11

The average climate sensitivity of CMIP6 models is higher so far — 3.8C compared to 3.2C in CMIP5. As more models have come in, however, the average sensitivity of CMIP6 models has fallen. carbonbrief.org/cmip6-the-next… 4/11

This may reflect a bias toward faster reporting by the more well-funded and staffed modeling groups that also likely made the largest changes between CMIP5 and CMIP6. As the remaining 70 models are released, the average climate sensitivity may move closer to that of CMIP5. 5/11

Models are an important way that climate scientists estimate sensitivity, but they’re far from the only one. Sensitivity can also be estimated through emergent constraints, instrumental temperature records, and from climate proxy records from the Earth’s more distant past. 6/11

Of 142 studies, only a third suggest an ECS of 5C+ is within the uncertainty range, and only six suggest a most likely ECS value of 5C+. Relatively few non-model studies – those using instrumental records or paleoclimate proxies – support 5C+ ECS. carbonbrief.org/explainer-how-… 7/11

Moreover, the new high sensitivity models in CMIP6 appear to do rather poorly in "hindcasting" past temperatures. Some show almost no 20th century warming; others show too much warming, particularly over the last 30 years. earth-syst-dynam-discuss.net/esd-2019-86/ 8/11

Models with a TCR of around 2.2C or less (in blue) tend to agree much better with observations. High ECS models tend to have high TCR. In this case, all the models in the figure with an ECS above 5C (except for one — CESM2) also have a TCR value above 2.5C. 9/11

Hindcasts are not a perfect test of model skill; natural variability can confound our interpretation of shorter-term diffs between models and obs. nicklutsko.github.io/blog/2020/02/0…

But its hard to reconcile no 20th century warming (for example) with the actual temperature record. 10/11

These new 5C ECS models should remind us that large uncertainties (and long tails of risk) remain, but they do not by themselves overturn the long-term consensus that climate sensitivity is likely somewhere around 3C (+/- 1.5C) per doubling of CO2. 11/11