We have described and made available through Nature Scientific Data our global high resolution bias corrected CMIP6 derived wet-bulb and wet-bulb globe temperature data and codes. This is currently the most accurate formulation of these metrics available. www.nature.com/articles/s41...
A global high-resolution and bias-corrected dataset of CMIP6 projected heat stress metrics - Scientific Data
Scientific Data - A global high-resolution and bias-corrected dataset of CMIP6 projected heat stress metrics
www.nature.com
February 23, 2025 at 5:51 PM
We have described and made available through Nature Scientific Data our global high resolution bias corrected CMIP6 derived wet-bulb and wet-bulb globe temperature data and codes. This is currently the most accurate formulation of these metrics available. www.nature.com/articles/s41...
For the future, the risk of #AMOC shutdown won't mostly come from meltwater discharge. Many of the standard CMIP6 future climate scenarios shut down the AMOC without even including Greenland mass loss, just from the CO2 emissions effect on ocean & atmosphere. 3/3
iopscience.iop.org/article/10.1...
iopscience.iop.org/article/10.1...
Shutdown of northern Atlantic overturning after 2100 following deep mixing collapse in CMIP6 projections
Shutdown of northern Atlantic overturning after 2100 following deep mixing collapse in CMIP6 projections, Drijfhout, Sybren, Angevaare, Joran R, Mecking, Jennifer, van Westen, René M, Rahmstorf, Stefan
iopscience.iop.org
May 27, 2026 at 1:12 PM
For the future, the risk of #AMOC shutdown won't mostly come from meltwater discharge. Many of the standard CMIP6 future climate scenarios shut down the AMOC without even including Greenland mass loss, just from the CO2 emissions effect on ocean & atmosphere. 3/3
iopscience.iop.org/article/10.1...
iopscience.iop.org/article/10.1...
Here's the latest comparison of global surface temperature observations (red) with IPCC climate model simulations (through March 2025, via @hausfath.bsky.social):
May 23, 2025 at 12:48 AM
Here's the latest comparison of global surface temperature observations (red) with IPCC climate model simulations (through March 2025, via @hausfath.bsky.social):
The #AMOC by 2100 will thus be so weak (~8 Sv) that it is very likely on the way to full shutdown, as our study last year showed: iopscience.iop.org/article/10.1...
In my view, this now makes a full #AMOC shutdown more likely than not. Definitely not a low likelihood risk, as we used to think.
In my view, this now makes a full #AMOC shutdown more likely than not. Definitely not a low likelihood risk, as we used to think.
Shutdown of northern Atlantic overturning after 2100 following deep mixing collapse in CMIP6 projections
Shutdown of northern Atlantic overturning after 2100 following deep mixing collapse in CMIP6 projections, Drijfhout, Sybren, Angevaare, Joran R, Mecking, Jennifer, van Westen, René M, Rahmstorf, Stefan
iopscience.iop.org
April 15, 2026 at 6:16 PM
The #AMOC by 2100 will thus be so weak (~8 Sv) that it is very likely on the way to full shutdown, as our study last year showed: iopscience.iop.org/article/10.1...
In my view, this now makes a full #AMOC shutdown more likely than not. Definitely not a low likelihood risk, as we used to think.
In my view, this now makes a full #AMOC shutdown more likely than not. Definitely not a low likelihood risk, as we used to think.
10/ And Drijfhout et al. (2025) found that the #AMOC could fully shut down after 2100 when deep-ocean mixing collapses, a mechanism most models still don't capture.
The IPCC's "unlikely" assessment is looking increasingly out of date.
The IPCC's "unlikely" assessment is looking increasingly out of date.
Shutdown of northern Atlantic overturning after 2100 following deep mixing collapse in CMIP6 projections
Shutdown of northern Atlantic overturning after 2100 following deep mixing collapse in CMIP6 projections, Drijfhout, Sybren, Angevaare, Joran R, Mecking, Jennifer, van Westen, René M, Rahmstorf, Stefan
iopscience.iop.org
April 20, 2026 at 7:47 AM
10/ And Drijfhout et al. (2025) found that the #AMOC could fully shut down after 2100 when deep-ocean mixing collapses, a mechanism most models still don't capture.
The IPCC's "unlikely" assessment is looking increasingly out of date.
The IPCC's "unlikely" assessment is looking increasingly out of date.
Has the Atlantic Meridional Overturning Circulation (AMOC) weakened over the last decades? In our new study, we combine state-of-the-art CMIP6 models and observation-based estimates of the air-sea heat flux in the North Atlantic to show that the AMOC has not declined since the 1960s! 🌊
Atlantic overturning inferred from air-sea heat fluxes indicates no decline since the 1960s - Nature Communications
The AMOC is crucial for the global ocean overturning circulation and controls the climate around the North Atlantic. Here, the authors use 24 Earth System Models from the CMIP6 to demonstrate tha...
www.nature.com
January 15, 2025 at 10:11 AM
Has the Atlantic Meridional Overturning Circulation (AMOC) weakened over the last decades? In our new study, we combine state-of-the-art CMIP6 models and observation-based estimates of the air-sea heat flux in the North Atlantic to show that the AMOC has not declined since the 1960s! 🌊
Seems important:
All ('IPCC') CMIP6 models are completely unable to reproduce what NASA satellites observed!
The rate of global warming (starting with Earth's Energy Imbalance and now surface air temperature) has more than doubled!
All ('IPCC') CMIP6 models are completely unable to reproduce what NASA satellites observed!
The rate of global warming (starting with Earth's Energy Imbalance and now surface air temperature) has more than doubled!
June 17, 2025 at 12:04 PM
Seems important:
All ('IPCC') CMIP6 models are completely unable to reproduce what NASA satellites observed!
The rate of global warming (starting with Earth's Energy Imbalance and now surface air temperature) has more than doubled!
All ('IPCC') CMIP6 models are completely unable to reproduce what NASA satellites observed!
The rate of global warming (starting with Earth's Energy Imbalance and now surface air temperature) has more than doubled!
Is there an acceleration in global warming expected in the CMIP6 climate models? yes! But not as easy to detect in the real world.
(from www.realclimate.org/index.php/ar...)
(see also: www.realclimate.org/index.php/ar... from 2024).
(from www.realclimate.org/index.php/ar...)
(see also: www.realclimate.org/index.php/ar... from 2024).
March 9, 2026 at 1:44 AM
Is there an acceleration in global warming expected in the CMIP6 climate models? yes! But not as easy to detect in the real world.
(from www.realclimate.org/index.php/ar...)
(see also: www.realclimate.org/index.php/ar... from 2024).
(from www.realclimate.org/index.php/ar...)
(see also: www.realclimate.org/index.php/ar... from 2024).
I have a new update to climate model-observation comparisons over at The Climate Brink, covering CMIP3, CMIP5, and CMIP6. Models perform well globally. The latest generation shows too much long-term warming but better reproduces recent trends: www.theclimatebrink....
December 6, 2025 at 8:42 PM
I have a new update to climate model-observation comparisons over at The Climate Brink, covering CMIP3, CMIP5, and CMIP6. Models perform well globally. The latest generation shows too much long-term warming but better reproduces recent trends: www.theclimatebrink....
Downscaled daily CMIP6 projections of Tmax, Tmin, P, and e_a at ~1km resolution over the contiguous USA 🌊🧪 #climate
NASA Earth Exchange Downscaled Climate Projections 30 Arcseconds CMIP6 - Scientific Data
Scientific Data - NASA Earth Exchange Downscaled Climate Projections 30 Arcseconds CMIP6
www.nature.com
December 21, 2024 at 2:15 PM
Downscaled daily CMIP6 projections of Tmax, Tmin, P, and e_a at ~1km resolution over the contiguous USA 🌊🧪 #climate
How likely is an #AMOC shutdown? Here's our latest study based on 768 runs with 38 standard climate models (the CMIP6 models). Note these do not include any meltwater from Greenland mass loss, so might underestimate the risk a bit. 2/2
iopscience.iop.org/article/10.1...
iopscience.iop.org/article/10.1...
December 5, 2025 at 11:42 AM
How likely is an #AMOC shutdown? Here's our latest study based on 768 runs with 38 standard climate models (the CMIP6 models). Note these do not include any meltwater from Greenland mass loss, so might underestimate the risk a bit. 2/2
iopscience.iop.org/article/10.1...
iopscience.iop.org/article/10.1...
"CMIP6 models show that the Earth likely will reach 2°C of global warming by the 2040s without significant policy changes"
agupubs.onlinelibrary.wiley.com/doi/10.1029/...
agupubs.onlinelibrary.wiley.com/doi/10.1029/...
What Does Global Land Climate Look Like at 2°C Warming?
CMIP6 models show that the Earth likely will reach 2°C of global warming by the 2040s without significant policy changes Geographic pattern of changes in key climate indicators portend unfavorabl...
agupubs.onlinelibrary.wiley.com
March 2, 2025 at 11:46 AM
"CMIP6 models show that the Earth likely will reach 2°C of global warming by the 2040s without significant policy changes"
agupubs.onlinelibrary.wiley.com/doi/10.1029/...
agupubs.onlinelibrary.wiley.com/doi/10.1029/...
Our new study is out:
The Arctic Beaufort Gyre is projected to weaken or vanish in most CMIP6 models under future climate change.
The gyre would thus stop storing freshwater in its centre, with possibly large impacts on the Arctic and beyond.
Read: agupubs.onlinelibrary.wiley.com/doi/full/10....
The Arctic Beaufort Gyre is projected to weaken or vanish in most CMIP6 models under future climate change.
The gyre would thus stop storing freshwater in its centre, with possibly large impacts on the Arctic and beyond.
Read: agupubs.onlinelibrary.wiley.com/doi/full/10....
The Arctic Beaufort Gyre in CMIP6 Models: Present and Future
CMIP6 models generally simulate a too strong, large, and northward-extending Beaufort Gyre (BG) during the historical period Among the least unrealistic models, most simulations project a decline...
agupubs.onlinelibrary.wiley.com
March 31, 2025 at 2:48 PM
Our new study is out:
The Arctic Beaufort Gyre is projected to weaken or vanish in most CMIP6 models under future climate change.
The gyre would thus stop storing freshwater in its centre, with possibly large impacts on the Arctic and beyond.
Read: agupubs.onlinelibrary.wiley.com/doi/full/10....
The Arctic Beaufort Gyre is projected to weaken or vanish in most CMIP6 models under future climate change.
The gyre would thus stop storing freshwater in its centre, with possibly large impacts on the Arctic and beyond.
Read: agupubs.onlinelibrary.wiley.com/doi/full/10....
Here is how the 2100 warming outcomes in the new CMIP7 emissions scenarios compare to their CMIP6-era SSP counterparts:
September 1, 2026 at 3:23 PM
Here is how the 2100 warming outcomes in the new CMIP7 emissions scenarios compare to their CMIP6-era SSP counterparts:
However, if we constrain CMIP6 to match recent observed global temperatures, we see no models reaching 3C until at least 2060: www.carbonbrief.org/...
September 28, 2025 at 8:47 PM
However, if we constrain CMIP6 to match recent observed global temperatures, we see no models reaching 3C until at least 2060: www.carbonbrief.org/...
2003 was once exceptional. Now it's average.
2024's record will be average by the mid-2040s.
The extreme summers of the 2040s don't bear thinking about!
🌡🥵😬
#ClimateEmergency
After @edhawkins.org & @stottpeter.bsky.social using ERA5 + CMIP6.
2024's record will be average by the mid-2040s.
The extreme summers of the 2040s don't bear thinking about!
🌡🥵😬
#ClimateEmergency
After @edhawkins.org & @stottpeter.bsky.social using ERA5 + CMIP6.
August 10, 2026 at 7:04 AM
2003 was once exceptional. Now it's average.
2024's record will be average by the mid-2040s.
The extreme summers of the 2040s don't bear thinking about!
🌡🥵😬
#ClimateEmergency
After @edhawkins.org & @stottpeter.bsky.social using ERA5 + CMIP6.
2024's record will be average by the mid-2040s.
The extreme summers of the 2040s don't bear thinking about!
🌡🥵😬
#ClimateEmergency
After @edhawkins.org & @stottpeter.bsky.social using ERA5 + CMIP6.
Here is what the 'cold blob' looks like in terms of full-depth ocean heat content trend since 1955, and the timeline over which the shutdown occurs in state-of-the-art (CMIP6) climate models. 🌊 Sources (open access):
agupubs.onlinelibrary.wiley.com/doi/10.1029/...
iopscience.iop.org/article/10.1...
agupubs.onlinelibrary.wiley.com/doi/10.1029/...
iopscience.iop.org/article/10.1...
July 7, 2026 at 9:45 AM
Here is what the 'cold blob' looks like in terms of full-depth ocean heat content trend since 1955, and the timeline over which the shutdown occurs in state-of-the-art (CMIP6) climate models. 🌊 Sources (open access):
agupubs.onlinelibrary.wiley.com/doi/10.1029/...
iopscience.iop.org/article/10.1...
agupubs.onlinelibrary.wiley.com/doi/10.1029/...
iopscience.iop.org/article/10.1...
Here's the latest model/obs global temperature comparison (through March 2025, via @hausfath.bsky.social). Interestingly, El Niño-spiked 2016 is a bigger outlier than (also El Niño-assisted) 2024/2025 relative to the simulation ensemble.
April 16, 2025 at 8:53 PM
Here's the latest model/obs global temperature comparison (through March 2025, via @hausfath.bsky.social). Interestingly, El Niño-spiked 2016 is a bigger outlier than (also El Niño-assisted) 2024/2025 relative to the simulation ensemble.
Since there are once again some pretty outlandish claims circulating based on papers that have not passed peer review, a reminder that the global temperature record of the past decade is fully consistent with state-of-the-art (CMIP6) climate model simulations: www.pnas.org/doi/10.1073/...
June 2, 2026 at 6:46 PM
Since there are once again some pretty outlandish claims circulating based on papers that have not passed peer review, a reminder that the global temperature record of the past decade is fully consistent with state-of-the-art (CMIP6) climate model simulations: www.pnas.org/doi/10.1073/...
We estimate the probability of shutdown as 70% for high emissions and 25% for low emissions based on the CMIP6 models. These are rough estimates, but main point is: this is a risk that I used to consider <10%, and given the devastating impacts lasting many centuries we really want this to be <1%.
August 28, 2025 at 7:03 PM
We estimate the probability of shutdown as 70% for high emissions and 25% for low emissions based on the CMIP6 models. These are rough estimates, but main point is: this is a risk that I used to consider <10%, and given the devastating impacts lasting many centuries we really want this to be <1%.
Here's an updated comparison (via @hausfath.bsky.social):
January 6, 2025 at 9:23 PM
Here's an updated comparison (via @hausfath.bsky.social):
Global mean surface temperatures are evolving precisely as predicted/projected by the IPCC CMIP6 anthropogenic forcing runs. Watch this space for further details on that...
March 20, 2026 at 2:37 PM
Global mean surface temperatures are evolving precisely as predicted/projected by the IPCC CMIP6 anthropogenic forcing runs. Watch this space for further details on that...
And here's another version of the comparison (via Grant Foster) showing the results using several different alternative observational surface temperature datasets.
May 23, 2025 at 10:02 PM
And here's another version of the comparison (via Grant Foster) showing the results using several different alternative observational surface temperature datasets.
There is much more warming in the pipeline than simple models suggest.
Especially regionally.
Observations trump simple models.
Global averaging of (outdated CMIP6) models is pretty much useless with regards to what most people care about:
understanding real world impacts.
Especially regionally.
Observations trump simple models.
Global averaging of (outdated CMIP6) models is pretty much useless with regards to what most people care about:
understanding real world impacts.
December 14, 2025 at 2:53 PM
There is much more warming in the pipeline than simple models suggest.
Especially regionally.
Observations trump simple models.
Global averaging of (outdated CMIP6) models is pretty much useless with regards to what most people care about:
understanding real world impacts.
Especially regionally.
Observations trump simple models.
Global averaging of (outdated CMIP6) models is pretty much useless with regards to what most people care about:
understanding real world impacts.
Data from CERES show that from 2001 to 2023, Earth absorbed more energy than any CMIP6 model predicted.
Models with ECS < 2.5°C fail to match observed LW & SW trends (LW↓, SW↑). This means future warming may be closer to higher-sensitivity projections.
doi.org/10.1126/scie...
Models with ECS < 2.5°C fail to match observed LW & SW trends (LW↓, SW↑). This means future warming may be closer to higher-sensitivity projections.
doi.org/10.1126/scie...
July 1, 2025 at 2:26 AM
Data from CERES show that from 2001 to 2023, Earth absorbed more energy than any CMIP6 model predicted.
Models with ECS < 2.5°C fail to match observed LW & SW trends (LW↓, SW↑). This means future warming may be closer to higher-sensitivity projections.
doi.org/10.1126/scie...
Models with ECS < 2.5°C fail to match observed LW & SW trends (LW↓, SW↑). This means future warming may be closer to higher-sensitivity projections.
doi.org/10.1126/scie...