Trying something a little different: why do we think of the Sun as yellow and the Moon as white? And why are Martian sunsets blue? notesonclimate.substack.com/p/white-moon...

My first sole-author paper is now published! 🎉🎉🎉 It's the culmination of a lot of thinking, so I hope you enjoy it. Thanks to the friends, colleagues, and reviewers who helped make this happen 😊 agupubs.onlinelibrary.wiley.com/doi/10.1029/...

Interesting new paper by @vtcoop.bsky.social argues high Pliocene sensitivity is mostly a pattern effect from non-CO2 factors, esp ice sheets: www.pnas.org/doi/10.1073/... Interpretation seems subtle: if ice sheets ~stable Pliocene may be a poor analogue for 21st-century warming (too hot)...

@agu.org has a tool to help you submit a comment to NSF on the importance of NCAR: agu.quorum.us/campaign/154... #SaveNCAR

New post on the shape of the water vapor feedback: notesonclimate.substack.com/p/what-contr... Surprisingly, the WV feedback is controlled by surface temperature, not how much moisture is in the air

Bridging Physics & AI in Climate Modeling 🌍🤖 Introducing JCM v1.0: A fully differentiable, intermediate-complexity atmospheric model built in Python/JAX. Training hybrid models is hard because legacy code lacks gradients. JCM solves this.

Do you like charts? Oh yes you do. I've just published hundreds of them, as I do every year. www.nathanielbullard.com/presentations

New piece trying to understand why it took <5 years for ML models to not just match, but beat traditional forecasts notesonclimate.substack.com/p/why-is-wea... My guess is ML's success comes from 3 factors: (1) infrastructure, (2) how it handles uncertainties inherent to forecasting, (3) compute

Some interesting papers from over the break: Karen McKinnon and Isla Simpson showed evidence of substantial continental brightening from 1980 to 2024 that seems to be due to loss of cloud cover (not aerosols), but isn't reproduced by CMIP models agupubs.onlinelibrary.wiley.com/doi/10.1029/...

I came to the US over 10 years ago for a postdoc at NCAR. It changed my life and supercharged my ability to do science. NCAR's unique community nature makes it a global node in atmospheric and climate research. It's really not possible to overstate its value for science and society.

Today, we're rolling out a new suite of AI-driven global weather models, including a "first of its kind" hybrid ensemble system! Want to get geeky with us? Come check it out: www.noaa.gov/news-release...

New paper by Wenyu Zhou finds recent poleward migration of TCs is dominated by Pacific variability www.nature.com/articles/s41... It's awkward that so many of the climate signals we observe are driven by the tropical Pacific, where models struggle but we also expect trends to reverse sign

I've started a Substack to cross-post writing I put up on my website. Hopefully this makes it easier to follow. First one is a re-post of my most recent piece on "Energy Conservation in Climate Models" notesonclimate.substack.com/p/energy-con...

Beautiful analysis once again @hausfath.bsky.social www.carbonbrief.org/analysis-wha...

“Our results enhance the confidence in the climate model projections that the tropics will win the ‘tug-of-war’ and that the zonal-mean jet stream will shift poleward in response to future warming.” 👏 Louder for the people at the back!!📣

It appears that the U.S. Environmental Protection Agency (EPA) has, within the past week, scrubbed a large amount of climate change content from its official website, as well as *removed human-caused warming* from the discussion on its "causes of climate change" page.

We have a new paper showing that Reinforcement Learning (RL) is a promising technique for optimizing solar geoengineering (SAI) deployment. agupubs.onlinelibrary.wiley.com/doi/10.1029/... Very much a proof of concept, this grew out of Heng's undergrad thesis!

Our paper is published! 🎉🎉🎉 If you want to understand changes in total rainfall under warming, give this a read! Happy to chat about it with people :) agupubs.onlinelibrary.wiley.com/doi/10.1029/...

Interesting, the "warming hole" over North America weakened rapidly during 2021-2024 agupubs.onlinelibrary.wiley.com/doi/10.1029/...

Will trying to find holes in catastrophe models replace trying to find holes in climate models?

Nice analysis of high res (10-25km) simulations shows super-CC scaling of precip extremes, driven by mesoscale dynamics not well captured by coarser models Mesoscale systems are an interesting challenge for ML, since they need non-local schemes + some memory www.nature.com/articles/s41...

It‘s strange that this piece by @elizkolbert.bsky.social on the solar geo startup Stardust goes through some of the arguments for/against solar geoengineering, but doesn‘t address whether we should be trusting private companies to develop these technologies www.newyorker.com/news/the-led...

My latest for @science.org: A remarkable set of high-resolution climate model runs, computed over 900 (!) days of supercomputing time, are revealing how warming-induced changes to Earth's wind patterns due can prime huge spikes in extreme rainfall. But the MESACLIP runs also do much more than that.

Google #DeepMind hurricane model has been the most accurate in 2025 www.theguardian.com/technology/2... One reason for this success is hurricane (& weather) forecasting has the perfect infrastructure for ML, but I think there are also some more subtle reasons for why this works so well:

I was part of a fun study led by Yu-Chiao Liang comparing the response to Arctic sea ice loss of Google's NeuralGCM & two models in PAMIP. This is one of the first studies evaluating a hybrid ML climate model at something it wasn't explicitly trained for. Main takeaway: ...

David Keith is right: the problem with solar geoengineering isn’t cost, it’s confidence and trust. SAI is already cheap enough to be a "free driver", which is why SAI research needs to be open, transparent & publicly led (1/2) heatmap.news/climate-tech...

"Through the first 6 months of this year, disasters across the US caused >$100 billion in damage, the most expensive start to any year" -> La Nina events tend to amplify wildfires and inland storms. Luckily it shifted to neutral before hurricane season so we'll likely avoid...

People are often surprised that climate models don’t strictly conserve energy (at least I was). A short post on why atmospheric models “leak” energy, why ocean modelers care more, and how these small biases show up in long control runs. www.sio-climatephysics.com/notes-1/ener...

New paper! How much does uncertainty in AMOC decline contribute to uncertainty in global warming? Led by Lily Hahn, we find: (1) intermodel spread in AMOC decline reproduces 20% of spread in global warming (2) ocean heat uptake & climate feedbacks contribute roughly equally

With about 80% of the precincts reporting I've seen enough. GDMI - Google DeepMind is going to win the seat for best track model in 2025. The race for best intensity model is still too close to call, but GDMI is right there with the consensus and OFCL. Quite a remarkable campaign.

There's a good essay in @science.org about the importance of free global access to climate and weather data. The atmosphere doesn't know about borders, so paywalls anywhere hurt predictions everywhere; even more so for those who can't pay www.science.org/doi/10.1126/...

Amazing news, Angel is such a deserving MacArthur Fellow! Angel has transformed how we think about tropical dynamics and wave-convection interactions, and his work does such a great job blending pen-and-paper theory with observations. Here are a few of my favorite papers of his:

Interesting new @science.org paper by Spencer Hill et al shows that El Nino events weaken India's average monsoon rainfall, but *intensify* extreme precip. They also give a neat mechanism: subsidence during El Niño dries the mid-troposphere over India...

Scientists reanalyzing Cassini data found evidence that Enceladus’ subsurface ocean contains complex prebiotic molecules. Work like this shows the importance of sustained investment in planetary science missions and of good data management and continuity www.theguardian.com/science/2025...

I wrote a short piece on the debate about the future of climate modeling: Should we pool resources to do km-scale "masterpiece" runs or focus on using ML to build data-driven models that can still do large ensembles? 1/3

Some context on the reduction in NIH overhead rates: In 2021-22 UC's operating budget was ~$47 billion of which $987 mill came from indirect cost recovery (~2%). 20% of this went to grant and contract administration, ~45% to UC General Fund and ~35% to the UC "Opportunity Fund"

Hopefully the damage is just 5-10% staff cuts and “searching high and low for DEI” It’s disappointing this is all a group of tech-minded people can think to do. Now should be an exciting time for thinking about modernizing NOAA: a golden age of earth obs +rapid ML advances in WF

2024 was the warmest year on record and it also had the largest recorded increase in atmospheric CO2 (+3.58ppm). These are related: a strong El Niño drove much of the warming and reduced carbon uptake by drying out Northern Hemisphere land

A good paper for learning about connection between California wildfires and climate change is Williams et al (2019). They look at the historical record, but provide a nice overview of the different types of California fires and their climatic drivers…

There’s a new paper in Science adding to the discussion of the “missing” 0.2K in 2023’s extreme warmth. The authors show that the 2023 featured unusually low planetary albedo, linked to a multi-year trend of declining low clouds, particularly over the North Atlantic and tropical oceans (1/5)